video_id
stringlengths
59
122
avg_aesthetic_score_laion_aesthetics
float64
2.78
6.46
frame_aesthetic_scores_laion_aesthetics
stringlengths
72
273
video_name
stringlengths
43
106
video_path
stringlengths
86
149
motion_fb_motion_score
float64
-13.69
24.8
motion_lk_motion_score
float64
0
9.76
frame_shot_categorization_shot_categorizer
stringlengths
276
1.91k
avg_subjects_subject_presence
float64
0
39.3
avg_humans_subject_presence
float64
0
26
avg_animals_subject_presence
float64
0
25.3
frame_counts_subject_presence
stringlengths
268
1.61k
avg_vision_reward
float64
-0.15
0.15
frame_wise_rewards
stringlengths
78
104
video_url
stringlengths
39
39
scene_name
stringclasses
79 values
SCENE_NUM
float64
1
79
START_FRAME
float64
24
13.7k
END_FRAME
float64
73
13.8k
START_TIMECODE
stringlengths
12
12
END_TIMECODE
stringlengths
12
12
START_SECONDS
float64
1
570
END_SECONDS
float64
3.04
574
DURATION_SECONDS
float64
2
6.97
GAP_FROM_PREVIOUS
float64
1
219
prompt_x
stringclasses
1 value
caption_t2v_style
stringlengths
31
2.47k
prompt_y
stringclasses
1 value
caption_t2v_style_short
stringlengths
34
3.66k
duration_seconds
float64
2
6.97
width
int64
640
1.92k
height
int64
262
1.08k
total_frames
int64
48
201
resolution
stringclasses
238 values
short_scene_001_short_A_Series_of_Unfortunate_Events_(1_5)_Movie_CLIP_-_The_Baudelaire_Children_(2004)_HD
4.861241
{'frame_00:00.png': 4.762124538421631, 'frame_00:01.png': 4.502547740936279, 'frame_00:02.png': 4.845088958740234, 'frame_00:03.png': 5.040969371795654, 'frame_00:04.png': 5.155473709106445}
short_A_Series_of_Unfortunate_Events_(1_5)_Movie_CLIP_-_The_Baudelaire_Children_(2004)_HD
../assets/split_scenes/short_A_Series_of_Unfortunate_Events_(1_5)_Movie_CLIP_-_The_Baudelaire_Children_(2004)_HD/short_scene_001.mp4
-0.112536
0.778475
{'frame_00:00.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Soft light, High contrast, Backlight'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Center'}], 'frame_00:01.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Soft light, High contrast, Silhouette, Backlight'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Center'}], 'frame_00:02.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Hard light, High contrast, Silhouette, Backlight'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Center'}], 'frame_00:03.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Soft light, Low contrast, Silhouette, Backlight'}, {'<LIGHTING_TYPE>': 'Daylight, Overcast'}, {'<COMPOSITION>': 'Center'}], 'frame_00:04.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Daylight, Overcast'}, {'<COMPOSITION>': 'Center'}]}
1.4
0.6
0.4
{"0": {"raw_counts": {"person": 1, "bird": 1, "knife": 1, "total": 3}, "bucketed_counts": {"humans": 1, "animals": 1, "vehicles": 0, "food": 0, "others": 1, "total": 3}}, "1": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "2": {"raw_counts": {"total": 0}, "bucketed_counts": {"humans": 0, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 0}}, "3": {"raw_counts": {"person": 1, "snowboard": 1, "total": 2}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 2}}, "4": {"raw_counts": {"bird": 1, "total": 1}, "bucketed_counts": {"humans": 0, "animals": 1, "vehicles": 0, "food": 0, "others": 0, "total": 1}}}
-0.068882
[-1, 1, 1, -1, -1, -1, 1, -1, -1, -1, -1, -1, 1, -1, 1, -1, 1, -1, -1, -1, -1, -1, -1, -1, -1, 1]
https://youtube.com/watch?v=6zqeWbDCXMM
scene_001
1
198
318
00:00:08.258
00:00:13.263
8.258258
13.263263
5.005005
8.258258
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A person closely examines a photograph, holding it carefully between their hands. The photograph depicts a figure standing in a bleak, desolate landscape, possibly a shoreline, under a gloomy sky. The figure is dressed in a long, dark garment. The barren landscape features skeletal trees and eerie, twisted formations that add to the scene's unsettling atmosphere. The lighting on the photograph is subdued, enhancing the somber mood. The person's focused gaze and the visible tension in their expression suggest a deep emotional connection or a significant narrative associated with the image.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A woman holds a photograph of a figure in a dark dress standing on a beach with gnarled trees. The scene is shot in muted tones, creating a somber and melancholic mood.
5.005005
1,604
810
120
1604x810
short_scene_001_short_A_Series_of_Unfortunate_Events_(2_5)_Movie_CLIP_-_Aunt_Josephine_(2004)_HD
4.510507
{'frame_00:00.png': 4.6050705909729, 'frame_00:01.png': 4.618217468261719, 'frame_00:02.png': 4.545652866363525, 'frame_00:03.png': 4.517613887786865, 'frame_00:04.png': 4.308032512664795, 'frame_00:05.png': 4.4684553146362305}
short_A_Series_of_Unfortunate_Events_(2_5)_Movie_CLIP_-_Aunt_Josephine_(2004)_HD
../assets/split_scenes/short_A_Series_of_Unfortunate_Events_(2_5)_Movie_CLIP_-_Aunt_Josephine_(2004)_HD/short_scene_001.mp4
1.566204
1.530879
{'frame_00:00.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Daylight, Overcast'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:01.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Daylight, Overcast'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:02.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Daylight, Overcast'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:03.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Daylight, Overcast'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:04.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Daylight, Overcast'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:05.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Daylight, Overcast'}, {'<COMPOSITION>': 'Left heavy'}]}
3
3
0
{"0": {"raw_counts": {"person": 3, "total": 3}, "bucketed_counts": {"humans": 3, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 3}}, "1": {"raw_counts": {"person": 3, "total": 3}, "bucketed_counts": {"humans": 3, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 3}}, "2": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "3": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "4": {"raw_counts": {"person": 4, "total": 4}, "bucketed_counts": {"humans": 4, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 4}}, "5": {"raw_counts": {"person": 4, "total": 4}, "bucketed_counts": {"humans": 4, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 4}}}
-0.027352
[-1, 1, 1, 1, -1, 1, 1, -1, -1, 1, -1, -1, 1, -1, 1, -1, 1, -1, -1, -1, -1, 1, 1, -1, -1, 1]
https://youtube.com/watch?v=nh8mjiSlAws
scene_001
1
197
332
00:00:08.217
00:00:13.847
8.21655
13.847181
5.630631
8.21655
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A young girl is lifted up by another girl who is holding a suitcase, while a boy stands beside her on a stone bridge that has bare twisted tree decorations; the background reveals large rock faces and a stony ground. The young girl is bundled in a gray cape and a little pink fabric shows around her neck. The second girl is wearing a long gray buttoned dress and her hair is parted in the center with bangs. The boy stands beside her wearing a dark coat and light colored pants.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A young girl is gently picked up and held by an older sister while a boy stands guard as the three survey an eerie landscape. The scene is marked by gloomy skies, rocky terrain, and menacing wrought iron details, contributing to the overall sense of foreboding and uncertainty.
5.630631
1,604
810
135
1604x810
short_scene_001_short_A_Series_of_Unfortunate_Events_(3_5)_Movie_CLIP_-_We're_Too_Late_(2004)_HD
4.612993
{'frame_00:00.png': 4.660000801086426, 'frame_00:01.png': 4.301303863525391, 'frame_00:02.png': 4.481979846954346, 'frame_00:03.png': 4.622464179992676, 'frame_00:04.png': 4.703913688659668, 'frame_00:05.png': 4.908297538757324}
short_A_Series_of_Unfortunate_Events_(3_5)_Movie_CLIP_-_We're_Too_Late_(2004)_HD
../assets/split_scenes/short_A_Series_of_Unfortunate_Events_(3_5)_Movie_CLIP_-_We're_Too_Late_(2004)_HD/short_scene_001.mp4
-0.200421
1.297256
{'frame_00:00.png': [{'<COLOR>': 'Cool, Desaturated, Blue'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Daylight, Overcast'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:01.png': [{'<COLOR>': 'Cool, Desaturated, Blue'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:02.png': [{'<COLOR>': 'Cool, Desaturated'}, {'<LIGHTING>': 'Soft light, Low contrast, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:03.png': [{'<COLOR>': 'Cool, Desaturated'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:04.png': [{'<COLOR>': 'Cool, Desaturated'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:05.png': [{'<COLOR>': 'Cool, Desaturated'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Daylight, Overcast'}, {'<COMPOSITION>': 'Balanced'}]}
4
2
0
{"0": {"raw_counts": {"person": 2, "refrigerator": 1, "total": 3}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 3}}, "1": {"raw_counts": {"refrigerator": 1, "person": 3, "tv": 2, "total": 6}, "bucketed_counts": {"humans": 3, "animals": 0, "vehicles": 0, "food": 0, "others": 3, "total": 6}}, "2": {"raw_counts": {"person": 1, "refrigerator": 1, "tv": 1, "total": 3}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 2, "total": 3}}, "3": {"raw_counts": {"person": 2, "tv": 2, "oven": 1, "total": 5}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 3, "total": 5}}, "4": {"raw_counts": {"person": 2, "refrigerator": 1, "oven": 1, "total": 4}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 2, "total": 4}}, "5": {"raw_counts": {"person": 2, "bottle": 1, "total": 3}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 3}}}
-0.062232
[-1, -1, 1, -1, -1, -1, 1, -1, -1, -1, -1, -1, 1, -1, 1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1, 1]
https://youtube.com/watch?v=mIvQRRVbt9E
scene_001
1
134
263
00:00:05.589
00:00:10.969
5.588922
10.969303
5.38038
5.588922
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
Inside a dimly lit, antiquated house, a young boy, dressed in a dark coat, initially cradles a toddler in a hooded coat. The boy carefully lowers the toddler to the ground, then the two of them look up, off camera, with expressions of concern. Following this, a young girl, wearing a blue coat, steps into the frame to stand next to the boy and she also looks up, off camera, with concern. The scene is framed against a backdrop of faded blue walls, adorned with framed pictures and a vintage refrigerator. A front door with a barred window above it looms in the background, casting long shadows, adding to the scene's eerie atmosphere. The lighting is subdued, highlighting the gravity of the moment as the children look upwards, perhaps at the source of some disquiet or danger.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
Inside a quaint, dimly lit house, a group of children enter a room looking worried. The lighting is soft and muted, setting a suspenseful mood.
5.38038
1,604
810
129
1604x810
short_scene_001_short_A_Series_of_Unfortunate_Events_(5_5)_Movie_CLIP_-_The_Letter_That_Never_Came_(2004)_HD
4.944017
{'frame_00:00.png': 4.831611633300781, 'frame_00:01.png': 4.835056304931641, 'frame_00:02.png': 4.995705604553223, 'frame_00:03.png': 4.989868640899658, 'frame_00:04.png': 5.067842483520508}
short_A_Series_of_Unfortunate_Events_(5_5)_Movie_CLIP_-_The_Letter_That_Never_Came_(2004)_HD
../assets/split_scenes/short_A_Series_of_Unfortunate_Events_(5_5)_Movie_CLIP_-_The_Letter_That_Never_Came_(2004)_HD/short_scene_001.mp4
-0.184805
0.437952
{'frame_00:00.png': [{'<COLOR>': 'Cool, Desaturated'}, {'<LIGHTING>': 'Hard light, High contrast, Backlight'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:01.png': [{'<COLOR>': 'Cool, Desaturated'}, {'<LIGHTING>': 'Hard light, High contrast, Backlight'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:02.png': [{'<COLOR>': 'Cool, Desaturated'}, {'<LIGHTING>': 'Hard light, High contrast, Backlight'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:03.png': [{'<COLOR>': 'Cool, Desaturated'}, {'<LIGHTING>': 'Hard light, High contrast, Backlight'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:04.png': [{'<COLOR>': 'Cool, Desaturated'}, {'<LIGHTING>': 'Hard light, High contrast, Backlight'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Right heavy'}]}
2.2
2
0.2
{"0": {"raw_counts": {"person": 2, "dog": 1, "total": 3}, "bucketed_counts": {"humans": 2, "animals": 1, "vehicles": 0, "food": 0, "others": 0, "total": 3}}, "1": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "2": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "3": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "4": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}}
-0.045726
[-1, 1, 1, -1, -1, 1, 1, -1, -1, -1, -1, -1, 1, -1, 1, -1, -1, -1, -1, -1, -1, 1, 1, -1, -1, 1]
https://youtube.com/watch?v=WrfxIymvEhQ
scene_001
1
89
209
00:00:03.712
00:00:08.717
3.712045
8.71705
5.005005
3.712045
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
Amidst a backdrop of devastation, three siblings stand together in a room marked by charred remains and debris, with a sense of solemnity, the eldest girl is seen holding a toddler protectively in her arms, her expression somber, while the boy beside her carefully examines a small object in his hands, bathed in light streaming from a window that illuminates the scene, highlighting the contrast between innocence and destruction.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
Three young children stand amidst the charred remains of a building, with dramatic lighting casting stark shadows across the scene as the boy examines a small object. The girl holds a toddler protectively as they look around at the devastation.
5.005005
1,604
810
120
1604x810
short_scene_001_short_A_Shot_in_the_Dark_(1964)_-_Naked_Right_Down_To_Your_Mustache_Scene_(6_11)_|_Movieclips
5.008996
{'frame_00:00.png': 4.53738260269165, 'frame_00:01.png': 4.952462673187256, 'frame_00:02.png': 5.11564826965332, 'frame_00:03.png': 5.430492401123047}
short_A_Shot_in_the_Dark_(1964)_-_Naked_Right_Down_To_Your_Mustache_Scene_(6_11)_|_Movieclips
../assets/split_scenes/short_A_Shot_in_the_Dark_(1964)_-_Naked_Right_Down_To_Your_Mustache_Scene_(6_11)_|_Movieclips/short_scene_001.mp4
0.024556
1.247115
{'frame_00:00.png': [{'<COLOR>': 'Cool, Green, Cyan'}, {'<LIGHTING>': 'Soft light'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:01.png': [{'<COLOR>': 'Green'}, {'<LIGHTING>': 'Hard light, Side light'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:02.png': [{'<COLOR>': 'Green'}, {'<LIGHTING>': 'Hard light, Side light'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Center'}], 'frame_00:03.png': [{'<COLOR>': 'Green'}, {'<LIGHTING>': 'Hard light, Side light'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Center'}]}
0.75
0.75
0
{"0": {"raw_counts": {"total": 0}, "bucketed_counts": {"humans": 0, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 0}}, "1": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "2": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "3": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}}
0.015997
[-1 1 1 1 -1 1 1 -1 -1 1 -1 1 1 -1 1 -1 1 1 1 -1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=ZjPPfwVPTDE
scene_001
1
226
320
00:00:09.426
00:00:13.347
9.426093
13.34668
3.920587
9.426093
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A man dressed in a tan trench coat and brown hat enters a log cabin through a partially open white door. The door leads to a lush green outdoor space filled with trees and bushes. The man pauses in the doorway, looking slightly to the right with a somewhat concerned expression on his face. The interior of the cabin is made of wood, and there's a shower stall covered with a clear plastic curtain to the right of the door. The lighting is soft, creating a somewhat subdued mood.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A person wearing a trench coat and fedora hat walks through a door and into a log cabin-style house. The individual looks towards the camera with a serious expression on their face.
3.920587
1,920
836
94
1920x836
short_scene_001_short_A_Shot_in_the_Dark_(1964)_-_Naked_Traffic_Jam_Scene_(8_11)_|_Movieclips
4.624056
{'frame_00:00.png': 4.583851337432861, 'frame_00:01.png': 4.634775638580322, 'frame_00:02.png': 4.65354061126709}
short_A_Shot_in_the_Dark_(1964)_-_Naked_Traffic_Jam_Scene_(8_11)_|_Movieclips
../assets/split_scenes/short_A_Shot_in_the_Dark_(1964)_-_Naked_Traffic_Jam_Scene_(8_11)_|_Movieclips/short_scene_001.mp4
0.338302
0.586158
{'frame_00:00.png': [{'<COLOR>': 'Blue'}, {'<LIGHTING>': 'Hard light, Top light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:01.png': [{'<COLOR>': 'Blue'}, {'<LIGHTING>': 'Hard light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:02.png': [{'<COLOR>': 'Blue'}, {'<LIGHTING>': 'Hard light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Balanced'}]}
5.333333
4.666667
0
{"0": {"raw_counts": {"person": 6, "total": 6}, "bucketed_counts": {"humans": 6, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 6}}, "1": {"raw_counts": {"person": 4, "bus": 1, "total": 5}, "bucketed_counts": {"humans": 4, "animals": 0, "vehicles": 1, "food": 0, "others": 0, "total": 5}}, "2": {"raw_counts": {"person": 4, "bus": 1, "total": 5}, "bucketed_counts": {"humans": 4, "animals": 0, "vehicles": 1, "food": 0, "others": 0, "total": 5}}}
0.072755
[-1 1 1 1 1 1 1 -1 -1 1 1 1 1 -1 1 -1 1 -1 1 1 -1 1 1 1 1 1]
https://youtube.com/watch?v=E3yX8lT2UAI
scene_001
1
95
158
00:00:03.962
00:00:06.590
3.962296
6.589923
2.627628
3.962296
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
Passengers ride a green bus with white trim, identifiable by its route signage displaying "PI GALLE - PI. DE CLICHY." Spectators stand along the roadside, capturing the scene with cameras. The sunlight illuminates the bustling urban environment, casting shadows that add depth to the composition. The atmosphere suggests a vibrant city scene, capturing a moment of movement and observation.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
The frame captures a vintage tour bus filled with passengers, presumably tourists, gazing out with expressions of awe and excitement. Onlookers with cameras are also gathered, perhaps capturing the bus as it moves through the European city, basking in daylight.
2.627628
1,920
784
63
1920x784
short_scene_001_short_A_Shot_in_the_Dark_(1964)_-_Nothing_Matters_But_The_Facts_Scene_(3_11)_|_Movieclips
4.958312
{'frame_00:00.png': 5.0350189208984375, 'frame_00:01.png': 4.915504455566406, 'frame_00:02.png': 4.943445205688477, 'frame_00:03.png': 4.9680986404418945, 'frame_00:04.png': 4.92949104309082}
short_A_Shot_in_the_Dark_(1964)_-_Nothing_Matters_But_The_Facts_Scene_(3_11)_|_Movieclips
../assets/split_scenes/short_A_Shot_in_the_Dark_(1964)_-_Nothing_Matters_But_The_Facts_Scene_(3_11)_|_Movieclips/short_scene_001.mp4
0.526914
0.135243
{'frame_00:00.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:01.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:02.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:03.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:04.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}]}
2.2
1.2
0
{"0": {"raw_counts": {"person": 1, "tie": 1, "total": 2}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 2}}, "1": {"raw_counts": {"person": 2, "tie": 1, "total": 3}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 3}}, "2": {"raw_counts": {"person": 1, "tie": 1, "total": 2}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 2}}, "3": {"raw_counts": {"person": 1, "tie": 1, "total": 2}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 2}}, "4": {"raw_counts": {"person": 1, "tie": 1, "total": 2}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 2}}}
0.084934
[-1 1 1 1 1 1 1 -1 -1 1 1 1 1 1 1 -1 -1 1 1 1 -1 1 1 1 1 1]
https://youtube.com/watch?v=0oFdsgLP8n8
scene_001
1
2,987
3,095
00:02:04.583
00:02:09.087
124.582916
129.087421
4.504505
124.582916
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A man in a dark suit stands in front of a large blackboard covered in chalk drawings and notes. The blackboard shows a detailed floor plan of a building, along with a list of names and job titles, and another list titled "Essential Elements of Information." To the right of the man, a coat rack holds a light-colored trench coat and a hat. The man looks thoughtful and slightly amused, glancing upward as if considering something. The scene is brightly lit, highlighting the details on the blackboard and the man's expression.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A man in a suit stands beside a chalkboard covered with diagrams and notes. The scene is bathed in neutral lighting.
4.504505
1,920
810
108
1920x810
short_scene_001_short_A_Shot_in_the_Dark_(1964)_-_Sneak_Attack_Scene_(2_11)_|_Movieclips
4.860338
{'frame_00:00.png': 4.812592029571533, 'frame_00:01.png': 4.8675007820129395, 'frame_00:02.png': 4.891608715057373, 'frame_00:03.png': 4.887730121612549, 'frame_00:04.png': 4.842260837554932}
short_A_Shot_in_the_Dark_(1964)_-_Sneak_Attack_Scene_(2_11)_|_Movieclips
../assets/split_scenes/short_A_Shot_in_the_Dark_(1964)_-_Sneak_Attack_Scene_(2_11)_|_Movieclips/short_scene_001.mp4
0.061716
0.009446
{'frame_00:00.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Hard light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:01.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Hard light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:02.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Hard light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:03.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Hard light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:04.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Hard light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Left heavy'}]}
2
1
1
{"0": {"raw_counts": {"dog": 1, "person": 1, "total": 2}, "bucketed_counts": {"humans": 1, "animals": 1, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "1": {"raw_counts": {"person": 1, "dog": 1, "total": 2}, "bucketed_counts": {"humans": 1, "animals": 1, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "2": {"raw_counts": {"dog": 1, "person": 1, "total": 2}, "bucketed_counts": {"humans": 1, "animals": 1, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "3": {"raw_counts": {"dog": 1, "person": 1, "total": 2}, "bucketed_counts": {"humans": 1, "animals": 1, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "4": {"raw_counts": {"person": 1, "dog": 1, "total": 2}, "bucketed_counts": {"humans": 1, "animals": 1, "vehicles": 0, "food": 0, "others": 0, "total": 2}}}
0.034188
[-1 1 1 1 -1 1 1 -1 -1 1 1 1 1 1 1 -1 -1 -1 1 -1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=TmALN8vdU0s
scene_001
1
313
412
00:00:13.055
00:00:17.184
13.054721
17.183851
4.129129
13.054721
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A man with a mustache lies in bed, covered by a white robe and a sheet, his head resting on a pillow. The lighting is dim, with shadows cast across his face. The bed's headboard is visible behind him, featuring decorative elements. He appears to be sleeping, his expression relaxed, as the camera focuses on his face and upper body.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A man with a mustache lies in bed, eyes closed at first but then open in the dim light. He wears a white robe and rests on a pillow in what appears to be a vintage or old-fashioned bedroom setting.
4.129129
1,920
810
99
1920x810
short_scene_001_short_A_Simple_Favor_(2018)_-_Drinks_and_Death_Scene_(6_10)_|_Movieclips
5.274651
{'frame_00:00.png': 5.02495813369751, 'frame_00:01.png': 5.284955978393555, 'frame_00:02.png': 5.514039039611816}
short_A_Simple_Favor_(2018)_-_Drinks_and_Death_Scene_(6_10)_|_Movieclips
../assets/split_scenes/short_A_Simple_Favor_(2018)_-_Drinks_and_Death_Scene_(6_10)_|_Movieclips/short_scene_001.mp4
-1.443049
0.302785
{'frame_00:00.png': [{'<COLOR>': 'Green, White'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Daylight, Overcast'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:01.png': [{'<COLOR>': 'Green, White'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Daylight, Overcast'}, {'<COMPOSITION>': 'Center'}], 'frame_00:02.png': [{'<COLOR>': 'Green, White'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Center'}]}
2.666667
1
0
{"0": {"raw_counts": {"person": 1, "bowl": 1, "total": 2}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 2}}, "1": {"raw_counts": {"person": 1, "bottle": 1, "bowl": 1, "total": 3}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 2, "total": 3}}, "2": {"raw_counts": {"person": 1, "bottle": 1, "cup": 1, "total": 3}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 2, "total": 3}}}
0.06925
[ 1 1 1 1 1 1 1 -1 1 1 1 1 1 -1 1 -1 1 1 1 1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=u6IAct0ow4c
scene_001
1
75
140
00:00:03.128
00:00:05.839
3.128128
5.839173
2.711044
3.128128
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A woman dressed in a white pinstripe suit is preparing drinks. She is standing at a table outside, with a cocktail shaker and two martini glasses in front of her. The glasses are garnished with lemon twists. She looks down at the glasses and then looks up, with a slightly concerned expression on her face. The background is a blurred out green, suggesting a garden or park. The light is soft and diffused, creating a calm atmosphere.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A woman wearing a white suit stands in an outdoor setting, with trees and foliage visible in the background, as she prepares drinks using a shaker and martini glasses. She has a serious expression, as though contemplating something while preparing the drinks.
2.711044
1,920
810
65
1920x810
short_scene_001_short_A_Simple_Favor_(2018)_-_Emily_Is_Dead_Scene_(1_10)_|_Movieclips
5.184376
{'frame_00:00.png': 5.239075660705566, 'frame_00:01.png': 5.195884704589844, 'frame_00:02.png': 5.154531478881836, 'frame_00:03.png': 5.327116012573242, 'frame_00:04.png': 5.005273818969727}
short_A_Simple_Favor_(2018)_-_Emily_Is_Dead_Scene_(1_10)_|_Movieclips
../assets/split_scenes/short_A_Simple_Favor_(2018)_-_Emily_Is_Dead_Scene_(1_10)_|_Movieclips/short_scene_001.mp4
0.100873
0.201597
{'frame_00:00.png': [{'<COLOR>': 'Green'}, {'<LIGHTING>': 'Soft light, Backlight'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Center, Symmetrical'}], 'frame_00:01.png': [{'<COLOR>': 'Green'}, {'<LIGHTING>': 'Soft light, Backlight'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Center, Symmetrical'}], 'frame_00:02.png': [{'<COLOR>': 'Green'}, {'<LIGHTING>': 'Soft light, Backlight'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Center, Symmetrical'}], 'frame_00:03.png': [{'<COLOR>': 'Green'}, {'<LIGHTING>': 'Soft light, Backlight'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Center, Symmetrical'}], 'frame_00:04.png': [{'<COLOR>': 'Green'}, {'<LIGHTING>': 'Soft light, Backlight'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Center'}]}
5.2
1.2
0
{"0": {"raw_counts": {"person": 1, "chair": 2, "bowl": 1, "microwave": 1, "total": 5}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 4, "total": 5}}, "1": {"raw_counts": {"person": 1, "chair": 2, "bowl": 1, "microwave": 1, "potted plant": 1, "total": 6}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 5, "total": 6}}, "2": {"raw_counts": {"person": 1, "chair": 2, "bowl": 1, "microwave": 1, "potted plant": 1, "total": 6}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 5, "total": 6}}, "3": {"raw_counts": {"chair": 2, "person": 2, "bowl": 1, "microwave": 1, "total": 6}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 4, "total": 6}}, "4": {"raw_counts": {"person": 1, "chair": 1, "bowl": 1, "total": 3}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 2, "total": 3}}}
0.125374
[-1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 -1 -1 1 1 1 1 1 1 1 1 1]
https://youtube.com/watch?v=xDZfwTsiLrk
scene_001
1
49
169
00:00:02.044
00:00:07.049
2.04371
7.048715
5.005005
2.04371
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A woman stands centered in a brightly lit kitchen, addressing someone off-screen. She wears a sleeveless top with a floral pattern, her hands clasped in front of her on the countertop. Behind her, the kitchen is well-appointed with white cabinets, a window looking out to the greenery, and various appliances and decorative items lining the counter. Child's artwork is attached to the cabinetry. The woman's expression shifts subtly as if contemplating what to say next.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A woman stands in a bright kitchen, speaking directly to the camera with a somber expression; she then looks down and clasps her hands. The domestic setting features a cluttered countertop and children's artwork, contributing to the overall mood of the scene.
5.005005
1,920
854
120
1920x854
short_scene_001_short_A_Simple_Favor_(2018)_-_I_Shot_My_Ex-Husband_Scene_(10_10)_|_Movieclips
5.315035
{'frame_00:00.png': 5.350308418273926, 'frame_00:01.png': 5.205815315246582, 'frame_00:02.png': 5.365424156188965, 'frame_00:03.png': 5.340038299560547, 'frame_00:04.png': 5.3135881423950195}
short_A_Simple_Favor_(2018)_-_I_Shot_My_Ex-Husband_Scene_(10_10)_|_Movieclips
../assets/split_scenes/short_A_Simple_Favor_(2018)_-_I_Shot_My_Ex-Husband_Scene_(10_10)_|_Movieclips/short_scene_001.mp4
-0.233022
0.036172
{'frame_00:00.png': [{'<COLOR>': 'Cool, Saturated, Cyan, Blue'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:01.png': [{'<COLOR>': 'Cool, Saturated, Cyan, Blue'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:02.png': [{'<COLOR>': 'Cool, Saturated, Cyan, Blue'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:03.png': [{'<COLOR>': 'Cool, Saturated, Cyan, Blue'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:04.png': [{'<COLOR>': 'Cool, Saturated, Cyan, Blue'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}]}
1
1
0
{"0": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "1": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "2": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "3": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "4": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}}
0.079458
[ 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 -1 -1 1 1 1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=-ZxtmDbqDRc
scene_001
1
108
206
00:00:04.505
00:00:08.592
4.504505
8.591925
4.087421
4.504505
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A woman stands indoors, wearing a dark, long-sleeved blouse with intricate embroidery down the front. The background is a light beige wall, decorated with a large, framed artwork depicting a woman swimming in a pool. To the right of the woman, a small, rectangular speaker is mounted on the wall. Her facial expressions shift from surprise to a smile, then to a more serious, assertive expression, before settling into a thoughtful, slightly concerned look. The lighting is bright and even, illuminating the scene.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A woman stands in front of a wall with a large framed picture. She cycles through different expressions and emotions.
4.087421
1,920
946
98
1920x946
short_scene_001_short_A_Simple_Favor_(2018)_-_Mile_High_Club_Scene_(4_10)_|_Movieclips
4.966726
{'frame_00:00.png': 4.707736492156982, 'frame_00:01.png': 5.026937484741211, 'frame_00:02.png': 5.090744972229004, 'frame_00:03.png': 5.0414838790893555}
short_A_Simple_Favor_(2018)_-_Mile_High_Club_Scene_(4_10)_|_Movieclips
../assets/split_scenes/short_A_Simple_Favor_(2018)_-_Mile_High_Club_Scene_(4_10)_|_Movieclips/short_scene_001.mp4
-0.034396
0.018669
{'frame_00:00.png': [{'<COLOR>': 'Mixed, Saturated'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:01.png': [{'<COLOR>': 'Mixed, Saturated'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:02.png': [{'<COLOR>': 'Mixed, Saturated'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:03.png': [{'<COLOR>': 'Mixed, Saturated'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Right heavy'}]}
1.5
1.5
0
{"0": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "1": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "2": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "3": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}}
0.069348
[-1 1 1 1 1 1 1 -1 -1 1 1 1 1 1 1 -1 -1 1 1 1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=jwjCPSUGPXU
scene_001
1
132
209
00:00:05.506
00:00:08.717
5.505506
8.71705
3.211545
5.505506
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A woman with fair skin and dark hair, dressed in a dark blue top, is seen in a close-up shot, intently listening to someone whose back is facing the camera. The background is a mix of a bright red wall and a neutral-toned wall. Her expression is one of serious consideration, with her eyes focused towards the person she's listening to. The scene is well-lit, highlighting her features and the slight details in her clothing. The tone of the scene suggests a serious conversation or confrontation.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
The shot shows a woman looking at a man, appearing to be in a serious conversation; the mood seems tense and thoughtful. The scene is captured in a close-up.
3.211545
1,920
854
77
1920x854
short_scene_001_short_A_Simple_Favor_(2018)_-_Murder_and_Manipulation_Scene_(8_10)_|_Movieclips
4.525802
{'frame_00:00.png': 4.624925136566162, 'frame_00:01.png': 4.822159767150879, 'frame_00:02.png': 4.344931602478027, 'frame_00:03.png': 4.311191558837891}
short_A_Simple_Favor_(2018)_-_Murder_and_Manipulation_Scene_(8_10)_|_Movieclips
../assets/split_scenes/short_A_Simple_Favor_(2018)_-_Murder_and_Manipulation_Scene_(8_10)_|_Movieclips/short_scene_001.mp4
-0.012168
0.758047
{'frame_00:00.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Center'}], 'frame_00:01.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:02.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Center'}], 'frame_00:03.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Center'}]}
0.75
0.75
0
{"0": {"raw_counts": {"total": 0}, "bucketed_counts": {"humans": 0, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 0}}, "1": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "2": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "3": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}}
0.004544
[-1 1 1 1 -1 1 1 -1 -1 1 -1 -1 1 -1 1 1 1 -1 1 -1 -1 -1 -1 -1 -1 1]
https://youtube.com/watch?v=frV4tvni8H0
scene_001
1
330
418
00:00:13.764
00:00:17.434
13.763764
17.434101
3.670337
13.763764
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A closed white door with a golden number on top is presented, before a person appears from the door wearing a white robe and a red hood, looking surprised as they glance around before exiting to the right. The person walks between two beige walls, with a brick pattern along the bottom.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A woman emerges from a doorway, wrapped in a white robe and a red hooded sweatshirt. She pauses momentarily, glances to the side, and then proceeds to walk away.
3.670337
1,920
810
88
1920x810
short_scene_001_short_A_Simple_Favor_(2018)_-_Skinny_Dipping_Slaughter_Scene_(7_10)_|_Movieclips
5.424387
{'frame_00:00.png': 5.450014114379883, 'frame_00:01.png': 5.398758888244629}
short_A_Simple_Favor_(2018)_-_Skinny_Dipping_Slaughter_Scene_(7_10)_|_Movieclips
../assets/split_scenes/short_A_Simple_Favor_(2018)_-_Skinny_Dipping_Slaughter_Scene_(7_10)_|_Movieclips/short_scene_001.mp4
-0.029273
0.268444
{'frame_00:00.png': [{'<COLOR>': 'Cool, Desaturated, Cyan'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Daylight, Overcast'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:01.png': [{'<COLOR>': 'Cool, Desaturated'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Daylight, Overcast'}, {'<COMPOSITION>': 'Right heavy'}]}
1
0
1
{"0": {"raw_counts": {"dog": 1, "total": 1}, "bucketed_counts": {"humans": 0, "animals": 1, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "1": {"raw_counts": {"dog": 1, "total": 1}, "bucketed_counts": {"humans": 0, "animals": 1, "vehicles": 0, "food": 0, "others": 0, "total": 1}}}
0.03995
[-1 1 1 1 1 1 1 -1 1 1 1 1 1 -1 1 -1 -1 -1 1 -1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=vxFr0xNspFU
scene_001
1
188
236
00:00:07.841
00:00:09.843
7.841175
9.843177
2.002002
7.841175
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A woman with wet hair, her shoulders and chest visible, is seen swimming in a lake surrounded by tall trees. The woman appears to be looking towards something in the distance. Her expression is one of concern or fear. The water is relatively still, with gentle ripples around her. The background features a dense forest, with tall green trees creating a natural backdrop. The scene has a slightly overcast quality, contributing to a sense of unease.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A woman is seen emerging from a lake, framed against a backdrop of lush green trees, her expression suggesting a mix of surprise and alertness. She is surrounded by ripples as the camera captures her movement through the water.
2.002002
1,920
890
48
1920x890
short_scene_001_short_A_Simple_Favor_(2018)_-_Violent_Confessions_Scene_(9_10)_|_Movieclips
5.368378
{'frame_00:00.png': 5.0557780265808105, 'frame_00:01.png': 5.333962440490723, 'frame_00:02.png': 5.563488006591797, 'frame_00:03.png': 5.520284652709961}
short_A_Simple_Favor_(2018)_-_Violent_Confessions_Scene_(9_10)_|_Movieclips
../assets/split_scenes/short_A_Simple_Favor_(2018)_-_Violent_Confessions_Scene_(9_10)_|_Movieclips/short_scene_001.mp4
0.119428
1.402914
{'frame_00:00.png': [{'<COLOR>': 'Green, Cyan, Blue'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:01.png': [{'<COLOR>': 'Green, Cyan, Blue'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:02.png': [{'<COLOR>': 'Green, Blue'}, {'<LIGHTING>': 'Soft light, Low contrast, Side light'}, {'<LIGHTING_TYPE>': 'Daylight, Artificial light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:03.png': [{'<COLOR>': 'Blue'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Center'}]}
1.5
1.5
0
{"0": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "1": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "2": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "3": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}}
0.083437
[ 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 -1 1 1 1 1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=pppK-fl9a2E
scene_001
1
39
131
00:00:01.627
00:00:05.464
1.626627
5.463797
3.837171
1.626627
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
The scene features a blonde woman with fair skin and pink lipstick, facing slightly to the right of the frame, with a noticeable bruise near her left eye. She is wearing a collared shirt with a floral pattern. The background is softly blurred, revealing hints of an indoor setting, possibly a kitchen or living space, with light-colored walls and cabinets. The lighting is soft and diffused, highlighting the woman's features while subtly casting shadows that accentuate her expression. As the scene progresses, a man's face, mostly out of focus, enters the frame from the left. He is wearing a white collared shirt and stands close to the woman, suggesting an intimate or confrontational interaction.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A woman with a bruise near her eye stands in what appears to be a kitchen, speaking to someone off-screen; the scene is intimate and tense. The woman looks concerned as she engages in conversation.
3.837171
1,920
936
92
1920x936
short_scene_001_short_A_Star_Is_Born_(2018)_-_It_Isn't_Your_Fault_Scene_(7_7)_|_Movieclips
4.724096
{'frame_00:00.png': 4.725419044494629, 'frame_00:01.png': 4.8302388191223145, 'frame_00:02.png': 4.790652751922607, 'frame_00:03.png': 4.5874924659729, 'frame_00:04.png': 4.686676502227783}
short_A_Star_Is_Born_(2018)_-_It_Isn't_Your_Fault_Scene_(7_7)_|_Movieclips
../assets/split_scenes/short_A_Star_Is_Born_(2018)_-_It_Isn't_Your_Fault_Scene_(7_7)_|_Movieclips/short_scene_001.mp4
0.016003
0.069463
{'frame_00:00.png': [{'<COLOR>': 'Warm'}, {'<LIGHTING>': 'Hard light, High contrast, Side light, Edge light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy, Short side'}], 'frame_00:01.png': [{'<COLOR>': 'Warm'}, {'<LIGHTING>': 'Hard light, High contrast, Side light, Edge light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy, Short side'}], 'frame_00:02.png': [{'<COLOR>': 'Warm'}, {'<LIGHTING>': 'Hard light, High contrast, Side light, Edge light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy, Short side'}], 'frame_00:03.png': [{'<COLOR>': 'Warm'}, {'<LIGHTING>': 'Hard light, High contrast, Side light, Edge light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy, Short side'}], 'frame_00:04.png': [{'<COLOR>': 'Warm'}, {'<LIGHTING>': 'Hard light, High contrast, Side light, Edge light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy, Short side'}]}
1.2
1
0
{"0": {"raw_counts": {"person": 1, "bottle": 1, "total": 2}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 2}}, "1": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "2": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "3": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "4": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}}
0.027422
[-1 1 1 1 -1 1 1 -1 1 1 1 1 1 -1 1 -1 -1 -1 1 -1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=Ixggr1vrTUQ
scene_001
1
165
285
00:00:06.882
00:00:11.887
6.881882
11.886887
5.005005
6.881882
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A close-up shows a woman with light orange hair pulled back, wearing a tan shirt. She sits indoors with her head slightly bowed and eyes closed. Sunlight illuminates the side of her face, creating a warm glow. Her expression is subtle, hinting at both sadness and a slight smile. Behind her, an out-of-focus tapestry with geometric designs in shades of red and tan adds depth to the scene.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A person with auburn hair is seen in close-up, with moist eyes, suggesting an emotional moment or scene. The scene is illuminated with soft, dramatic lighting, creating a tender and melancholic atmosphere.
5.005005
1,920
810
120
1920x810
short_scene_001_short_A_Star_Is_Born_(2018)_-_Jackson_in_Rehab_Scene_(5_7)_|_Movieclips
5.299374
{'frame_00:00.png': 5.570040702819824, 'frame_00:01.png': 5.604582786560059, 'frame_00:02.png': 5.405454158782959, 'frame_00:03.png': 5.014078140258789, 'frame_00:04.png': 4.902716159820557}
short_A_Star_Is_Born_(2018)_-_Jackson_in_Rehab_Scene_(5_7)_|_Movieclips
../assets/split_scenes/short_A_Star_Is_Born_(2018)_-_Jackson_in_Rehab_Scene_(5_7)_|_Movieclips/short_scene_001.mp4
-0.060661
0.229963
{'frame_00:00.png': [{'<COLOR>': 'Blue, White'}, {'<LIGHTING>': 'Soft light, Side light, Edge light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:01.png': [{'<COLOR>': 'Blue, White'}, {'<LIGHTING>': 'Soft light, Side light, Edge light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:02.png': [{'<COLOR>': 'Blue, White'}, {'<LIGHTING>': 'Soft light, Side light, Edge light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:03.png': [{'<COLOR>': 'Blue, White'}, {'<LIGHTING>': 'Hard light, Side light, Edge light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:04.png': [{'<COLOR>': 'Blue, White'}, {'<LIGHTING>': 'Soft light, Backlight, Edge light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Right heavy'}]}
1
1
0
{"0": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "1": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "2": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "3": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "4": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}}
0.072037
[ 1 1 1 1 1 1 1 -1 -1 1 1 1 1 1 1 -1 -1 1 1 1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=RzKO_yFUe4I
scene_001
1
54
174
00:00:02.252
00:00:07.257
2.252252
7.257257
5.005005
2.252252
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A man with light eyes and medium-length brown hair combed back away from his forehead appears seated indoors, wearing a blue t-shirt, as he looks directly into the camera with an intense expression, then seems to soften with a slight downturn of the lips as he looks down and to the side; the scene is brightly lit by natural light coming from vertical blinds in the background, with the warm light of a table lamp visible on the right.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A bearded man wearing a blue shirt is captured in close-up. He appears to be looking downwards before briefly making eye contact with the camera.
5.005005
1,920
774
120
1920x774
short_scene_001_short_A_Streetcar_Named_Desire_(1951)_-_A_Streetcar_Named_Desire_Scene_|_Movieclips
5.279939
{'frame_00:00.png': 5.74923038482666, 'frame_00:01.png': 5.585073947906494, 'frame_00:02.png': 4.964493274688721, 'frame_00:03.png': 4.998251438140869, 'frame_00:04.png': 5.102645397186279}
short_A_Streetcar_Named_Desire_(1951)_-_A_Streetcar_Named_Desire_Scene_|_Movieclips
../assets/split_scenes/short_A_Streetcar_Named_Desire_(1951)_-_A_Streetcar_Named_Desire_Scene_|_Movieclips/short_scene_001.mp4
0.032387
0.719013
{'frame_00:00.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Center'}], 'frame_00:01.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Center'}], 'frame_00:02.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:03.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:04.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Balanced'}]}
1.4
1.4
0
{"0": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "1": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "2": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "3": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "4": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}}
-0.013384
[-1 1 1 -1 -1 -1 1 -1 -1 1 -1 -1 1 -1 1 -1 1 -1 1 -1 -1 -1 1 -1 -1 1]
https://youtube.com/watch?v=xC9lZejXs4g
scene_001
1
164
284
00:00:06.840
00:00:11.845
6.840174
11.845179
5.005005
6.840174
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A woman is featured in a medium shot, standing indoors and holding a lit cigarette while draped in a semi-transparent shawl that reveals her shoulders. The woman's expression conveys a look of concern or apprehension, and she glances off-camera to her right, where a man in a suit appears. The man is looking at her with a half-smile as they engage in conversation. The scene is shot in black and white, and the background is softly blurred, focusing the viewer's attention on the interaction between the two individuals.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A woman smoking a cigarette looks to her right with a concerned expression; a man in a suit stands nearby and looks at her. The scene is shot in black and white with dramatic lighting and evokes a sense of tension or uncertainty.
5.005005
1,440
810
120
1440x810
short_scene_001_short_A_Streetcar_Named_Desire_(1951)_-_He's_Like_an_Animal_Scene_|_Movieclips
5.540249
{'frame_00:00.png': 5.508567810058594, 'frame_00:01.png': 5.601825714111328, 'frame_00:02.png': 5.548437118530273, 'frame_00:03.png': 5.502163410186768}
short_A_Streetcar_Named_Desire_(1951)_-_He's_Like_an_Animal_Scene_|_Movieclips
../assets/split_scenes/short_A_Streetcar_Named_Desire_(1951)_-_He's_Like_an_Animal_Scene_|_Movieclips/short_scene_001.mp4
-0.107919
0.759708
{'frame_00:00.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Center'}], 'frame_00:01.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Side light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Center'}], 'frame_00:02.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Side light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Center'}], 'frame_00:03.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Side light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Left heavy'}]}
3
3
0
{"0": {"raw_counts": {"person": 4, "total": 4}, "bucketed_counts": {"humans": 4, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 4}}, "1": {"raw_counts": {"person": 3, "total": 3}, "bucketed_counts": {"humans": 3, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 3}}, "2": {"raw_counts": {"person": 3, "total": 3}, "bucketed_counts": {"humans": 3, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 3}}, "3": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}}
0.032628
[-1 1 1 -1 -1 1 1 -1 -1 1 1 1 1 -1 1 -1 1 -1 1 -1 -1 1 1 -1 1 1]
https://youtube.com/watch?v=yHs5FYRJTQw
scene_001
1
140
222
00:00:05.839
00:00:09.259
5.839173
9.259259
3.420087
5.839173
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
The frame presents a man, caught in a medium shot, his gaze directed slightly downward and to his left. He's clad in a light-colored tank top and a cap, his attire smudged with dark marks. The setting is an outdoor, urban environment, indicated by the presence of buildings and a vintage gas pump to his right, complete with a warning label. In the background, other figures are visible, suggesting a bustling yet indistinct scene. The shot is a monochrome rendering, lending it a classic, timeless feel.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A sweaty man wearing a tank top and cap stands outside with a serious expression, his clothes slightly dirty. The scene is in black and white, with a gas pump visible in the background.
3.420087
1,440
810
82
1440x810
short_scene_001_short_A_Streetcar_Named_Desire_(1951)_-_I'm_the_King_Around_Here_Scene_|_Movieclips
5.116955
{'frame_00:00.png': 5.090612411499023, 'frame_00:01.png': 5.078578948974609, 'frame_00:02.png': 5.1094560623168945, 'frame_00:03.png': 5.101639747619629, 'frame_00:04.png': 5.2044878005981445}
short_A_Streetcar_Named_Desire_(1951)_-_I'm_the_King_Around_Here_Scene_|_Movieclips
../assets/split_scenes/short_A_Streetcar_Named_Desire_(1951)_-_I'm_the_King_Around_Here_Scene_|_Movieclips/short_scene_001.mp4
-0.749
0.16451
{'frame_00:00.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:01.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:02.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:03.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:04.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Balanced'}]}
3.2
2.2
0
{"0": {"raw_counts": {"person": 2, "cup": 1, "total": 3}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 3}}, "1": {"raw_counts": {"person": 2, "cup": 1, "total": 3}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 3}}, "2": {"raw_counts": {"person": 3, "cup": 1, "total": 4}, "bucketed_counts": {"humans": 3, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 4}}, "3": {"raw_counts": {"person": 2, "cup": 1, "total": 3}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 3}}, "4": {"raw_counts": {"person": 2, "cup": 1, "total": 3}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 3}}}
-0.03381
[-1 1 1 -1 -1 1 1 -1 -1 1 -1 -1 1 -1 1 -1 -1 -1 -1 -1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=JpNLKb3rHEg
scene_001
1
432
552
00:00:18.018
00:00:23.023
18.018018
23.023023
5.005005
18.018018
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
The scene unfolds in black and white, capturing a moment of intense dialogue between a man and a woman seated at a table. The man, positioned on the left, is deep in conversation, gesturing with his hand as he speaks. The woman, on the right, listens intently, her expression a mix of concern and captivation. She clutches her hands close to her chest, a ring visible on her finger. The setting appears to be indoors, with decorative elements in the background adding depth to the composition. The lighting casts dramatic shadows, emphasizing the emotional weight of their exchange. The scene feels intimate and fraught with unspoken tensions.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A man is shown sitting at a table, making hand gestures while speaking to a woman who is sitting across from him, smiling with her hands clasped together. The black-and-white still is set indoors, with the subjects sitting in front of some furniture and what seems to be a window.
5.005005
1,440
836
120
1440x836
short_scene_001_short_A_Streetcar_Named_Desire_(1951)_-_Meetings_with_Strangers_Scene_|_Movieclips
4.79398
{'frame_00:00.png': 4.632222652435303, 'frame_00:01.png': 4.545118808746338, 'frame_00:02.png': 4.914468288421631, 'frame_00:03.png': 4.980693340301514, 'frame_00:04.png': 4.897396564483643}
short_A_Streetcar_Named_Desire_(1951)_-_Meetings_with_Strangers_Scene_|_Movieclips
../assets/split_scenes/short_A_Streetcar_Named_Desire_(1951)_-_Meetings_with_Strangers_Scene_|_Movieclips/short_scene_001.mp4
0.539881
0.340468
{'frame_00:00.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, High contrast, Top light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:01.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, High contrast, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:02.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, High contrast, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:03.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, High contrast, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:04.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, High contrast, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}]}
2.4
2
0
{"0": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "1": {"raw_counts": {"person": 3, "bottle": 1, "cell phone": 1, "total": 5}, "bucketed_counts": {"humans": 3, "animals": 0, "vehicles": 0, "food": 0, "others": 2, "total": 5}}, "2": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "3": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "4": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}}
-0.068882
[-1 1 1 -1 -1 -1 1 -1 -1 -1 -1 -1 1 -1 1 -1 1 -1 -1 -1 -1 -1 -1 -1 -1 1]
https://youtube.com/watch?v=-TVicuP4BZQ
scene_001
1
995
1,115
00:00:41.500
00:00:46.505
41.499833
46.504838
5.005005
41.499833
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A dramatic scene unfolds between a woman and a man, filmed in black and white, with dramatic lighting casting deep shadows. The woman, with wavy hair and wearing a necklace, stands close to the man, who is partially obscured by darkness. The camera captures a sense of tension and intimacy between them, as they stand near a table adorned with flowers and fans, enhancing the scene's depth and complexity. Their expressions are serious, suggesting a profound exchange or confrontation is taking place.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A man and a woman stand indoors together, sharing a drink in a scene with dramatic lighting and a tense mood. The man looks sternly at the woman, who appears contemplative.
5.005005
1,440
810
120
1440x810
short_scene_001_short_A_Streetcar_Named_Desire_(1951)_-_Pearls_Before_Swine_Scene_|_Movieclips
5.742463
{'frame_00:00.png': 5.783283233642578, 'frame_00:01.png': 5.7016425132751465}
short_A_Streetcar_Named_Desire_(1951)_-_Pearls_Before_Swine_Scene_|_Movieclips
../assets/split_scenes/short_A_Streetcar_Named_Desire_(1951)_-_Pearls_Before_Swine_Scene_|_Movieclips/short_scene_001.mp4
0.190184
0.014635
{'frame_00:00.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Top light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light'}, {'<COMPOSITION>': 'Center'}], 'frame_00:01.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Top light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light'}, {'<COMPOSITION>': 'Center'}]}
1
1
0
{"0": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "1": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}}
-0.000041
[-1 1 1 -1 -1 1 1 -1 -1 1 -1 -1 1 -1 1 -1 -1 -1 1 -1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=pnoNj9OU3_g
scene_001
1
207
255
00:00:08.634
00:00:10.636
8.633634
10.635636
2.002002
8.633634
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A man wearing a white tank top is shown in a black and white medium shot, with dramatic lighting casting shadows across his face and body. He is in the middle of eating something, as evidenced by the food held in his hand and a piece visible in his open mouth. Behind him, an ornate chandelier hangs from the ceiling, and a dark wooden chest sits on a table in the background. The scene has a tense, theatrical atmosphere.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A man wearing a tank top is eating while standing in a room. The shot is captured in black and white with a chandelier visible in the background.
2.002002
1,440
810
48
1440x810
short_scene_001_short_A_Streetcar_Named_Desire_(1951)_-_The_Napoleonic_Code_Scene_|_Movieclips
4.908684
{'frame_00:00.png': 5.103056907653809, 'frame_00:01.png': 5.380152702331543, 'frame_00:02.png': 5.2197265625, 'frame_00:03.png': 4.212281703948975, 'frame_00:04.png': 4.628204345703125}
short_A_Streetcar_Named_Desire_(1951)_-_The_Napoleonic_Code_Scene_|_Movieclips
../assets/split_scenes/short_A_Streetcar_Named_Desire_(1951)_-_The_Napoleonic_Code_Scene_|_Movieclips/short_scene_001.mp4
0.150564
2.060045
{'frame_00:00.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:01.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Center'}], 'frame_00:02.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Top light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Center'}], 'frame_00:03.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Center'}], 'frame_00:04.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Top light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Center'}]}
2
1.6
0
{"0": {"raw_counts": {"person": 1, "chair": 1, "total": 2}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 2}}, "1": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "2": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "3": {"raw_counts": {"person": 2, "bottle": 1, "total": 3}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 3}}, "4": {"raw_counts": {"person": 3, "total": 3}, "bucketed_counts": {"humans": 3, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 3}}}
-0.028464
[-1 -1 1 -1 -1 -1 1 -1 -1 -1 -1 -1 1 -1 1 1 1 -1 1 -1 -1 -1 -1 -1 -1 1]
https://youtube.com/watch?v=AyBH76_kDYo
scene_001
1
259
379
00:00:10.802
00:00:15.807
10.802469
15.807474
5.005005
10.802469
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A black and white scene presents a muscular man wearing a tight t-shirt, standing in what appears to be a modest interior setting with draperies and a framed window in the background. He initially looks down, seemingly lost in thought, before glancing to his right and beginning to speak. As the scene progresses, a woman with short, wavy hair enters the frame, wearing a slip and a string of pearls. She avoids eye contact, looking downwards. The room is cluttered with objects on a table, including a small statue of an athlete and a framed photograph of a military man. The lighting is moody, emphasizing the textures of the room and the emotions of the characters. The scene evokes a sense of tension and quiet drama, focusing on the body language and interactions of the two individuals.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A man in a t-shirt stands in an indoor setting; he walks around as a woman enters the frame. The scene, filmed in black and white, creates a dramatic mood.
5.005005
1,440
810
120
1440x810
short_scene_001_short_A_Streetcar_Named_Desire_(1951)_-_You_Must_Be_Stanley_Scene_|_Movieclips
5.384301
{'frame_00:00.png': 5.30588436126709, 'frame_00:01.png': 5.32750129699707, 'frame_00:02.png': 5.304106712341309, 'frame_00:03.png': 5.599710464477539}
short_A_Streetcar_Named_Desire_(1951)_-_You_Must_Be_Stanley_Scene_|_Movieclips
../assets/split_scenes/short_A_Streetcar_Named_Desire_(1951)_-_You_Must_Be_Stanley_Scene_|_Movieclips/short_scene_001.mp4
0.688243
0.24473
{'frame_00:00.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:01.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:02.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Center'}], 'frame_00:03.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Center'}]}
3
1
0
{"0": {"raw_counts": {"person": 1, "bottle": 2, "potted plant": 1, "total": 4}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 3, "total": 4}}, "1": {"raw_counts": {"person": 1, "bottle": 2, "potted plant": 1, "total": 4}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 3, "total": 4}}, "2": {"raw_counts": {"person": 1, "bottle": 1, "total": 2}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 2}}, "3": {"raw_counts": {"person": 1, "bottle": 1, "total": 2}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 2}}}
0.010897
[-1 1 1 -1 -1 1 1 -1 -1 1 -1 1 1 -1 1 -1 1 -1 1 -1 -1 -1 1 -1 -1 1]
https://youtube.com/watch?v=djXsEQ5CDOE
scene_001
1
133
216
00:00:05.547
00:00:09.009
5.547214
9.009009
3.461795
5.547214
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A monochromatic medium shot captures a man in a light-colored, satin bomber jacket. He is centrally framed against a textured wall, adjacent to draped curtains. A vintage oscillating fan sits on a table to his right, near a few bottles. The man's expression shifts subtly, conveying a sense of contemplation, as he turns his head slightly. The backdrop features a softly lit street scene, visible through an open doorway or window, adding depth to the composition.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A man stands in a dimly lit room, looking off-camera with an earnest expression. The scene is characterized by dramatic lighting and a sense of introspective tension.
3.461795
1,440
810
83
1440x810
short_scene_001_short_A_Thousand_Words_(2012)_-_Baby_Back_Ribs_Scene_(2_10)_|_Movieclips
5.696349
{'frame_00:00.png': 5.57009220123291, 'frame_00:01.png': 5.9205780029296875, 'frame_00:02.png': 5.598377227783203}
short_A_Thousand_Words_(2012)_-_Baby_Back_Ribs_Scene_(2_10)_|_Movieclips
../assets/split_scenes/short_A_Thousand_Words_(2012)_-_Baby_Back_Ribs_Scene_(2_10)_|_Movieclips/short_scene_001.mp4
0.180977
0.246664
{'frame_00:00.png': [{'<COLOR>': 'Warm, Saturated'}, {'<LIGHTING>': 'Soft light, Low contrast, Side light'}, {'<LIGHTING_TYPE>': 'Daylight, Artificial light, Practical light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:01.png': [{'<COLOR>': 'Warm, Saturated'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Daylight, Artificial light, Practical light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:02.png': [{'<COLOR>': 'Warm, Saturated'}, {'<LIGHTING>': 'Soft light, Low contrast, Top light'}, {'<LIGHTING_TYPE>': 'Daylight, Artificial light'}, {'<COMPOSITION>': 'Right heavy'}]}
2
1
0
{"0": {"raw_counts": {"person": 1, "tie": 1, "total": 2}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 2}}, "1": {"raw_counts": {"person": 1, "tie": 1, "total": 2}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 2}}, "2": {"raw_counts": {"person": 1, "tie": 1, "total": 2}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 2}}}
0.069348
[-1 1 1 1 1 1 1 -1 -1 1 1 1 1 1 1 -1 -1 1 1 1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=4vJ6xB6ctaA
scene_001
1
135
186
00:00:05.631
00:00:07.758
5.630631
7.757758
2.127127
5.630631
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A medium shot captures a man wearing a light-colored suit, a burgundy shirt, and a patterned tie. His expression shifts as he looks slightly downward, then directly forward with a neutral gaze. The background includes a blurred object resembling a set of lights and a partition with vertical lines, adding depth to the composition.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
An individual in a suit is looking at something with curiosity in their eyes; the scene has a bright and airy feel.
2.127127
1,920
790
51
1920x790
short_scene_001_short_A_Thousand_Words_(2012)_-_Freestyle_Chanting_Scene_(1_10)_|_Movieclips
4.918449
{'frame_00:00.png': 4.8570475578308105, 'frame_00:01.png': 4.879134654998779, 'frame_00:02.png': 4.815728664398193, 'frame_00:03.png': 5.121883869171143}
short_A_Thousand_Words_(2012)_-_Freestyle_Chanting_Scene_(1_10)_|_Movieclips
../assets/split_scenes/short_A_Thousand_Words_(2012)_-_Freestyle_Chanting_Scene_(1_10)_|_Movieclips/short_scene_001.mp4
-0.143154
0.180015
{'frame_00:00.png': [{'<COLOR>': 'Warm, Saturated, Yellow'}, {'<LIGHTING>': 'Soft light, Side light, Edge light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:01.png': [{'<COLOR>': 'Warm, Saturated, Yellow'}, {'<LIGHTING>': 'Soft light, Side light, Edge light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:02.png': [{'<COLOR>': 'Warm, Saturated, Yellow'}, {'<LIGHTING>': 'Soft light, Side light, Edge light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:03.png': [{'<COLOR>': 'Warm, Saturated, Yellow'}, {'<LIGHTING>': 'Soft light, Side light, Edge light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Right heavy'}]}
5
4.5
0
{"0": {"raw_counts": {"person": 4, "couch": 1, "total": 5}, "bucketed_counts": {"humans": 4, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 5}}, "1": {"raw_counts": {"person": 5, "couch": 1, "total": 6}, "bucketed_counts": {"humans": 5, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 6}}, "2": {"raw_counts": {"person": 5, "total": 5}, "bucketed_counts": {"humans": 5, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 5}}, "3": {"raw_counts": {"person": 4, "total": 4}, "bucketed_counts": {"humans": 4, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 4}}}
0.066561
[-1 1 1 1 1 1 1 -1 1 1 1 1 1 -1 1 -1 1 1 1 1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=PurV6BzV3fI
scene_001
1
66
142
00:00:02.753
00:00:05.923
2.752753
5.922589
3.169837
2.752753
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A group of people are participating in a group meditation session outdoors. The main subject is a woman in the foreground with light brown hair, wearing a light-colored top and a necklace. She is in a meditative pose with her eyes closed. Behind her, slightly out of focus, are other participants, dressed in white, also with their eyes closed and in meditative poses. The background consists of a stone wall and some greenery. The overall atmosphere is calm and serene, enhanced by the soft, natural lighting. The camera focuses closely on the woman in the foreground, highlighting her peaceful expression.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
The scene depicts a group of individuals participating in a meditative practice outdoors, wearing simple attire, with a serene and focused atmosphere. The individuals appear to be aligned in a calm and peaceful state.
3.169837
1,920
774
76
1920x774
short_scene_001_short_A_Thousand_Words_(2012)_-_Nailed_It_Scene_(6_10)_|_Movieclips
4.87026
{'frame_00:00.png': 4.911582946777344, 'frame_00:01.png': 4.850337505340576, 'frame_00:02.png': 5.027199745178223, 'frame_00:03.png': 4.770320892333984, 'frame_00:04.png': 4.791860103607178}
short_A_Thousand_Words_(2012)_-_Nailed_It_Scene_(6_10)_|_Movieclips
../assets/split_scenes/short_A_Thousand_Words_(2012)_-_Nailed_It_Scene_(6_10)_|_Movieclips/short_scene_001.mp4
0.715358
1.071984
{'frame_00:00.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Hard light, High contrast, Backlight, Edge light'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:01.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Hard light, High contrast, Backlight, Edge light'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:02.png': [{'<COLOR>': 'Cool, Desaturated'}, {'<LIGHTING>': 'Hard light, High contrast, Backlight, Edge light'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:03.png': [{'<COLOR>': 'Cool, Desaturated, Yellow, White'}, {'<LIGHTING>': 'Hard light, High contrast, Backlight, Edge light'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:04.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Hard light, High contrast, Backlight, Edge light'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Balanced'}]}
7
4.2
0
{"0": {"raw_counts": {"person": 4, "laptop": 1, "tie": 1, "total": 6}, "bucketed_counts": {"humans": 4, "animals": 0, "vehicles": 0, "food": 0, "others": 2, "total": 6}}, "1": {"raw_counts": {"laptop": 1, "person": 5, "tie": 1, "total": 7}, "bucketed_counts": {"humans": 5, "animals": 0, "vehicles": 0, "food": 0, "others": 2, "total": 7}}, "2": {"raw_counts": {"person": 4, "tie": 1, "laptop": 1, "chair": 1, "total": 7}, "bucketed_counts": {"humans": 4, "animals": 0, "vehicles": 0, "food": 0, "others": 3, "total": 7}}, "3": {"raw_counts": {"person": 4, "laptop": 1, "tie": 1, "chair": 1, "total": 7}, "bucketed_counts": {"humans": 4, "animals": 0, "vehicles": 0, "food": 0, "others": 3, "total": 7}}, "4": {"raw_counts": {"person": 4, "tie": 1, "chair": 1, "laptop": 2, "total": 8}, "bucketed_counts": {"humans": 4, "animals": 0, "vehicles": 0, "food": 0, "others": 4, "total": 8}}}
0.034501
[-1 1 1 1 1 1 1 -1 -1 1 1 1 1 -1 1 -1 -1 1 1 -1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=XPsddyf2EcY
scene_001
1
61
165
00:00:02.544
00:00:06.882
2.544211
6.881882
4.337671
2.544211
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
The scene unfolds in a brightly lit, modern office boardroom, with a large window offering a view of a cityscape. A bald man in a dark suit and yellow tie stands at the head of a dark wooden conference table, gesturing as if giving a presentation. Seated around the table are four other individuals: an older man in a light-colored suit is seated on the left, listening intently and holding glasses in his hand, the backs of two women are visible, and a man in a dark shirt sits on the right, looking on with a contemplative expression. The speaker appears to be in the middle of a lively explanation, his mouth open and his hand raised, engaging the others in the discussion. The atmosphere is one of serious business, with papers and a telephone on the table suggesting active engagement in a project or negotiation.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A man in a suit stands and speaks at a conference table where others are seated, in a room with large windows showing a city. The scene is brightly lit, suggesting a formal or professional environment.
4.337671
1,920
756
104
1920x756
short_scene_001_short_A_Thousand_Words_(2012)_-_Show_Her_That_You_Love_Her_Scene_(9_10)_|_Movieclips
5.428637
{'frame_00:00.png': 5.44879150390625, 'frame_00:01.png': 5.4514079093933105, 'frame_00:02.png': 5.385711669921875}
short_A_Thousand_Words_(2012)_-_Show_Her_That_You_Love_Her_Scene_(9_10)_|_Movieclips
../assets/split_scenes/short_A_Thousand_Words_(2012)_-_Show_Her_That_You_Love_Her_Scene_(9_10)_|_Movieclips/short_scene_001.mp4
-0.14359
0.170453
{'frame_00:00.png': [{'<COLOR>': 'Warm, Desaturated'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:01.png': [{'<COLOR>': 'Warm, Desaturated'}, {'<LIGHTING>': 'Soft light, Side light, Edge light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:02.png': [{'<COLOR>': 'Warm, Desaturated'}, {'<LIGHTING>': 'Soft light, Side light, Edge light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}]}
1.333333
1.333333
0
{"0": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "1": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "2": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}}
0.020001
[-1 1 1 1 -1 1 1 -1 -1 1 1 1 1 -1 1 -1 -1 -1 1 -1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=Nuwu9VTP3Ak
scene_001
1
144
205
00:00:06.006
00:00:08.550
6.006006
8.550217
2.544211
6.006006
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
The scene is tightly framed around two individuals engaged in a serious conversation. One figure, with dark hair and a light-colored shirt, is shown in a close-up, his face etched with concern. His intense gaze is directed at the other person, whose back is to the camera. This figure is wearing a jacket, and the focus is clearly on the exchange between the two, their expressions suggesting a moment of gravity and emotional weight. The lighting is dim, adding to the somber mood of the encounter.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
Two individuals are engaged in intense conversation indoors, with one man facing the camera displaying a solemn expression. The scene is lit with a warm, muted light, focusing on their interaction.
2.544211
1,920
724
61
1920x724
short_scene_001_short_A_Thousand_Words_(2012)_-_Squirrels_in_My_Pants_Scene_(5_10)_|_Movieclips
4.882479
{'frame_00:00.png': 4.974362373352051, 'frame_00:01.png': 4.814692974090576, 'frame_00:02.png': 4.700078010559082, 'frame_00:03.png': 4.814118385314941, 'frame_00:04.png': 4.8506293296813965, 'frame_00:05.png': 4.933575630187988, 'frame_00:06.png': 5.089893817901611}
short_A_Thousand_Words_(2012)_-_Squirrels_in_My_Pants_Scene_(5_10)_|_Movieclips
../assets/split_scenes/short_A_Thousand_Words_(2012)_-_Squirrels_in_My_Pants_Scene_(5_10)_|_Movieclips/short_scene_001.mp4
0.220655
1.317405
{'frame_00:00.png': [{'<COLOR>': 'Cool, Cyan, Blue'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Daylight, Artificial light, Fluorescent'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:01.png': [{'<COLOR>': 'Cyan, Blue'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:02.png': [{'<COLOR>': 'Cyan, Blue'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:03.png': [{'<COLOR>': 'Cyan, Blue'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:04.png': [{'<COLOR>': 'Cyan, Blue'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:05.png': [{'<COLOR>': 'Cyan, Blue'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:06.png': [{'<COLOR>': 'Cool, Cyan, Blue'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Balanced'}]}
9.428571
6.714286
0
{"0": {"raw_counts": {"person": 7, "tie": 3, "total": 10}, "bucketed_counts": {"humans": 7, "animals": 0, "vehicles": 0, "food": 0, "others": 3, "total": 10}}, "1": {"raw_counts": {"person": 6, "tie": 5, "total": 11}, "bucketed_counts": {"humans": 6, "animals": 0, "vehicles": 0, "food": 0, "others": 5, "total": 11}}, "2": {"raw_counts": {"person": 7, "tie": 2, "total": 9}, "bucketed_counts": {"humans": 7, "animals": 0, "vehicles": 0, "food": 0, "others": 2, "total": 9}}, "3": {"raw_counts": {"person": 8, "tie": 2, "total": 10}, "bucketed_counts": {"humans": 8, "animals": 0, "vehicles": 0, "food": 0, "others": 2, "total": 10}}, "4": {"raw_counts": {"person": 7, "tie": 2, "total": 9}, "bucketed_counts": {"humans": 7, "animals": 0, "vehicles": 0, "food": 0, "others": 2, "total": 9}}, "5": {"raw_counts": {"person": 6, "tie": 2, "total": 8}, "bucketed_counts": {"humans": 6, "animals": 0, "vehicles": 0, "food": 0, "others": 2, "total": 8}}, "6": {"raw_counts": {"person": 6, "tie": 3, "total": 9}, "bucketed_counts": {"humans": 6, "animals": 0, "vehicles": 0, "food": 0, "others": 3, "total": 9}}}
0.05914
[-1 1 1 1 1 1 1 -1 -1 1 1 1 1 -1 1 -1 1 1 1 1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=B2KSjVAskxM
scene_001
1
168
321
00:00:07.007
00:00:13.388
7.007007
13.388388
6.381381
7.007007
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
Three men in suits engage in a discussion, one holding a blue folder, amidst a bustling office environment, with several other people in business attire walking around in the background. The man on the left gestures with his hand while speaking. The man in the center looks ahead with a slight smile, gripping the folder. The camera focuses on the three men, briefly revealing other participants in the conversation as they shift their positions. The men appear confident and engaged as the discussion progresses, smiles exchanged, and an overall sense of optimism fills the room.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
Three men in business suits are talking to someone, while other people walk around in the background. The men appear to be smiling and in an upbeat mood, in an office setting.
6.381381
1,920
774
153
1920x774
short_scene_001_short_A_Thousand_Words_(2012)_-_Talk_Dirty_to_Me_Scene_(7_10)_|_Movieclips
4.550344
{'frame_00:00.png': 4.570397853851318, 'frame_00:01.png': 4.506475925445557, 'frame_00:02.png': 4.555637359619141, 'frame_00:03.png': 4.5688652992248535}
short_A_Thousand_Words_(2012)_-_Talk_Dirty_to_Me_Scene_(7_10)_|_Movieclips
../assets/split_scenes/short_A_Thousand_Words_(2012)_-_Talk_Dirty_to_Me_Scene_(7_10)_|_Movieclips/short_scene_001.mp4
0.019498
0.259169
{'frame_00:00.png': [{'<COLOR>': 'Warm, Saturated, Yellow'}, {'<LIGHTING>': 'Soft light, Backlight'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:01.png': [{'<COLOR>': 'Warm, Saturated, Yellow'}, {'<LIGHTING>': 'Soft light, Backlight'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:02.png': [{'<COLOR>': 'Warm, Saturated, Yellow'}, {'<LIGHTING>': 'Soft light, Backlight'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:03.png': [{'<COLOR>': 'Warm, Saturated, Yellow'}, {'<LIGHTING>': 'Soft light, Backlight'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light'}, {'<COMPOSITION>': 'Balanced'}]}
3
2
0
{"0": {"raw_counts": {"person": 2, "bottle": 1, "total": 3}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 3}}, "1": {"raw_counts": {"person": 2, "bottle": 1, "total": 3}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 3}}, "2": {"raw_counts": {"person": 2, "bottle": 1, "total": 3}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 3}}, "3": {"raw_counts": {"person": 2, "vase": 1, "total": 3}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 3}}}
0.001771
[ 1 1 1 1 1 1 1 1 1 1 -1 -1 1 -1 1 -1 -1 -1 -1 -1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=B_cE7WEW5gM
scene_001
1
345
441
00:00:14.389
00:00:18.393
14.389389
18.393393
4.004004
14.389389
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
Through a dimly lit room with flickering candles, a man enters a room, wearing a button-up shirt and tie, while in the foreground a woman stands with her back to the camera, wearing a black two-piece outfit with studded gloves. The man removes his tie, his face displaying a mixture of surprise and intrigue as he fully enters the room, the scene illuminated by the soft glow of the candlelight. The woman remains still, enhancing the intimate atmosphere of the scene, as the scene is lit with dramatic lighting.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A man in formal wear enters a room to find a woman in suggestive attire posing in the light of many candles. He then removes his tie in what appears to be a heated moment.
4.004004
1,920
756
96
1920x756
short_scene_001_short_A_Thousand_Words_(2012)_-_Triple_Shot,_No_Assassinations_Scene_(3_10)_|_Movieclips
5.592396
{'frame_00:00.png': 5.5867204666137695, 'frame_00:01.png': 5.757889747619629, 'frame_00:02.png': 5.7227020263671875, 'frame_00:03.png': 5.302270412445068}
short_A_Thousand_Words_(2012)_-_Triple_Shot,_No_Assassinations_Scene_(3_10)_|_Movieclips
../assets/split_scenes/short_A_Thousand_Words_(2012)_-_Triple_Shot,_No_Assassinations_Scene_(3_10)_|_Movieclips/short_scene_001.mp4
-0.056946
0.146267
{'frame_00:00.png': [{'<COLOR>': 'Desaturated, Cyan, Blue'}, {'<LIGHTING>': 'Soft light, Backlight, Edge light'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:01.png': [{'<COLOR>': 'Cyan, Blue'}, {'<LIGHTING>': 'Soft light, Backlight, Edge light'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:02.png': [{'<COLOR>': 'Cyan, Blue'}, {'<LIGHTING>': 'Soft light, Backlight'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:03.png': [{'<COLOR>': 'Cyan, Blue'}, {'<LIGHTING>': 'Soft light, Backlight'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Right heavy'}]}
3
1.75
0
{"0": {"raw_counts": {"person": 2, "tie": 1, "car": 1, "total": 4}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 1, "food": 0, "others": 1, "total": 4}}, "1": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "2": {"raw_counts": {"person": 3, "tie": 1, "car": 1, "total": 5}, "bucketed_counts": {"humans": 3, "animals": 0, "vehicles": 1, "food": 0, "others": 1, "total": 5}}, "3": {"raw_counts": {"person": 1, "tie": 1, "total": 2}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 2}}}
0.061829
[ 1 1 1 1 1 1 1 -1 -1 1 1 1 1 -1 1 -1 1 1 1 1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=8rX7fkDLEx0
scene_001
1
92
173
00:00:03.837
00:00:07.216
3.837171
7.215549
3.378378
3.837171
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A man in a gray suit and blue tie stands inside an office. He faces the camera and looks straight ahead. He then puts up three fingers next to his face with a smile. He then turns to face the camera again before making a gesture with his hands, holding them out in front of him as if holding a small object. A glass door is behind him, where a dark figure can be seen walking by in the background. Other people can be seen, out of focus, in the background.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A man in a suit is looking directly at the camera, then playfully makes gestures with his hands and face. The scene appears to be taking place inside an office or public building.
3.378378
1,920
810
81
1920x810
short_scene_001_short_A_Thousand_Words_(2012)_-_You_Found_My_Furry_Tapes_Scene_(4_10)_|_Movieclips
4.826297
{'frame_00:00.png': 4.422775745391846, 'frame_00:01.png': 4.434165954589844, 'frame_00:02.png': 5.329416275024414, 'frame_00:03.png': 4.959760665893555, 'frame_00:04.png': 4.985365390777588}
short_A_Thousand_Words_(2012)_-_You_Found_My_Furry_Tapes_Scene_(4_10)_|_Movieclips
../assets/split_scenes/short_A_Thousand_Words_(2012)_-_You_Found_My_Furry_Tapes_Scene_(4_10)_|_Movieclips/short_scene_001.mp4
0.526818
1.558542
{'frame_00:00.png': [{'<COLOR>': 'Warm, Desaturated'}, {'<LIGHTING>': 'Soft light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:01.png': [{'<COLOR>': 'Warm, Desaturated'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light'}, {'<COMPOSITION>': 'Center'}], 'frame_00:02.png': [{'<COLOR>': 'Warm, Desaturated'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:03.png': [{'<COLOR>': 'Warm, Desaturated, Blue'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:04.png': [{'<COLOR>': 'Mixed, Saturated, Blue'}, {'<LIGHTING>': 'Soft light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light'}, {'<COMPOSITION>': 'Balanced'}]}
3.4
1.4
0
{"0": {"raw_counts": {"bottle": 1, "total": 1}, "bucketed_counts": {"humans": 0, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 1}}, "1": {"raw_counts": {"person": 1, "bottle": 1, "chair": 1, "total": 3}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 2, "total": 3}}, "2": {"raw_counts": {"person": 2, "tie": 1, "bottle": 1, "tv": 1, "total": 5}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 3, "total": 5}}, "3": {"raw_counts": {"person": 2, "tie": 1, "total": 3}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 3}}, "4": {"raw_counts": {"person": 2, "remote": 1, "tie": 1, "tv": 1, "total": 5}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 3, "total": 5}}}
0.032197
[ 1 1 1 1 1 1 1 -1 -1 1 -1 -1 1 -1 1 1 1 1 1 -1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=6AJH6b91rF8
scene_001
1
161
281
00:00:06.715
00:00:11.720
6.715048
11.720053
5.005005
6.715048
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
An office space is revealed, featuring a dark shelving unit filled with various objects, including a photo, trophy, and decorative box, alongside a television set. A framed certificate adorns the adjacent wall. A man strides into view, carrying a briefcase. He begins to remove his suit jacket, engaging in conversation with another man dressed in a dark suit, who enters the frame to the right. The scene is lit by a combination of natural and artificial light, creating a professional and somewhat serious atmosphere.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A man enters an office and takes off his suit jacket while another man speaks to him. The office setting has modern decor with brown wood paneled walls and shelving with various decorative objects.
5.005005
1,920
774
120
1920x774
short_scene_001_short_A_View_To_a_Kill_Movie_CLIP_-_Buying_or_Selling_(1985)_HD
4.714863
{'frame_00:00.png': 4.970160007476807, 'frame_00:01.png': 4.490785121917725, 'frame_00:02.png': 4.783849716186523, 'frame_00:03.png': 4.861402988433838, 'frame_00:04.png': 4.456238269805908, 'frame_00:05.png': 4.72674036026001}
short_A_View_To_a_Kill_Movie_CLIP_-_Buying_or_Selling_(1985)_HD
../assets/split_scenes/short_A_View_To_a_Kill_Movie_CLIP_-_Buying_or_Selling_(1985)_HD/short_scene_001.mp4
-0.361885
4.232993
{'frame_00:00.png': [{'<COLOR>': 'Green, Cyan'}, {'<LIGHTING>': 'Hard light, Top light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:01.png': [{'<COLOR>': 'Green'}, {'<LIGHTING>': 'Hard light, Top light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:02.png': [{'<COLOR>': 'Green'}, {'<LIGHTING>': 'Hard light, Top light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:03.png': [{'<COLOR>': 'Green'}, {'<LIGHTING>': 'Hard light, Top light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:04.png': [{'<COLOR>': 'Green'}, {'<LIGHTING>': 'Hard light, Top light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:05.png': [{'<COLOR>': 'Red'}, {'<LIGHTING>': 'Hard light, Top light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Balanced'}]}
12.666667
11
0
{"0": {"raw_counts": {"person": 12, "tie": 1, "total": 13}, "bucketed_counts": {"humans": 12, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 13}}, "1": {"raw_counts": {"person": 10, "tie": 1, "total": 11}, "bucketed_counts": {"humans": 10, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 11}}, "2": {"raw_counts": {"person": 11, "tie": 1, "total": 12}, "bucketed_counts": {"humans": 11, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 12}}, "3": {"raw_counts": {"person": 14, "total": 14}, "bucketed_counts": {"humans": 14, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 14}}, "4": {"raw_counts": {"person": 9, "tie": 3, "cell phone": 1, "total": 13}, "bucketed_counts": {"humans": 9, "animals": 0, "vehicles": 0, "food": 0, "others": 4, "total": 13}}, "5": {"raw_counts": {"person": 10, "tie": 3, "total": 13}, "bucketed_counts": {"humans": 10, "animals": 0, "vehicles": 0, "food": 0, "others": 3, "total": 13}}}
0.059055
[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, -1, 1, -1, 1, 1, 1, -1, -1, 1, 1, -1, -1, 1]
https://youtube.com/watch?v=eLOVOy0CCv8
scene_001
1
410
542
00:00:17.100
00:00:22.606
17.100434
22.605939
5.505506
17.100434
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
The scene unfolds at a lavish outdoor party, densely populated with formally dressed guests. A fair-haired man wearing sunglasses and a tuxedo engages in conversation with another man in a similar suit, standing face-to-face amidst the crowd. Between them, a bald man smiles faintly, holding a champagne flute. The shot gradually focuses to include a woman with striking features in a red wrap, who appears to join their discussion, her expression serious. The background is adorned with classical sculptures and ornate architecture, underscoring the opulence of the event.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
The scene captures a formal outdoor gathering with people dressed in elegant attire, some engaged in conversation, set against a backdrop of a grand estate. A sense of sophistication and intrigue pervades the atmosphere.
5.505506
1,280
522
132
1280x522
short_scene_001_short_A_View_To_a_Kill_Movie_CLIP_-_Skyship_(1985)_HD
5.233846
{'frame_00:00.png': 5.299782752990723, 'frame_00:01.png': 5.216463565826416, 'frame_00:02.png': 5.117559432983398, 'frame_00:03.png': 5.301576614379883}
short_A_View_To_a_Kill_Movie_CLIP_-_Skyship_(1985)_HD
../assets/split_scenes/short_A_View_To_a_Kill_Movie_CLIP_-_Skyship_(1985)_HD/short_scene_001.mp4
0.410488
0.113197
{'frame_00:00.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Soft light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:01.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Soft light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:02.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Soft light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:03.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Soft light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}]}
3
2
0
{"0": {"raw_counts": {"person": 2, "tie": 1, "total": 3}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 3}}, "1": {"raw_counts": {"person": 2, "tie": 1, "total": 3}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 3}}, "2": {"raw_counts": {"person": 2, "tie": 1, "total": 3}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 3}}, "3": {"raw_counts": {"person": 2, "tie": 1, "total": 3}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 3}}}
0.020001
[-1, 1, 1, 1, -1, 1, 1, -1, -1, 1, 1, 1, 1, -1, 1, -1, -1, -1, 1, -1, -1, 1, 1, -1, -1, 1]
https://youtube.com/watch?v=VEGSjiQq998
scene_001
1
214
292
00:00:08.926
00:00:12.179
8.925592
12.178846
3.253253
8.925592
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A person with light hair is seen wearing a gray suit and red tie, looking off to the side and then directly at the camera. Behind them, another individual stands with their arms crossed, wearing a black outfit. They both appear to be in a modern-looking interior, characterized by sleek lines and a large, stylized symbol on the wall. The overall atmosphere suggests a scene of intrigue or confrontation, with a blend of formality and tension.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A fair-skinned man in a suit stands looking off-camera with a dark-skinned woman standing in the background. The scene is set in what appears to be an office building with a high-tech logo in the background.
3.253253
1,280
540
78
1280x540
short_scene_001_short_A_Woman_in_a_Man's_Profession_-_Shakespeare_in_Love_(6_8)_Movie_CLIP_(1998)_HD
4.845563
{'frame_00:00.png': 5.028652667999268, 'frame_00:01.png': 4.7965264320373535, 'frame_00:02.png': 4.559311866760254, 'frame_00:03.png': 4.997760772705078}
short_A_Woman_in_a_Man's_Profession_-_Shakespeare_in_Love_(6_8)_Movie_CLIP_(1998)_HD
../assets/split_scenes/short_A_Woman_in_a_Man's_Profession_-_Shakespeare_in_Love_(6_8)_Movie_CLIP_(1998)_HD/short_scene_001.mp4
0.342896
1.903347
{'frame_00:00.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Hard light, Side light'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:01.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Hard light, Side light'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:02.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Hard light, Side light'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:03.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Hard light, Side light'}, {'<LIGHTING_TYPE>': 'Daylight, Overcast'}, {'<COMPOSITION>': 'Right heavy'}]}
2
2
0
{"0": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "1": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "2": {"raw_counts": {"person": 3, "total": 3}, "bucketed_counts": {"humans": 3, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 3}}, "3": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}}
-0.096997
[-1, -1, 1, -1, -1, -1, 1, -1, -1, -1, -1, -1, -1, -1, -1, 1, 1, -1, -1, -1, -1, -1, -1, -1, -1, 1]
https://youtube.com/watch?v=o_KXbKa2crI
scene_001
1
98
172
00:00:04.087
00:00:07.174
4.087421
7.173841
3.08642
4.087421
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A man in a period costume is clapping enthusiastically and appears to be shouting or cheering, with his hand covering his mouth in excitement. The background shows a tiered seating area filled with people who are also applauding, suggesting a lively and engaging performance or event taking place. The setting suggests a theater or public gathering spot from a historical period, as evidenced by the architecture and the attire of the people. The overall atmosphere conveys a sense of celebration and enjoyment among the audience.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A group of people at a theatre, sitting in the boxes and standing on the main floor, clap enthusiastically while wearing Elizabethan era clothing. A few people on the side appear particularly excited with their hands raised.
3.08642
1,280
522
74
1280x522
short_scene_001_short_Abominable_(2019)_-_Bridge_Battle_Scene_(9_10)_|_Movieclips
4.803887
{'frame_00:00.png': 4.800931930541992, 'frame_00:01.png': 4.738840103149414, 'frame_00:02.png': 4.871889591217041}
short_Abominable_(2019)_-_Bridge_Battle_Scene_(9_10)_|_Movieclips
../assets/split_scenes/short_Abominable_(2019)_-_Bridge_Battle_Scene_(9_10)_|_Movieclips/short_scene_001.mp4
-0.430786
0.3488
{'frame_00:00.png': [{'<COLOR>': 'Cool, Desaturated, Blue'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:01.png': [{'<COLOR>': 'Cool, Desaturated, Blue'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Moonlight, Artificial light'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:02.png': [{'<COLOR>': 'Cool, Desaturated, Blue'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Balanced'}]}
6
5.666667
0
{"0": {"raw_counts": {"person": 2, "chair": 1, "total": 3}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 3}}, "1": {"raw_counts": {"person": 9, "total": 9}, "bucketed_counts": {"humans": 9, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 9}}, "2": {"raw_counts": {"person": 6, "total": 6}, "bucketed_counts": {"humans": 6, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 6}}}
-0.059698
[-1 -1 1 -1 -1 -1 1 -1 -1 -1 -1 -1 1 -1 1 -1 1 -1 -1 -1 -1 -1 -1 -1 1 1]
https://youtube.com/watch?v=GU1zIn2BRN0
scene_001
1
123
175
00:00:05.130
00:00:07.299
5.13013
7.298966
2.168836
5.13013
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A dimly lit, snow-covered environment features a gated entrance, beyond which a vehicle with a cage-like structure stands prominently. A figure in dark clothing and sunglasses exits the vehicle, holding what appears to be a firearm. Subsequently, two additional figures emerge—one in a lab coat and another in dark attire. They are positioned between the imposing vehicle and a group of individuals wearing helmets in the foreground. The scene is heavily stylized, with a focus on dramatic lighting and shadowy figures, enhancing the sense of tension and mystery.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A group of uniformed figures stands guard as several individuals, including one in a white coat, exit a vehicle under a cold, blue-toned light. The scene suggests a tense or controlled environment.
2.168836
1,920
810
52
1920x810
short_scene_001_short_Abominable_(2019)_-_Drone_Attack_Scene_(4_10)_|_Movieclips
4.990292
{'frame_00:00.png': 5.174649238586426, 'frame_00:01.png': 4.622650146484375, 'frame_00:02.png': 5.173577308654785}
short_Abominable_(2019)_-_Drone_Attack_Scene_(4_10)_|_Movieclips
../assets/split_scenes/short_Abominable_(2019)_-_Drone_Attack_Scene_(4_10)_|_Movieclips/short_scene_001.mp4
0.097647
0.742538
{'frame_00:00.png': [{'<COLOR>': 'Cool, Green, Blue'}, {'<LIGHTING>': 'Hard light, High contrast, Backlight'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:01.png': [{'<COLOR>': 'Cool, Green, Blue'}, {'<LIGHTING>': 'Hard light, High contrast, Backlight'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:02.png': [{'<COLOR>': 'Mixed, Saturated, Green, Blue'}, {'<LIGHTING>': 'Hard light, High contrast, Side light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Right heavy'}]}
0.333333
0.333333
0
{"0": {"raw_counts": {"total": 0}, "bucketed_counts": {"humans": 0, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 0}}, "1": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "2": {"raw_counts": {"total": 0}, "bucketed_counts": {"humans": 0, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 0}}}
-0.01328
[-1 -1 1 1 -1 1 1 1 1 1 -1 -1 1 -1 1 1 1 -1 -1 -1 -1 -1 -1 -1 -1 1]
https://youtube.com/watch?v=rRUVmTXpPyg
scene_001
1
87
144
00:00:03.629
00:00:06.006
3.628629
6.006006
2.377377
3.628629
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
Dandelions sway gently in a sunlit meadow beside a dirt path, where tall grasses and bushy foliage thrive. A child approaches, their hands reaching out to pluck a dandelion. Close-up, the child examines the delicate seed head with a gentle expression, the light catching in their eyes as they focus intently on the flower.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A person is seen picking dandelions from the ground in a sunlit grassy area. The peaceful setting has a dreamy, animated quality.
2.377377
1,920
876
57
1920x876
short_scene_001_short_Abominable_(2019)_-_Magic_Boat_Chase_Scene_(7_10)_|_Movieclips
4.659883
{'frame_00:00.png': 4.675813674926758, 'frame_00:01.png': 4.378181457519531, 'frame_00:02.png': 4.408347129821777, 'frame_00:03.png': 5.177190780639648}
short_Abominable_(2019)_-_Magic_Boat_Chase_Scene_(7_10)_|_Movieclips
../assets/split_scenes/short_Abominable_(2019)_-_Magic_Boat_Chase_Scene_(7_10)_|_Movieclips/short_scene_001.mp4
-0.022438
1.088159
{'frame_00:00.png': [{'<COLOR>': 'Cool, Saturated, Green, Blue, White'}, {'<LIGHTING>': 'Hard light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:01.png': [{'<COLOR>': 'Cool, Green, Blue, White'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Center'}], 'frame_00:02.png': [{'<COLOR>': 'Cool, Desaturated, White'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Daylight, Overcast'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:03.png': [{'<COLOR>': 'Cool, Green, Blue'}, {'<LIGHTING>': 'Hard light, Low contrast'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Balanced'}]}
0.25
0
0
{"0": {"raw_counts": {"train": 1, "total": 1}, "bucketed_counts": {"humans": 0, "animals": 0, "vehicles": 1, "food": 0, "others": 0, "total": 1}}, "1": {"raw_counts": {"total": 0}, "bucketed_counts": {"humans": 0, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 0}}, "2": {"raw_counts": {"total": 0}, "bucketed_counts": {"humans": 0, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 0}}, "3": {"raw_counts": {"total": 0}, "bucketed_counts": {"humans": 0, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 0}}}
-0.07441
[-1 -1 -1 1 -1 -1 1 1 1 1 -1 -1 -1 -1 -1 1 1 -1 -1 -1 -1 -1 -1 -1 -1 1]
https://youtube.com/watch?v=LCj92toBBBE
scene_001
1
266
340
00:00:11.094
00:00:14.181
11.094428
14.180848
3.08642
11.094428
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A town built along a river is presented in a wide shot, featuring small buildings with brown roofs lined up next to the water, and in the background, green hills and a blue sky with white clouds are visible; a small boat with a person aboard floats gently on the water. Abruptly, a massive splash of water erupts, filling the frame and obscuring the serene landscape. The wave crashes forward, engulfing the scene with water droplets, and then, as the water recedes, revealing a figure standing atop a structure near the river amidst the turbulent water, under a sky now pouring rain.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A village with water canals is disturbed by a large splash, creating waves that inundate the area. The scene depicts rising water levels and turbulent currents, as the village's buildings and surrounding mountains provide a backdrop to the aquatic chaos.
3.08642
1,920
1,032
74
1920x1032
short_scene_001_short_Abominable_(2019)_-_Meeting_Everest_Scene_(1_10)_|_Movieclips
4.839776
{'frame_00:00.png': 4.814322471618652, 'frame_00:01.png': 4.8786845207214355, 'frame_00:02.png': 4.826322078704834}
short_Abominable_(2019)_-_Meeting_Everest_Scene_(1_10)_|_Movieclips
../assets/split_scenes/short_Abominable_(2019)_-_Meeting_Everest_Scene_(1_10)_|_Movieclips/short_scene_001.mp4
-0.072729
0.022969
{'frame_00:00.png': [{'<COLOR>': 'Mixed, Desaturated, Blue'}, {'<LIGHTING>': 'Hard light, High contrast, Underlight'}, {'<LIGHTING_TYPE>': 'Moonlight, Artificial light, Practical light, Mixed light'}, {'<COMPOSITION>': 'Center'}], 'frame_00:01.png': [{'<COLOR>': 'Mixed, Desaturated, Cyan, Blue'}, {'<LIGHTING>': 'Hard light, High contrast, Underlight'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light'}, {'<COMPOSITION>': 'Center'}], 'frame_00:02.png': [{'<COLOR>': 'Mixed, Saturated, Cyan, Blue'}, {'<LIGHTING>': 'Hard light, High contrast, Underlight'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light, Mixed light'}, {'<COMPOSITION>': 'Center'}]}
0
0
0
{"0": {"raw_counts": {"total": 0}, "bucketed_counts": {"humans": 0, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 0}}, "1": {"raw_counts": {"total": 0}, "bucketed_counts": {"humans": 0, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 0}}, "2": {"raw_counts": {"total": 0}, "bucketed_counts": {"humans": 0, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 0}}}
-0.051106
[-1 -1 1 -1 -1 -1 1 -1 -1 -1 -1 -1 1 -1 1 1 1 -1 -1 -1 -1 -1 -1 -1 1 1]
https://youtube.com/watch?v=A-Ckh0slywY
scene_001
1
74
125
00:00:03.086
00:00:05.214
3.08642
5.213547
2.127127
3.08642
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A figure stands atop a building at night, with dramatic lighting casting long shadows across the scene. Nearby, a makeshift structure is adorned with colorful lights. A helicopter approaches, its spotlight illuminating the rooftop, and the words "Burnish Industries" are visible on its side. The helicopter's arrival suggests a confrontation or intervention within the urban setting.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
An individual stands on a rooftop as a helicopter approaches a structure decorated with lights; the scene is set at night. The helicopter shines a spotlight as it descends.
2.127127
1,920
810
51
1920x810
short_scene_001_short_Abominable_(2019)_-_The_Magic_Violin_Scene_(8_10)_|_Movieclips
5.68883
{'frame_00:00.png': 5.530084133148193, 'frame_00:01.png': 5.868206977844238, 'frame_00:02.png': 5.772580623626709, 'frame_00:03.png': 5.584449291229248}
short_Abominable_(2019)_-_The_Magic_Violin_Scene_(8_10)_|_Movieclips
../assets/split_scenes/short_Abominable_(2019)_-_The_Magic_Violin_Scene_(8_10)_|_Movieclips/short_scene_001.mp4
5.087858
0.028614
{'frame_00:00.png': [{'<COLOR>': 'Cool, Saturated, Green, Blue'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Daylight, Overcast'}, {'<COMPOSITION>': 'Center'}], 'frame_00:01.png': [{'<COLOR>': 'Cool, Saturated, Green, Blue'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Daylight, Overcast'}, {'<COMPOSITION>': 'Center'}], 'frame_00:02.png': [{'<COLOR>': 'Cool, Saturated, Green, Blue'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Daylight, Overcast'}, {'<COMPOSITION>': 'Center'}], 'frame_00:03.png': [{'<COLOR>': 'Cool, Saturated, Green, Blue'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Daylight, Overcast'}, {'<COMPOSITION>': 'Center'}]}
2.25
1.25
0.75
{"0": {"raw_counts": {"total": 0}, "bucketed_counts": {"humans": 0, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 0}}, "1": {"raw_counts": {"person": 1, "elephant": 1, "total": 2}, "bucketed_counts": {"humans": 1, "animals": 1, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "2": {"raw_counts": {"person": 1, "fire hydrant": 1, "total": 2}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 2}}, "3": {"raw_counts": {"person": 3, "hot dog": 1, "elephant": 1, "total": 5}, "bucketed_counts": {"humans": 3, "animals": 2, "vehicles": 0, "food": 0, "others": 0, "total": 5}}}
-0.042397
[-1 1 1 -1 -1 1 1 -1 -1 -1 -1 -1 1 -1 1 -1 1 -1 -1 -1 -1 -1 1 -1 -1 1]
https://youtube.com/watch?v=chvjkV0jRh8
scene_001
1
101
194
00:00:04.213
00:00:08.091
4.212546
8.091425
3.878879
4.212546
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
The animated scene shows a lush, fantastical landscape where massive stone hands form part of the terrain, covered in moss and greenery. A small figure is carefully climbing one of the fingers, while two other figures are standing on another finger in the distance, appearing to observe the climber. The lighting is soft and diffused, creating a dreamy and slightly mysterious atmosphere. The overall composition suggests a journey or exploration within a world of colossal proportions. A white, furry creature emerges from below one of the stone hands.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
The scene depicts an animated figure standing on a large, stone-like finger, while other figures are visible on another giant finger in the background, all set against a misty, mountainous backdrop. A large, white, furry creature is shown approaching the figures.
3.878879
1,920
1,032
93
1920x1032
short_scene_001_short_Abominable_(2019)_-_Yak_Attack!_Scene_(6_10)_|_Movieclips
5.104567
{'frame_00:00.png': 4.261094570159912, 'frame_00:01.png': 5.518312454223633, 'frame_00:02.png': 5.534294128417969}
short_Abominable_(2019)_-_Yak_Attack!_Scene_(6_10)_|_Movieclips
../assets/split_scenes/short_Abominable_(2019)_-_Yak_Attack!_Scene_(6_10)_|_Movieclips/short_scene_001.mp4
0.321191
0.367391
{'frame_00:00.png': [{'<COLOR>': 'Mixed, Desaturated, Blue'}, {'<LIGHTING>': 'Soft light'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Center'}], 'frame_00:01.png': [{'<COLOR>': 'Mixed, Saturated, Blue'}, {'<LIGHTING>': 'Soft light'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Center'}], 'frame_00:02.png': [{'<COLOR>': 'Mixed, Saturated, Cyan, Blue'}, {'<LIGHTING>': 'Soft light'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Center'}]}
0.333333
0
0.333333
{"0": {"raw_counts": {"total": 0}, "bucketed_counts": {"humans": 0, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 0}}, "1": {"raw_counts": {"teddy bear": 1, "total": 1}, "bucketed_counts": {"humans": 0, "animals": 1, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "2": {"raw_counts": {"total": 0}, "bucketed_counts": {"humans": 0, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 0}}}
-0.027352
[-1 1 1 1 -1 1 1 -1 -1 1 -1 -1 1 -1 1 -1 1 -1 -1 -1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=eNCK08cInIE
scene_001
1
147
213
00:00:06.131
00:00:08.884
6.131131
8.883884
2.752753
6.131131
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
Peeking through a slightly open red wooden door, a person's hands are visible, hinting at someone hidden behind it, with a stone wall and a window to the side, suggesting an old building exterior, soon a young girl with short black hair and a red shirt cautiously peeks out, her expression a mix of curiosity and concern, and moments later, she's joined by a young boy and a large-eyed creature, creating a sense of a group, perhaps a family or team, cautiously observing something outside, with a focus on their shared curiosity and slight apprehension, captured in a medium shot that emphasizes their expressions and the texture of the weathered door.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A mysterious scene unfolds as two children cautiously peek through a cracked wooden door, with a curious creature joining them, all set against a textured wall and window backdrop. The shot evokes a sense of secrecy and anticipation.
2.752753
1,920
864
66
1920x864
short_scene_001_short_About_Time_(2013)_-_A_Friend_Who_Is_a_Girl_Scene_(4_10)_|_Movieclips
5.378282
{'frame_00:00.png': 5.224994659423828, 'frame_00:01.png': 5.440554618835449, 'frame_00:02.png': 5.531447410583496, 'frame_00:03.png': 5.316131591796875}
short_About_Time_(2013)_-_A_Friend_Who_Is_a_Girl_Scene_(4_10)_|_Movieclips
../assets/split_scenes/short_About_Time_(2013)_-_A_Friend_Who_Is_a_Girl_Scene_(4_10)_|_Movieclips/short_scene_001.mp4
0.068871
0.506386
{'frame_00:00.png': [{'<COLOR>': 'Warm, Desaturated'}, {'<LIGHTING>': 'Soft light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Center'}], 'frame_00:01.png': [{'<COLOR>': 'Warm, Desaturated'}, {'<LIGHTING>': 'Soft light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Center'}], 'frame_00:02.png': [{'<COLOR>': 'Warm, Desaturated'}, {'<LIGHTING>': 'Soft light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Center'}], 'frame_00:03.png': [{'<COLOR>': 'Warm, Desaturated'}, {'<LIGHTING>': 'Soft light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Center'}]}
6.5
5.75
0
{"0": {"raw_counts": {"person": 5, "tie": 2, "total": 7}, "bucketed_counts": {"humans": 5, "animals": 0, "vehicles": 0, "food": 0, "others": 2, "total": 7}}, "1": {"raw_counts": {"person": 6, "total": 6}, "bucketed_counts": {"humans": 6, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 6}}, "2": {"raw_counts": {"person": 7, "tie": 1, "total": 8}, "bucketed_counts": {"humans": 7, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 8}}, "3": {"raw_counts": {"person": 5, "total": 5}, "bucketed_counts": {"humans": 5, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 5}}}
-0.05594
[-1 1 1 1 -1 -1 1 -1 -1 -1 -1 -1 1 -1 1 -1 1 -1 -1 -1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=_oygxA1aaRU
scene_001
1
175
256
00:00:07.299
00:00:10.677
7.298966
10.677344
3.378378
7.298966
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
Amidst a bustling crowd, a man with ginger hair and a blue sweater looks concerned as he turns to his right, while another man with curly hair and a suit peers around his shoulder with a worried expression; both are framed by a diverse group of people, including those in suits and casual attire, against a backdrop of a dark wall. The man with ginger hair is mid-motion, having just removed or about to wear a jacket over his shoulder. The surrounding crowd, a mix of ages and ethnicities, adds to the sense of urgency or concern suggested by the two men's expressions.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
In a crowded room with dim lighting, two men stand out as one gestures anxiously and pulls the jacket off his companion, both looking surprised. The scene is filled with other indistinct figures milling around, creating a sense of urgency and confusion.
3.378378
1,920
810
81
1920x810
short_scene_001_short_About_Time_(2013)_-_Best_Man_Speeches_Scene_(7_10)_|_Movieclips
5.41422
{'frame_00:00.png': 5.5111541748046875, 'frame_00:01.png': 5.411203384399414, 'frame_00:02.png': 5.429483413696289, 'frame_00:03.png': 5.3050384521484375}
short_About_Time_(2013)_-_Best_Man_Speeches_Scene_(7_10)_|_Movieclips
../assets/split_scenes/short_About_Time_(2013)_-_Best_Man_Speeches_Scene_(7_10)_|_Movieclips/short_scene_001.mp4
-0.455978
0.161339
{'frame_00:00.png': [{'<COLOR>': 'Warm, Saturated'}, {'<LIGHTING>': 'Soft light, Low contrast, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:01.png': [{'<COLOR>': 'Warm, Saturated'}, {'<LIGHTING>': 'Soft light, Low contrast, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:02.png': [{'<COLOR>': 'Warm, Saturated'}, {'<LIGHTING>': 'Soft light, Low contrast, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:03.png': [{'<COLOR>': 'Warm, Saturated'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Balanced'}]}
9.25
7.25
0
{"0": {"raw_counts": {"person": 8, "tie": 2, "total": 10}, "bucketed_counts": {"humans": 8, "animals": 0, "vehicles": 0, "food": 0, "others": 2, "total": 10}}, "1": {"raw_counts": {"person": 6, "tie": 2, "total": 8}, "bucketed_counts": {"humans": 6, "animals": 0, "vehicles": 0, "food": 0, "others": 2, "total": 8}}, "2": {"raw_counts": {"person": 7, "tie": 2, "total": 9}, "bucketed_counts": {"humans": 7, "animals": 0, "vehicles": 0, "food": 0, "others": 2, "total": 9}}, "3": {"raw_counts": {"person": 8, "tie": 2, "total": 10}, "bucketed_counts": {"humans": 8, "animals": 0, "vehicles": 0, "food": 0, "others": 2, "total": 10}}}
0.053189
[-1 1 1 1 1 1 1 -1 -1 1 1 1 1 -1 1 -1 -1 -1 1 1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=x_Dephm0Z9I
scene_001
1
84
159
00:00:03.504
00:00:06.632
3.503504
6.631632
3.128128
3.503504
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
At a wedding reception, a couple sits in the foreground, the woman in a red dress and the man in a blue suit, both smiling as they look towards the wedding ceremony. Behind them, the wedding guests are sitting and listening to the wedding ceremony. The guest on the left wears a pink suit and a black tie, while the woman behind him wears a tan dress. In the background, more guests are visible, all focused on the ceremony. The light is bright and natural, creating a warm and happy atmosphere.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A group of people sits in a formal setting, possibly at a wedding, wearing suits and dresses; the mood appears joyful. The shot focuses on two individuals in the foreground, capturing their expressions and attire in detail, while others fill the background.
3.128128
1,920
810
75
1920x810
short_scene_001_short_About_Time_(2013)_-_First_Night_Together_Scene_(3_10)_|_Movieclips
4.446966
{'frame_00:00.png': 4.308161735534668, 'frame_00:01.png': 4.48919153213501, 'frame_00:02.png': 4.543545722961426}
short_About_Time_(2013)_-_First_Night_Together_Scene_(3_10)_|_Movieclips
../assets/split_scenes/short_About_Time_(2013)_-_First_Night_Together_Scene_(3_10)_|_Movieclips/short_scene_001.mp4
0.152581
0.068494
{'frame_00:00.png': [{'<COLOR>': 'Warm, Desaturated'}, {'<LIGHTING>': 'Soft light, Low contrast, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:01.png': [{'<COLOR>': 'Warm, Desaturated'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:02.png': [{'<COLOR>': 'Warm, Desaturated'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light'}, {'<COMPOSITION>': 'Left heavy'}]}
1.666667
0.333333
0
{"0": {"raw_counts": {"suitcase": 1, "total": 1}, "bucketed_counts": {"humans": 0, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 1}}, "1": {"raw_counts": {"suitcase": 2, "total": 2}, "bucketed_counts": {"humans": 0, "animals": 0, "vehicles": 0, "food": 0, "others": 2, "total": 2}}, "2": {"raw_counts": {"person": 1, "suitcase": 1, "total": 2}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 2}}}
-0.06101
[-1 -1 1 -1 -1 -1 1 -1 -1 -1 -1 -1 1 -1 1 -1 1 -1 -1 -1 -1 -1 1 -1 -1 1]
https://youtube.com/watch?v=AX-b2bm93lY
scene_001
1
658
725
00:00:27.444
00:00:30.239
27.444111
30.238572
2.794461
27.444111
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
The interior shot reveals a dimly lit room, where patterned curtains hang to the left of the frame, partially obscuring what lies beyond. Adjacent to the curtains, a pale green wall creates a narrow corridor leading deeper into the space. A chair, draped with red clothing, sits amidst the shadows, with a vague shape above it. To the right, a bedside table is illuminated by a lamp, its wire snaking across the surface, hinting at a lived-in space. Suddenly, a person emerges from behind the curtains, their face and form silhouetted against the brighter light from the window, creating a dramatic contrast in the muted color palette of the room.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A figure in a dark outfit cautiously peeks out from behind patterned curtains, creating a tense and suspenseful atmosphere within the dimly lit room. The scene is composed with contrasting textures and shadows that contribute to a sense of mystery and potential danger.
2.794461
1,920
810
67
1920x810
short_scene_001_short_About_Time_(2013)_-_Last_Night_Was_Never_Going_to_Work_Scene_(1_10)_|_Movieclips
5.093426
{'frame_00:00.png': 5.160425186157227, 'frame_00:01.png': 4.605077266693115, 'frame_00:02.png': 5.475815296173096, 'frame_00:03.png': 5.111115455627441, 'frame_00:04.png': 5.114694595336914}
short_About_Time_(2013)_-_Last_Night_Was_Never_Going_to_Work_Scene_(1_10)_|_Movieclips
../assets/split_scenes/short_About_Time_(2013)_-_Last_Night_Was_Never_Going_to_Work_Scene_(1_10)_|_Movieclips/short_scene_001.mp4
0.142039
0.284641
{'frame_00:00.png': [{'<COLOR>': 'Warm, Desaturated'}, {'<LIGHTING>': 'Soft light, Low contrast, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:01.png': [{'<COLOR>': 'Warm, Desaturated'}, {'<LIGHTING>': 'Soft light, Low contrast, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:02.png': [{'<COLOR>': 'Warm, Desaturated'}, {'<LIGHTING>': 'Soft light, Low contrast, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:03.png': [{'<COLOR>': 'Warm, Desaturated'}, {'<LIGHTING>': 'Soft light, Low contrast, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Center'}], 'frame_00:04.png': [{'<COLOR>': 'Warm, Desaturated'}, {'<LIGHTING>': 'Soft light, Low contrast, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Center'}]}
1
1
0
{"0": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "1": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "2": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "3": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "4": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}}
0.021972
[-1 1 1 1 -1 1 1 -1 -1 1 1 1 1 -1 1 -1 -1 1 1 -1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=SaOjJ3HjbDs
scene_001
1
53
173
00:00:02.211
00:00:07.216
2.210544
7.215549
5.005005
2.210544
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A man with reddish hair opens a door and steps partway into a room. He wears a dark t-shirt and is framed by the white doorframe. To the left, there's a framed piece of art on the wall with a pale, abstract design. To his right, there's a dark wooden door. The room beyond is dimly lit. The man appears to be speaking, looking directly ahead with a slight smile. He then briefly lowers his head before looking up again and continuing to speak. As he does, the camera zooms in slightly, providing a tighter shot of him in the doorway, before zooming back to the original shot. His posture and expression suggest he is in the midst of a casual conversation.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A man stands in a doorway, wearing a black t-shirt, as he seems to be communicating something to someone off-screen. The warm-toned scene showcases the man framed by the doorway.
5.005005
1,920
810
120
1920x810
short_scene_001_short_About_Time_(2013)_-_Live_Every_Day_Twice_Scene_(8_10)_|_Movieclips
4.740745
{'frame_00:00.png': 4.813418865203857, 'frame_00:01.png': 4.754183292388916, 'frame_00:02.png': 4.598450660705566, 'frame_00:03.png': 4.770415306091309, 'frame_00:04.png': 4.767258644104004}
short_About_Time_(2013)_-_Live_Every_Day_Twice_Scene_(8_10)_|_Movieclips
../assets/split_scenes/short_About_Time_(2013)_-_Live_Every_Day_Twice_Scene_(8_10)_|_Movieclips/short_scene_001.mp4
1.503092
0.630107
{'frame_00:00.png': [{'<COLOR>': 'Green'}, {'<LIGHTING>': 'Soft light, Low contrast, Backlight'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny, Artificial light, Mixed light, Fluorescent'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:01.png': [{'<COLOR>': 'Green'}, {'<LIGHTING>': 'Soft light, Low contrast, Backlight'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:02.png': [{'<COLOR>': 'Green'}, {'<LIGHTING>': 'Soft light, Backlight'}, {'<LIGHTING_TYPE>': 'Daylight, Artificial light, Fluorescent'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:03.png': [{'<COLOR>': 'Cyan'}, {'<LIGHTING>': 'Soft light, Backlight'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny, Artificial light, Fluorescent'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:04.png': [{'<COLOR>': 'Green'}, {'<LIGHTING>': 'Soft light, Backlight'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Balanced'}]}
16.2
5.4
0
{"0": {"raw_counts": {"person": 7, "chair": 4, "cup": 1, "tie": 1, "book": 2, "total": 15}, "bucketed_counts": {"humans": 7, "animals": 0, "vehicles": 0, "food": 0, "others": 8, "total": 15}}, "1": {"raw_counts": {"person": 6, "chair": 4, "book": 3, "tie": 1, "cup": 1, "laptop": 1, "total": 16}, "bucketed_counts": {"humans": 6, "animals": 0, "vehicles": 0, "food": 0, "others": 10, "total": 16}}, "2": {"raw_counts": {"person": 4, "chair": 5, "tie": 2, "book": 5, "mouse": 2, "cup": 2, "total": 20}, "bucketed_counts": {"humans": 4, "animals": 0, "vehicles": 0, "food": 0, "others": 16, "total": 20}}, "3": {"raw_counts": {"person": 4, "chair": 3, "cup": 1, "mouse": 2, "tie": 3, "book": 2, "total": 15}, "bucketed_counts": {"humans": 4, "animals": 0, "vehicles": 0, "food": 0, "others": 11, "total": 15}}, "4": {"raw_counts": {"person": 6, "cup": 2, "chair": 4, "book": 1, "mouse": 1, "tie": 1, "total": 15}, "bucketed_counts": {"humans": 6, "animals": 0, "vehicles": 0, "food": 0, "others": 9, "total": 15}}}
0.056706
[-1 1 1 1 1 1 1 -1 -1 1 1 -1 1 -1 1 -1 -1 -1 1 1 -1 1 1 -1 1 1]
https://youtube.com/watch?v=T1uRDmrPbp4
scene_001
1
310
430
00:00:12.930
00:00:17.935
12.929596
17.934601
5.005005
12.929596
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A group of five professionals is gathered around a round conference table in a brightly lit office. All are dressed in dark suits, suggesting a formal or business setting. The table is cluttered with stacks of file folders marked "Queen's Court Chambers," along with glasses of water and coffee cups. The individual at the left of the frame, with curly hair, picks up a document and hands it across the table to a red-haired colleague. As he hands the document, he gestures emphatically with his other hand, adding a sense of urgency or importance to the exchange. The woman between them continues to look down and write. The overall atmosphere suggests a busy, focused, and perhaps tense meeting, with a clear sense of collaboration and discussion taking place.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A group of legal professionals are sitting around a table in a conference room, with one man standing and passing out documents. The scene portrays an atmosphere of serious discussion and collaboration.
5.005005
1,920
810
120
1920x810
short_scene_001_short_About_Time_(2013)_-_Meeting_Mary_Scene_(2_10)_|_Movieclips
5.107698
{'frame_00:00.png': 5.202531337738037, 'frame_00:01.png': 5.314784049987793, 'frame_00:02.png': 4.978193759918213, 'frame_00:03.png': 5.213949680328369, 'frame_00:04.png': 4.829032897949219}
short_About_Time_(2013)_-_Meeting_Mary_Scene_(2_10)_|_Movieclips
../assets/split_scenes/short_About_Time_(2013)_-_Meeting_Mary_Scene_(2_10)_|_Movieclips/short_scene_001.mp4
-0.197665
0.116018
{'frame_00:00.png': [{'<COLOR>': 'Warm, Desaturated'}, {'<LIGHTING>': 'Soft light, Side light, Backlight, Edge light'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:01.png': [{'<COLOR>': 'Warm, Desaturated'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:02.png': [{'<COLOR>': 'Warm, Saturated'}, {'<LIGHTING>': 'Soft light, Side light, Edge light'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:03.png': [{'<COLOR>': 'Warm, Desaturated'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:04.png': [{'<COLOR>': 'Warm, Desaturated'}, {'<LIGHTING>': 'Soft light, Side light, Edge light'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light'}, {'<COMPOSITION>': 'Left heavy'}]}
3
2
0
{"0": {"raw_counts": {"person": 2, "tie": 1, "total": 3}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 3}}, "1": {"raw_counts": {"person": 2, "tie": 1, "total": 3}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 3}}, "2": {"raw_counts": {"person": 2, "tie": 2, "total": 4}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 2, "total": 4}}, "3": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "4": {"raw_counts": {"person": 2, "microwave": 1, "total": 3}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 3}}}
0.012416
[-1 -1 1 1 -1 1 1 -1 -1 1 -1 -1 1 -1 1 1 1 -1 1 -1 -1 -1 1 -1 -1 1]
https://youtube.com/watch?v=dgMKzky9S4I
scene_001
1
266
385
00:00:11.094
00:00:16.058
11.094428
16.057724
4.963297
11.094428
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
Two men are seen standing outside at night, one in a black suit with a patterned shirt and the other in a black suit with a red shirt. The man on the left is gesturing with his hands as he speaks, while the man on the right is looking at him with a serious expression. The man on the left begins to move out of the frame and the man on the right starts to speak. The scene then focuses on the man on the right, who continues to speak while looking directly at the camera. He is standing in front of a storefront with windows displaying clothing and accessories. The background is dark and blurry, with some lights visible in the distance.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
Two men are seen in a dark setting, with one individual in a patterned shirt and the other in a solid red shirt, both wearing dark blazers. They appear to be engaged in a conversation, with one turning to address the other.
4.963297
1,920
810
119
1920x810
short_scene_001_short_About_Time_(2013)_-_Stripping_for_Decisions_Scene_(5_10)_|_Movieclips
4.605716
{'frame_00:00.png': 4.338836193084717, 'frame_00:01.png': 4.470173358917236, 'frame_00:02.png': 4.782464981079102, 'frame_00:03.png': 4.831387519836426}
short_About_Time_(2013)_-_Stripping_for_Decisions_Scene_(5_10)_|_Movieclips
../assets/split_scenes/short_About_Time_(2013)_-_Stripping_for_Decisions_Scene_(5_10)_|_Movieclips/short_scene_001.mp4
-0.016094
0.244797
{'frame_00:00.png': [{'<COLOR>': 'Warm, Desaturated'}, {'<LIGHTING>': 'Soft light, Low contrast, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Center'}], 'frame_00:01.png': [{'<COLOR>': 'Warm, Desaturated'}, {'<LIGHTING>': 'Soft light, Low contrast, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Center'}], 'frame_00:02.png': [{'<COLOR>': 'Warm, Saturated'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:03.png': [{'<COLOR>': 'Warm, Desaturated'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}]}
1
1
0
{"0": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "1": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "2": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "3": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}}
0.015008
[-1 1 1 1 -1 1 1 -1 -1 1 -1 -1 1 -1 1 1 1 -1 1 -1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=keKt8-uDTN0
scene_001
1
36
119
00:00:01.502
00:00:04.963
1.501502
4.963297
3.461795
1.501502
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
The scene unfolds indoors with a woman at the center, dressed in a bright red zip-up hoodie. The room is dimly lit, creating a cozy yet somewhat somber atmosphere. In the background, a wall is adorned with what appears to be a floral patterned tapestry, adding depth and texture to the setting. The camera slightly shifts its angle as the woman seems to be in conversation, her expressions changing from an initial open-mouthed pose to a more contemplative look. The lighting casts soft shadows, highlighting her features as she engages with someone off-screen, transitioning from a direct gaze to looking downwards, suggesting thoughtfulness or concern. Towards the end, a window with light filtering through it is visible behind her.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A woman stands indoors wearing a red zip-up hoodie. Her expression shifts, suggesting an internal conflict or realization as she engages in a conversation.
3.461795
1,920
810
83
1920x810
short_scene_001_short_Accepted_(10_10)_Movie_CLIP_-_Bartleby's_Speech_(2006)_HD
5.062279
{'frame_00:00.png': 5.045755386352539, 'frame_00:01.png': 5.0430588722229, 'frame_00:02.png': 5.128672122955322, 'frame_00:03.png': 4.987903594970703, 'frame_00:04.png': 5.106006622314453}
short_Accepted_(10_10)_Movie_CLIP_-_Bartleby's_Speech_(2006)_HD
../assets/split_scenes/short_Accepted_(10_10)_Movie_CLIP_-_Bartleby's_Speech_(2006)_HD/short_scene_001.mp4
0.284083
1.213081
{'frame_00:00.png': [{'<COLOR>': 'Warm, Saturated, Green'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Daylight, Artificial light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:01.png': [{'<COLOR>': 'Warm, Saturated, Green'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Center'}], 'frame_00:02.png': [{'<COLOR>': 'Warm, Saturated, Green'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:03.png': [{'<COLOR>': 'Warm, Saturated, Green'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Center'}], 'frame_00:04.png': [{'<COLOR>': 'Warm, Saturated, Green'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Center'}]}
8.8
7.8
0
{"0": {"raw_counts": {"person": 8, "tie": 1, "total": 9}, "bucketed_counts": {"humans": 8, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 9}}, "1": {"raw_counts": {"person": 7, "tie": 1, "total": 8}, "bucketed_counts": {"humans": 7, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 8}}, "2": {"raw_counts": {"person": 8, "tie": 1, "total": 9}, "bucketed_counts": {"humans": 8, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 9}}, "3": {"raw_counts": {"person": 8, "tie": 1, "total": 9}, "bucketed_counts": {"humans": 8, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 9}}, "4": {"raw_counts": {"person": 8, "tie": 1, "total": 9}, "bucketed_counts": {"humans": 8, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 9}}}
0.021972
[-1 1 1 1 -1 1 1 -1 -1 1 1 1 1 -1 1 -1 -1 1 1 -1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=KZtF_gT3Sgg
scene_001
1
105
210
00:00:04.379
00:00:08.759
4.379379
8.758759
4.379379
4.379379
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
In a brightly lit, auditorium-like setting, a man wearing a dark blazer over a patterned shirt is seated, initially looking down before glancing up and gesturing with his fingers, first displaying three fingers, then one. The scene is densely populated with people of various ages and attire filling the rows behind him, some appearing attentive, others with more neutral or serious expressions. One man, seated to the left of the gesturing individual, wears a light tan blazer and a collared shirt. Further back, more individuals are visible, filling the frame and contributing to the atmosphere of a crowded event.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A man sits in an audience and gestures with his hands, seemingly contributing to a conversation or discussion. The scene is set in a lecture hall or auditorium filled with a diverse group of people.
4.379379
1,280
522
105
1280x522
short_scene_001_short_Accepted_(1_10)_Movie_CLIP_-_An_Invitation_from_Monica_(2006)_HD
4.471018
{'frame_00:00.png': 4.67167854309082, 'frame_00:01.png': 4.207024574279785, 'frame_00:02.png': 4.415505409240723, 'frame_00:03.png': 4.589863300323486}
short_Accepted_(1_10)_Movie_CLIP_-_An_Invitation_from_Monica_(2006)_HD
../assets/split_scenes/short_Accepted_(1_10)_Movie_CLIP_-_An_Invitation_from_Monica_(2006)_HD/short_scene_001.mp4
-0.067386
4.148759
{'frame_00:00.png': [{'<COLOR>': 'Blue'}, {'<LIGHTING>': 'Soft light'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:01.png': [{'<COLOR>': 'Blue'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Center'}], 'frame_00:02.png': [{'<COLOR>': 'Yellow'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:03.png': [{'<COLOR>': 'Yellow'}, {'<LIGHTING>': 'Soft light, Low contrast, Top light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Balanced'}]}
3.75
3.75
0
{"0": {"raw_counts": {"person": 5, "total": 5}, "bucketed_counts": {"humans": 5, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 5}}, "1": {"raw_counts": {"person": 4, "total": 4}, "bucketed_counts": {"humans": 4, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 4}}, "2": {"raw_counts": {"person": 3, "total": 3}, "bucketed_counts": {"humans": 3, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 3}}, "3": {"raw_counts": {"person": 3, "total": 3}, "bucketed_counts": {"humans": 3, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 3}}}
-0.035844
[-1 -1 1 1 -1 1 1 -1 -1 1 -1 -1 -1 -1 1 1 1 -1 -1 -1 -1 -1 1 -1 -1 1]
https://youtube.com/watch?v=RehkdOxmytI
scene_001
1
38
126
00:00:01.585
00:00:05.255
1.584918
5.255255
3.670337
1.584918
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
The scene is set outside a building with "William McKinley" inscribed above the doorway. A young man dashes out of the building's doors, appearing frantic, as other students casually exit behind him. A second young man with curly hair and glasses walks into the frame, looking bewildered and slightly concerned. He is accompanied by another student with a backpack, who seems to be gesturing with his hand as if in conversation or explanation. In the background, a man in a suit walks away from the building, adding to the bustling atmosphere of a school dismissal. The lighting is bright and sunny, suggesting an upbeat, everyday scenario despite the initial frantic action.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
Several young people are seen walking out of a school building with the name "William McKinley" on the entrance; one student is running excitedly as a group of others are walking and conversing calmly. The atmosphere appears to be ordinary for a school setting.
3.670337
1,280
558
88
1280x558
short_scene_001_short_Accepted_(3_10)_Movie_CLIP_-_The_Birthplace_of_Crack_(2006)_HD
4.888119
{'frame_00:00.png': 4.836962699890137, 'frame_00:01.png': 4.982293128967285, 'frame_00:02.png': 4.668661117553711, 'frame_00:03.png': 5.010646820068359, 'frame_00:04.png': 4.942032337188721}
short_Accepted_(3_10)_Movie_CLIP_-_The_Birthplace_of_Crack_(2006)_HD
../assets/split_scenes/short_Accepted_(3_10)_Movie_CLIP_-_The_Birthplace_of_Crack_(2006)_HD/short_scene_001.mp4
3.542548
1.80945
{'frame_00:00.png': [{'<COLOR>': 'Cyan'}, {'<LIGHTING>': 'Soft light, Low contrast, Side light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:01.png': [{'<COLOR>': 'Cyan'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:02.png': [{'<COLOR>': 'Cyan'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:03.png': [{'<COLOR>': 'Cyan'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:04.png': [{'<COLOR>': 'Cyan, Blue'}, {'<LIGHTING>': 'Soft light, Low contrast, Side light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Balanced'}]}
8.6
3.4
0
{"0": {"raw_counts": {"person": 3, "car": 5, "total": 8}, "bucketed_counts": {"humans": 3, "animals": 0, "vehicles": 5, "food": 0, "others": 0, "total": 8}}, "1": {"raw_counts": {"person": 3, "car": 5, "total": 8}, "bucketed_counts": {"humans": 3, "animals": 0, "vehicles": 5, "food": 0, "others": 0, "total": 8}}, "2": {"raw_counts": {"person": 3, "car": 6, "total": 9}, "bucketed_counts": {"humans": 3, "animals": 0, "vehicles": 6, "food": 0, "others": 0, "total": 9}}, "3": {"raw_counts": {"person": 4, "car": 6, "total": 10}, "bucketed_counts": {"humans": 4, "animals": 0, "vehicles": 6, "food": 0, "others": 0, "total": 10}}, "4": {"raw_counts": {"person": 4, "car": 4, "total": 8}, "bucketed_counts": {"humans": 4, "animals": 0, "vehicles": 4, "food": 0, "others": 0, "total": 8}}}
0.038481
[-1 1 1 1 1 1 1 -1 -1 1 1 1 1 -1 1 -1 1 1 1 -1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=TWjtzvFrA8c
scene_001
1
79
181
00:00:03.295
00:00:07.549
3.294962
7.549216
4.254254
3.294962
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
Three people stand together in an outdoor parking lot while having a conversation. The person on the right is wearing a green and white striped polo shirt, and the person on the left is wearing a sleeveless white shirt. The person in the middle has their back to the camera and is wearing a gray t-shirt. They stand near the back of a white car. In the background, other parked cars are visible, along with buildings and trees under a sunny sky.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
Three people stand beside a car, seemingly in conversation on a street, with other cars driving on the street. One person gestures while talking, with buildings and trees visible in the background.
4.254254
1,280
540
102
1280x540
short_scene_001_short_Accepted_(4_10)_Movie_CLIP_-_We_Gotta_Find_a_Dean_(2006)_HD
5.032938
{'frame_00:00.png': 5.003552436828613, 'frame_00:01.png': 4.821330547332764, 'frame_00:02.png': 5.122524738311768, 'frame_00:03.png': 5.184344291687012}
short_Accepted_(4_10)_Movie_CLIP_-_We_Gotta_Find_a_Dean_(2006)_HD
../assets/split_scenes/short_Accepted_(4_10)_Movie_CLIP_-_We_Gotta_Find_a_Dean_(2006)_HD/short_scene_001.mp4
-0.04475
0.715043
{'frame_00:00.png': [{'<COLOR>': 'Green, Cyan'}, {'<LIGHTING>': 'Hard light, Low contrast, Side light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Center'}], 'frame_00:01.png': [{'<COLOR>': 'Green, Cyan'}, {'<LIGHTING>': 'Hard light, Low contrast, Side light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:02.png': [{'<COLOR>': 'Green, Cyan'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:03.png': [{'<COLOR>': 'Green, Cyan'}, {'<LIGHTING>': 'Soft light, Low contrast, Side light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Left heavy'}]}
1
1
0
{"0": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "1": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "2": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "3": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}}
0.080473
[-1 1 1 1 1 1 1 -1 -1 1 1 1 1 1 1 -1 -1 1 1 1 -1 1 1 -1 1 1]
https://youtube.com/watch?v=mHJH39bjALk
scene_001
1
109
192
00:00:04.546
00:00:08.008
4.546213
8.008008
3.461795
4.546213
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A person with brown hair stands outside against a backdrop of trees and a clear sky, wearing a gray T-shirt with a graphic on it and a partially unbuttoned light-colored patterned shirt over the tee, looking slightly upward with a somewhat concerned or thoughtful expression on their face, as a flag on a pole waves gently in the background, enhancing the outdoor atmosphere of the scene.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A young man is standing outdoors under a bright sky, wearing a gray t-shirt and a white patterned shirt over it. The background features trees and a flag pole, suggesting a public park or school setting.
3.461795
1,280
516
83
1280x516
short_scene_001_short_Accepted_(5_10)_Movie_CLIP_-_Breeding_Pimps_and_Whores_(2006)_HD
4.578389
{'frame_00:00.png': 4.4029951095581055, 'frame_00:01.png': 4.425415992736816, 'frame_00:02.png': 4.649081230163574, 'frame_00:03.png': 4.641293048858643, 'frame_00:04.png': 4.773157119750977}
short_Accepted_(5_10)_Movie_CLIP_-_Breeding_Pimps_and_Whores_(2006)_HD
../assets/split_scenes/short_Accepted_(5_10)_Movie_CLIP_-_Breeding_Pimps_and_Whores_(2006)_HD/short_scene_001.mp4
0.720904
3.468104
{'frame_00:00.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:01.png': [{'<COLOR>': 'Cool, Desaturated, Blue'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:02.png': [{'<COLOR>': 'Cool, Desaturated, Blue'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:03.png': [{'<COLOR>': 'Cool, Desaturated, Blue'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:04.png': [{'<COLOR>': 'Blue'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Daylight, Artificial light'}, {'<COMPOSITION>': 'Right heavy'}]}
2
1
0
{"0": {"raw_counts": {"person": 1, "tie": 2, "total": 3}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 2, "total": 3}}, "1": {"raw_counts": {"person": 1, "tie": 1, "total": 2}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 2}}, "2": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "3": {"raw_counts": {"person": 1, "tie": 1, "total": 2}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 2}}, "4": {"raw_counts": {"person": 1, "tie": 1, "total": 2}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 2}}}
0.073328
[-1 1 1 1 1 1 1 -1 -1 1 1 1 1 1 1 -1 1 1 1 1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=e8HRcYG3Ftg
scene_001
1
46
166
00:00:01.919
00:00:06.924
1.918585
6.92359
5.005005
1.918585
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A man with curly gray hair and glasses is seated in a black leather office chair, wearing a blue blazer over a patterned collared shirt. He is positioned behind a desk in an office setting. The background includes blinds on a window and a lamp. Throughout the scene, the man gestures with his hands as he speaks, his facial expressions changing from a slight smile to a more intense, almost angry look. The camera remains focused on him, capturing his expressions and gestures in detail as the scene progresses.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A man sits in an office and talks, making hand gestures as he speaks. He appears to be in a serious conversation.
5.005005
1,280
522
120
1280x522
short_scene_001_short_Accepted_(7_10)_Movie_CLIP_-_Welcome_to_S_H_I_T__(2006)_HD
4.725347
{'frame_00:00.png': 4.597132205963135, 'frame_00:01.png': 4.844489574432373, 'frame_00:02.png': 4.774776935577393, 'frame_00:03.png': 4.684988975524902}
short_Accepted_(7_10)_Movie_CLIP_-_Welcome_to_S_H_I_T__(2006)_HD
../assets/split_scenes/short_Accepted_(7_10)_Movie_CLIP_-_Welcome_to_S_H_I_T__(2006)_HD/short_scene_001.mp4
2.214407
0.169556
{'frame_00:00.png': [{'<COLOR>': 'Cool, Desaturated, Cyan, Blue'}, {'<LIGHTING>': 'Soft light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:01.png': [{'<COLOR>': 'Blue'}, {'<LIGHTING>': 'Soft light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:02.png': [{'<COLOR>': 'Blue'}, {'<LIGHTING>': 'Soft light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:03.png': [{'<COLOR>': 'Blue'}, {'<LIGHTING>': 'Soft light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Balanced'}]}
8.5
8.25
0.25
{"0": {"raw_counts": {"person": 8, "total": 8}, "bucketed_counts": {"humans": 8, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 8}}, "1": {"raw_counts": {"person": 8, "horse": 1, "total": 9}, "bucketed_counts": {"humans": 8, "animals": 1, "vehicles": 0, "food": 0, "others": 0, "total": 9}}, "2": {"raw_counts": {"person": 6, "total": 6}, "bucketed_counts": {"humans": 6, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 6}}, "3": {"raw_counts": {"person": 11, "total": 11}, "bucketed_counts": {"humans": 11, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 11}}}
0.016892
[-1 1 1 1 -1 1 1 -1 -1 1 -1 -1 1 -1 1 -1 1 -1 1 -1 -1 -1 1 -1 1 1]
https://youtube.com/watch?v=QZjmr4_K8To
scene_001
1
40
118
00:00:01.668
00:00:04.922
1.668335
4.921588
3.253253
1.668335
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
An eye-level shot shows a crowded classroom with students and a professor. Most of the students are in the front of the room and face forward, and some are looking at the professor as he walks in. He is wearing a straw fedora, a tan button-down shirt over a white t-shirt, and glasses. He stands in the back of the room with his arms crossed. The classroom is bland and has posters on the wall.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A man enters a classroom full of students, looking perturbed. The students turn their heads towards him, watching him as he walks into the room.
3.253253
1,280
540
78
1280x540
short_scene_001_short_Accepted_(8_10)_Movie_CLIP_-_What_Do_You_Want_to_Learn?_(2006)_HD
4.163639
{'frame_00:00.png': 4.316036224365234, 'frame_00:01.png': 3.9258599281311035, 'frame_00:02.png': 4.1168928146362305, 'frame_00:03.png': 4.206591606140137, 'frame_00:04.png': 4.252816677093506}
short_Accepted_(8_10)_Movie_CLIP_-_What_Do_You_Want_to_Learn?_(2006)_HD
../assets/split_scenes/short_Accepted_(8_10)_Movie_CLIP_-_What_Do_You_Want_to_Learn?_(2006)_HD/short_scene_001.mp4
-0.813918
6.795064
{'frame_00:00.png': [{'<COLOR>': 'Cool, Desaturated, Cyan, Blue'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:01.png': [{'<COLOR>': 'Cool, Desaturated, Cyan, Blue'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Center'}], 'frame_00:02.png': [{'<COLOR>': 'Cool, Cyan, Blue'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:03.png': [{'<COLOR>': 'Cool, Desaturated, Cyan, Blue'}, {'<LIGHTING>': 'Soft light, Low contrast, Backlight'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:04.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Daylight'}, {'<COMPOSITION>': 'Right heavy'}]}
1.6
1.6
0
{"0": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "1": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "2": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "3": {"raw_counts": {"person": 3, "total": 3}, "bucketed_counts": {"humans": 3, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 3}}, "4": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}}
-0.047838
[-1 -1 1 1 -1 -1 1 -1 -1 1 -1 -1 1 -1 1 1 1 -1 -1 -1 -1 -1 -1 -1 -1 1]
https://youtube.com/watch?v=lQfG0D7wPKA
scene_001
1
30
150
00:00:01.251
00:00:06.256
1.251251
6.256256
5.005005
1.251251
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A man in a light-colored, short-sleeved shirt stands in a room with large blinds covering the windows, conversing with others as they enter the frame. The camera follows his movement as he turns and walks, engaging with people who approach him. One individual, wearing a blue and red striped shirt, stands with their arms crossed, observing the interaction. Another person with curly hair and a red headband enters the scene, engaging briefly with the man in the light-colored shirt before he continues moving across the room. The lighting in the room is bright, creating shadows from the blinds, adding depth to the scene.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
Several people stand around in an indoor room setting, with blinds covering a window in the background. The individual in the center appears to be walking toward the side.
5.005005
1,280
486
120
1280x486
short_scene_001_short_Accepted_(9_10)_Movie_CLIP_-_Ask_Me_About_My_Wiener_(2006)_HD
4.784815
{'frame_00:00.png': 4.633172035217285, 'frame_00:01.png': 4.746738910675049, 'frame_00:02.png': 4.929875373840332, 'frame_00:03.png': 4.7999091148376465, 'frame_00:04.png': 4.814380645751953}
short_Accepted_(9_10)_Movie_CLIP_-_Ask_Me_About_My_Wiener_(2006)_HD
../assets/split_scenes/short_Accepted_(9_10)_Movie_CLIP_-_Ask_Me_About_My_Wiener_(2006)_HD/short_scene_001.mp4
-0.773618
2.626359
{'frame_00:00.png': [{'<COLOR>': 'Saturated, Red, Green'}, {'<LIGHTING>': 'Soft light, Low contrast, Side light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:01.png': [{'<COLOR>': 'Green, Blue, White'}, {'<LIGHTING>': 'Soft light, Low contrast, Side light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:02.png': [{'<COLOR>': 'Warm, Saturated, Green'}, {'<LIGHTING>': 'Soft light, Low contrast, Side light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:03.png': [{'<COLOR>': 'Green'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:04.png': [{'<COLOR>': 'Green'}, {'<LIGHTING>': 'Soft light, Low contrast, Side light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Balanced'}]}
2.4
2.4
0
{"0": {"raw_counts": {"person": 4, "total": 4}, "bucketed_counts": {"humans": 4, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 4}}, "1": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "2": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "3": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "4": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}}
0.079715
[ 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 -1 1 1 1 1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=eP52omnnZmg
scene_001
1
61
181
00:00:02.544
00:00:07.549
2.544211
7.549216
5.005005
2.544211
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A young man and woman are walking together outside, seemingly engaged in conversation. The man, wearing a blue t-shirt under a light-striped button-down shirt, is looking attentively at the woman. She is blonde and wearing a brown top under a grey shirt with the sleeves rolled up. She gestures with her hands as she speaks, her expression animated. Two other girls are walking slightly behind them to the right, one with glasses and a light-colored shirt, and the other with blonde hair wearing a red cardigan. They are all outdoors in what appears to be a sunny day, with lush green bushes and trees serving as the backdrop. The setting has a casual, friendly vibe.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A young man and woman are walking together outdoors while talking; there are other girls walking in the background. The lighting is bright and natural, casting subtle shadows as the group strolls along a path lined with lush greenery.
5.005005
1,280
540
120
1280x540
short_scene_001_short_Action_Point_(2018)_-_Bad_Publicity_is_Good_Publicity_Scene_(4_10)_|_Movieclips
4.86844
{'frame_00:00.png': 4.707772731781006, 'frame_00:01.png': 4.995969772338867, 'frame_00:02.png': 4.836220741271973, 'frame_00:03.png': 4.93379545211792}
short_Action_Point_(2018)_-_Bad_Publicity_is_Good_Publicity_Scene_(4_10)_|_Movieclips
../assets/split_scenes/short_Action_Point_(2018)_-_Bad_Publicity_is_Good_Publicity_Scene_(4_10)_|_Movieclips/short_scene_001.mp4
-0.394311
1.025623
{'frame_00:00.png': [{'<COLOR>': 'Warm, Desaturated'}, {'<LIGHTING>': 'Soft light, Low contrast, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:01.png': [{'<COLOR>': 'Warm, Desaturated'}, {'<LIGHTING>': 'Soft light, Low contrast, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:02.png': [{'<COLOR>': 'Warm, Desaturated'}, {'<LIGHTING>': 'Soft light, Low contrast, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:03.png': [{'<COLOR>': 'Warm, Desaturated'}, {'<LIGHTING>': 'Soft light, Low contrast, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Left heavy'}]}
3.25
3.25
0
{"0": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "1": {"raw_counts": {"person": 3, "total": 3}, "bucketed_counts": {"humans": 3, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 3}}, "2": {"raw_counts": {"person": 4, "total": 4}, "bucketed_counts": {"humans": 4, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 4}}, "3": {"raw_counts": {"person": 4, "total": 4}, "bucketed_counts": {"humans": 4, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 4}}}
-0.04186
[-1 -1 1 1 -1 1 1 -1 -1 -1 -1 -1 1 -1 1 -1 1 -1 -1 -1 -1 -1 1 -1 -1 1]
https://youtube.com/watch?v=9EFxRx8EbDk
scene_001
1
77
155
00:00:03.212
00:00:06.465
3.211545
6.464798
3.253253
3.211545
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A group of people is captured riding an escalator. The camera captures them from a slightly lower angle as they are ascending. On the left side of the frame, there are individuals in vintage clothing. One person is wearing a shirt with a bright, colorful pattern. Beside them, a young woman with long blonde hair tied in pigtails is wearing a yellow top with a ruffled collar. To her right, a man with dark hair and a mustache is wearing a brown checkered suit. He is looking down with a serious expression. The background is a plain, light-colored wall. The lighting is even and slightly warm, giving the scene a nostalgic feel.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A group of people are seen riding down an escalator as one character looks back with a serious expression. The scene conveys a sense of tension and anticipation.
3.253253
1,920
774
78
1920x774
short_scene_001_short_Action_Point_(2018)_-_Bus_Jump_Scene_(9_10)_|_Movieclips
4.839356
{'frame_00:00.png': 4.912605285644531, 'frame_00:01.png': 4.9318037033081055, 'frame_00:02.png': 4.9138946533203125, 'frame_00:03.png': 4.70851993560791, 'frame_00:04.png': 4.72995662689209}
short_Action_Point_(2018)_-_Bus_Jump_Scene_(9_10)_|_Movieclips
../assets/split_scenes/short_Action_Point_(2018)_-_Bus_Jump_Scene_(9_10)_|_Movieclips/short_scene_001.mp4
-0.750331
1.486316
{'frame_00:00.png': [{'<COLOR>': 'Warm'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:01.png': [{'<COLOR>': 'Warm'}, {'<LIGHTING>': 'Soft light, Backlight, Edge light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:02.png': [{'<COLOR>': 'Warm'}, {'<LIGHTING>': 'Soft light, Backlight, Edge light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:03.png': [{'<COLOR>': 'Warm'}, {'<LIGHTING>': 'Soft light, Backlight, Edge light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:04.png': [{'<COLOR>': 'Warm, Saturated'}, {'<LIGHTING>': 'Soft light, Backlight, Edge light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Right heavy'}]}
2.8
2
0
{"0": {"raw_counts": {"person": 2, "car": 1, "total": 3}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 1, "food": 0, "others": 0, "total": 3}}, "1": {"raw_counts": {"person": 2, "car": 1, "total": 3}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 1, "food": 0, "others": 0, "total": 3}}, "2": {"raw_counts": {"person": 2, "car": 1, "total": 3}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 1, "food": 0, "others": 0, "total": 3}}, "3": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "4": {"raw_counts": {"person": 2, "car": 1, "total": 3}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 1, "food": 0, "others": 0, "total": 3}}}
0.033975
[-1 1 1 1 1 1 1 -1 1 1 -1 1 1 -1 1 -1 1 -1 1 -1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=E4rFRdmeWqU
scene_001
1
42
162
00:00:01.752
00:00:06.757
1.751752
6.756757
5.005005
1.751752
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
Two men are in a white convertible, driving down a road with trees in the background. The man in the passenger seat is wearing a red and white striped tank top and has a beard and a braid in his hair. He is animatedly talking, while the man driving is listening intently. The driver is wearing a dark blazer and has a concerned expression on his face. The lighting suggests it is late afternoon or early evening. The scene is captured in medium close-up, focusing on the expressions and interaction between the two men.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
Two men are seen in a white convertible, seemingly engaged in an intense conversation or argument; one man drives with a serious expression, while the other, wearing a tank top, gestures and speaks animatedly. The scene is set against a backdrop of open land and trees, with the sun casting a warm glow over the setting.
5.005005
1,920
790
120
1920x790
short_scene_001_short_Action_Point_(2018)_-_Catching_Animals_Scene_(5_10)_|_Movieclips
5.056557
{'frame_00:00.png': 4.951317310333252, 'frame_00:01.png': 5.174883842468262, 'frame_00:02.png': 5.008605003356934, 'frame_00:03.png': 5.091423034667969}
short_Action_Point_(2018)_-_Catching_Animals_Scene_(5_10)_|_Movieclips
../assets/split_scenes/short_Action_Point_(2018)_-_Catching_Animals_Scene_(5_10)_|_Movieclips/short_scene_001.mp4
-0.058892
0.827031
{'frame_00:00.png': [{'<COLOR>': 'Warm, Saturated, Yellow'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:01.png': [{'<COLOR>': 'Warm, Saturated, Yellow'}, {'<LIGHTING>': 'Soft light, Backlight, Edge light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:02.png': [{'<COLOR>': 'Warm, Saturated, Green'}, {'<LIGHTING>': 'Soft light, Side light, Edge light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:03.png': [{'<COLOR>': 'Warm, Saturated, Green'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Right heavy'}]}
4
4
0
{"0": {"raw_counts": {"person": 3, "total": 3}, "bucketed_counts": {"humans": 3, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 3}}, "1": {"raw_counts": {"person": 4, "total": 4}, "bucketed_counts": {"humans": 4, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 4}}, "2": {"raw_counts": {"person": 5, "total": 5}, "bucketed_counts": {"humans": 5, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 5}}, "3": {"raw_counts": {"person": 4, "total": 4}, "bucketed_counts": {"humans": 4, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 4}}}
0.078167
[-1 1 1 1 1 1 1 -1 1 1 1 1 1 -1 1 -1 -1 1 1 1 -1 1 1 1 1 1]
https://youtube.com/watch?v=0Dp--gKKMJ8
scene_001
1
99
181
00:00:04.129
00:00:07.549
4.129129
7.549216
3.420087
4.129129
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A woman in a white t-shirt with green trim and the words "TEENAGE KICKS" printed on it stands in a sun-drenched outdoor setting, facing a man whose back is to the camera. She has headphones around her neck and a yellow purse strap over her shoulder. Two young boys are standing nearby, their heads visible near the man. The background includes trees and colorful triangular flags hanging from a string, suggesting a festive or outdoor event. The lighting is bright, casting warm tones across the scene, indicating it is daytime. The woman is looking at the man and appears to be in conversation.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A fair-haired person wearing a "Teenage Kicks" t-shirt is seen looking at someone, appearing to talk or listen intently during an outdoor gathering with colorful pennants hanging behind them. The scene shows other people nearby, including children and someone in a suit, suggesting a casual, possibly celebratory, event.
3.420087
1,920
810
82
1920x810
short_scene_001_short_Action_Point_(2018)_-_Shooting_the_Commercial_Scene_(6_10)_|_Movieclips
4.961846
{'frame_00:00.png': 4.789478302001953, 'frame_00:01.png': 4.964664459228516, 'frame_00:02.png': 5.131393909454346}
short_Action_Point_(2018)_-_Shooting_the_Commercial_Scene_(6_10)_|_Movieclips
../assets/split_scenes/short_Action_Point_(2018)_-_Shooting_the_Commercial_Scene_(6_10)_|_Movieclips/short_scene_001.mp4
0.537986
0.418805
{'frame_00:00.png': [{'<COLOR>': 'Warm'}, {'<LIGHTING>': 'Hard light, Top light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:01.png': [{'<COLOR>': 'Warm'}, {'<LIGHTING>': 'Hard light, Top light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:02.png': [{'<COLOR>': 'Warm'}, {'<LIGHTING>': 'Hard light, Top light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Balanced'}]}
9
7.333333
0
{"0": {"raw_counts": {"person": 6, "umbrella": 1, "skis": 1, "total": 8}, "bucketed_counts": {"humans": 6, "animals": 0, "vehicles": 0, "food": 0, "others": 2, "total": 8}}, "1": {"raw_counts": {"person": 10, "suitcase": 1, "total": 11}, "bucketed_counts": {"humans": 10, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 11}}, "2": {"raw_counts": {"person": 6, "suitcase": 1, "umbrella": 1, "total": 8}, "bucketed_counts": {"humans": 6, "animals": 0, "vehicles": 0, "food": 0, "others": 2, "total": 8}}}
-0.007677
[-1 1 1 1 1 1 1 -1 -1 1 -1 -1 1 -1 1 -1 -1 -1 -1 -1 -1 1 1 -1 1 1]
https://youtube.com/watch?v=PyUqYOB4ey8
scene_001
1
111
171
00:00:04.630
00:00:07.132
4.62963
7.132132
2.502503
4.62963
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A group of individuals are captured in an outdoor setting, amidst a dirt-covered landscape suggesting a go-kart track, with low-lying sheds, tires and hills in the background; one person, donned in a blazer, stands assertively while holding a megaphone, while another is crouched behind a vintage camera on a tripod, seemingly preparing to film; nearby, a person in retro attire stands casually, while a fourth individual sits in a go-kart, ready for action, all of them contributing to an atmosphere of relaxed concentration before the start of the shoot or race.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A group of young adults is shown outdoors on a sunny day, possibly at a go-kart track, with one man speaking into a megaphone while another is filming. There is a lighthearted and youthful energy present.
2.502503
1,920
774
60
1920x774
short_scene_001_short_Action_Point_(2018)_-_The_Sh**birds_Scene_(1_10)_|_Movieclips
4.875343
{'frame_00:00.png': 4.994297981262207, 'frame_00:01.png': 4.803251266479492, 'frame_00:02.png': 4.828478813171387}
short_Action_Point_(2018)_-_The_Sh**birds_Scene_(1_10)_|_Movieclips
../assets/split_scenes/short_Action_Point_(2018)_-_The_Sh**birds_Scene_(1_10)_|_Movieclips/short_scene_001.mp4
0.044858
1.163457
{'frame_00:00.png': [{'<COLOR>': 'Warm, Cyan'}, {'<LIGHTING>': 'Hard light, Backlight, Edge light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:01.png': [{'<COLOR>': 'Warm, Cyan'}, {'<LIGHTING>': 'Soft light, Low contrast, Backlight'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:02.png': [{'<COLOR>': 'Warm, Cyan'}, {'<LIGHTING>': 'Soft light, Backlight'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Balanced'}]}
5.666667
5.666667
0
{"0": {"raw_counts": {"person": 6, "total": 6}, "bucketed_counts": {"humans": 6, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 6}}, "1": {"raw_counts": {"person": 5, "total": 5}, "bucketed_counts": {"humans": 5, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 5}}, "2": {"raw_counts": {"person": 6, "total": 6}, "bucketed_counts": {"humans": 6, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 6}}}
0.033975
[-1 1 1 1 1 1 1 -1 1 1 -1 1 1 -1 1 -1 1 -1 1 -1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=Vchw3dbVeUo
scene_001
1
61
124
00:00:02.544
00:00:05.172
2.544211
5.171839
2.627628
2.544211
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A group of people ride in the back of an old pick-up truck through a rural landscape. The truck's cab is black, and the bed is filled with several young individuals who are smiling and appear to be enjoying the ride. A woman is seated towards the right side, with her face expressing laughter and happiness as she looks towards the others. The landscape is vast and open, featuring dry grass, dirt, and sparse trees in the background under a clear, bright sky. The overall composition of the shot conveys a sense of carefree joy and camaraderie among the group as they journey through the countryside.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A group of people are riding in the back of a pickup truck, with open fields in the background and clear sky above them, appearing to be in good spirits and enjoying the ride. The warm lighting and candid shots suggest a carefree, joyful journey.
2.627628
1,920
790
63
1920x790
short_scene_001_short_Action_Point_(2018)_-_Trebuchet_Mishap_Scene_(8_10)_|_Movieclips
4.607521
{'frame_00:00.png': 4.580532550811768, 'frame_00:01.png': 4.629220008850098, 'frame_00:02.png': 4.612810134887695}
short_Action_Point_(2018)_-_Trebuchet_Mishap_Scene_(8_10)_|_Movieclips
../assets/split_scenes/short_Action_Point_(2018)_-_Trebuchet_Mishap_Scene_(8_10)_|_Movieclips/short_scene_001.mp4
-3.051006
0.237446
{'frame_00:00.png': [{'<COLOR>': 'Warm, Desaturated'}, {'<LIGHTING>': 'Hard light, High contrast, Backlight'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:01.png': [{'<COLOR>': 'Warm, Desaturated'}, {'<LIGHTING>': 'Hard light, Backlight'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:02.png': [{'<COLOR>': 'Warm, Desaturated, Yellow'}, {'<LIGHTING>': 'Hard light, Top light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Balanced'}]}
4.666667
3.666667
0
{"0": {"raw_counts": {"person": 4, "train": 1, "total": 5}, "bucketed_counts": {"humans": 4, "animals": 0, "vehicles": 1, "food": 0, "others": 0, "total": 5}}, "1": {"raw_counts": {"person": 2, "suitcase": 1, "total": 3}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 3}}, "2": {"raw_counts": {"person": 5, "train": 1, "total": 6}, "bucketed_counts": {"humans": 5, "animals": 0, "vehicles": 1, "food": 0, "others": 0, "total": 6}}}
-0.031331
[-1 1 1 1 -1 1 1 -1 -1 1 -1 -1 1 -1 1 -1 -1 -1 -1 -1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=UUAQ3T09_os
scene_001
1
124
193
00:00:05.172
00:00:08.050
5.171839
8.049716
2.877878
5.171839
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
The scene unfolds in a sun-drenched outdoor space, where a peculiar, oversized vehicle cobbled together from various scrap parts dominates the foreground. The vehicle, painted in a patchwork of colors, features a large scoop at the front, a rotating barrel-like structure at the top, and several wheels. People are scattered around the vehicle, some standing near it, others walking by in the background, creating a sense of activity and purpose around the strange machine. The atmosphere is relaxed and informal, suggesting a collaborative or community-driven project. The setting is rustic, with trees, wooden buildings, and scattered hay bales contributing to a sense of rural ingenuity.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
Several people are near a large, bizarre machine with various colors and metal pieces, set in an outdoor, wooded area with structures made of wood and metal. A person is standing atop the machine, and hay bales are nearby, suggesting a farm setting.
2.877878
1,920
810
69
1920x810
short_scene_001_short_Adam's_Rib_(1949)_-_A_Little_Slap_Scene_(5_10)_|_Movieclips
5.064623
{'frame_00:00.png': 4.913707256317139, 'frame_00:01.png': 5.1745476722717285, 'frame_00:02.png': 5.043380260467529, 'frame_00:03.png': 4.986936569213867, 'frame_00:04.png': 5.23340368270874, 'frame_00:05.png': 5.035763740539551}
short_Adam's_Rib_(1949)_-_A_Little_Slap_Scene_(5_10)_|_Movieclips
../assets/split_scenes/short_Adam's_Rib_(1949)_-_A_Little_Slap_Scene_(5_10)_|_Movieclips/short_scene_001.mp4
0.073651
2.028725
{'frame_00:00.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Center'}], 'frame_00:01.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Center'}], 'frame_00:02.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Center'}], 'frame_00:03.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Center'}], 'frame_00:04.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Center'}], 'frame_00:05.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Center'}]}
1
1
0
{"0": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "1": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "2": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "3": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "4": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "5": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}}
0.017523
[-1 1 1 -1 -1 1 1 -1 -1 1 1 1 1 -1 1 -1 -1 -1 1 -1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=6oqDO7aHVFo
scene_001
1
2,656
2,780
00:01:50.777
00:01:55.949
110.777444
115.949283
5.171839
110.777444
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A man with short, light-colored hair stands in a dimly lit room, wearing a light-colored robe tied at the waist. The camera focuses on the man, capturing his somber expression as he looks down and to his left. In the background, a bed with a dark headboard is visible, along with striped curtains and a small lamp on a bedside table. To the right of the man, the details of a paneled door with framed insets are visible. The scene is presented in black and white, contributing to a serious and introspective mood.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A man in a bathrobe stands in a bedroom with a serious expression, illuminated by a stark, high-contrast monochrome lighting. The scene conveys a sense of tension and contemplation within the confines of a traditional interior.
5.171839
1,280
540
124
1280x540
short_scene_001_short_Adam's_Rib_(1949)_-_A_Woman_Scorned_Scene_(4_10)_|_Movieclips
5.319023
{'frame_00:00.png': 5.230763912200928, 'frame_00:01.png': 5.329659461975098, 'frame_00:02.png': 5.211307048797607, 'frame_00:03.png': 5.479994773864746, 'frame_00:04.png': 5.343388557434082}
short_Adam's_Rib_(1949)_-_A_Woman_Scorned_Scene_(4_10)_|_Movieclips
../assets/split_scenes/short_Adam's_Rib_(1949)_-_A_Woman_Scorned_Scene_(4_10)_|_Movieclips/short_scene_001.mp4
0.599545
0.175468
{'frame_00:00.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Soft light, Low contrast, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Center'}], 'frame_00:01.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Soft light, Low contrast, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Center'}], 'frame_00:02.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Soft light, Low contrast, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Center'}], 'frame_00:03.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Soft light, Low contrast, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Center'}], 'frame_00:04.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Center'}]}
1
1
0
{"0": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "1": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "2": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "3": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "4": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}}
0.03171
[-1 1 1 -1 -1 1 1 -1 -1 1 1 1 1 1 1 -1 -1 -1 1 -1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=JxbvOwAB-xI
scene_001
1
112
232
00:00:04.671
00:00:09.676
4.671338
9.676343
5.005005
4.671338
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A woman is captured in a medium shot against a monochromatic backdrop, her gaze directed towards the right, conveying a sense of attentive listening. Her blonde hair is styled in elegant waves, complementing her high collared, dark colored shirt. The lighting accentuates her facial features, highlighting the subtle contours of her face and her expression that conveys her focus. Positioned slightly behind and to the left of her is a chair with distinct geometric patterns on its backrest, adding depth to the composition.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A woman with styled blonde hair and a collared shirt is the focus of a black and white film segment. Her expression shifts from surprised to serious, accompanied by subdued lighting and a simple backdrop.
5.005005
1,280
540
120
1280x540
short_scene_001_short_Adam's_Rib_(1949)_-_Contempt_for_the_Law_Scene_(6_10)_|_Movieclips
4.672602
{'frame_00:00.png': 4.78949499130249, 'frame_00:01.png': 4.4613356590271, 'frame_00:02.png': 4.723841667175293, 'frame_00:03.png': 4.7715349197387695, 'frame_00:04.png': 4.6168012619018555}
short_Adam's_Rib_(1949)_-_Contempt_for_the_Law_Scene_(6_10)_|_Movieclips
../assets/split_scenes/short_Adam's_Rib_(1949)_-_Contempt_for_the_Law_Scene_(6_10)_|_Movieclips/short_scene_001.mp4
-0.081019
1.117944
{'frame_00:00.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:01.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Center'}], 'frame_00:02.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:03.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:04.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Balanced'}]}
2.4
1.8
0
{"0": {"raw_counts": {"person": 1, "tie": 3, "total": 4}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 3, "total": 4}}, "1": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "2": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "3": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "4": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}}
-0.070824
[-1 -1 1 -1 -1 -1 1 -1 -1 -1 -1 -1 1 -1 1 -1 1 -1 -1 -1 -1 -1 -1 -1 -1 1]
https://youtube.com/watch?v=POu3JnhEmT4
scene_001
1
424
544
00:00:17.684
00:00:22.689
17.684351
22.689356
5.005005
17.684351
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A man in a dark suit and tie is seen entering a room, while a woman in a white shirt and black skirt stands in the doorway, observing him. The man proceeds to a closet where several dark-colored coats hang. He pauses, taking one of the coats off its hanger. The woman watches the man and the scene is presented in monochrome.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A man is seen entering a room in a black-and-white shot, soon followed by a woman, as they navigate toward a closet. The scene unfolds with a blend of formality and curiosity.
5.005005
1,280
540
120
1280x540
short_scene_001_short_Adam's_Rib_(1949)_-_Equality_Scene_(7_10)_|_Movieclips
5.130747
{'frame_00:00.png': 5.10405158996582, 'frame_00:01.png': 5.155367851257324, 'frame_00:02.png': 5.1325154304504395, 'frame_00:03.png': 5.170531272888184, 'frame_00:04.png': 5.0912675857543945}
short_Adam's_Rib_(1949)_-_Equality_Scene_(7_10)_|_Movieclips
../assets/split_scenes/short_Adam's_Rib_(1949)_-_Equality_Scene_(7_10)_|_Movieclips/short_scene_001.mp4
0.023896
0.032625
{'frame_00:00.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, High contrast, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Center'}], 'frame_00:01.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Center'}], 'frame_00:02.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Center'}], 'frame_00:03.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Center'}], 'frame_00:04.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Center'}]}
3.2
3.2
0
{"0": {"raw_counts": {"person": 3, "total": 3}, "bucketed_counts": {"humans": 3, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 3}}, "1": {"raw_counts": {"person": 3, "total": 3}, "bucketed_counts": {"humans": 3, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 3}}, "2": {"raw_counts": {"person": 3, "total": 3}, "bucketed_counts": {"humans": 3, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 3}}, "3": {"raw_counts": {"person": 4, "total": 4}, "bucketed_counts": {"humans": 4, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 4}}, "4": {"raw_counts": {"person": 3, "total": 3}, "bucketed_counts": {"humans": 3, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 3}}}
-0.004348
[-1 1 1 -1 -1 1 1 -1 -1 -1 -1 1 1 -1 1 -1 -1 -1 1 -1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=hdQOL2aFufE
scene_001
1
381
501
00:00:15.891
00:00:20.896
15.890891
20.895896
5.005005
15.890891
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A woman sits in a chair, looking directly at the camera with a serious expression, wearing a dark dress with a high collar and a decorative hat with netting, while a man sits behind her looking down, and another woman sits to the right with her chin in her hand and a hat on her head, all in a room with wood-paneled walls and a table. The scene is captured in black and white with dramatic lighting, enhancing the somber mood and focusing attention on the woman's intense gaze. The composition is carefully arranged to draw the viewer's eye to her, emphasizing her as the central figure in the scene.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A black and white clip showcases a woman adorned with an elaborate hat and veil, seated with a serious expression, while other individuals sit in the background. The somber atmosphere is enhanced by the classic monochrome palette.
5.005005
1,280
540
120
1280x540
short_scene_001_short_Adam's_Rib_(1949)_-_Mutual_Everything_(8_10)_|_Movieclips
5.202393
{'frame_00:00.png': 5.280555725097656, 'frame_00:01.png': 5.149528503417969, 'frame_00:02.png': 5.409953594207764, 'frame_00:03.png': 5.019542694091797, 'frame_00:04.png': 5.1523847579956055}
short_Adam's_Rib_(1949)_-_Mutual_Everything_(8_10)_|_Movieclips
../assets/split_scenes/short_Adam's_Rib_(1949)_-_Mutual_Everything_(8_10)_|_Movieclips/short_scene_001.mp4
-0.554382
0.015893
{'frame_00:00.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:01.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:02.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:03.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:04.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Balanced'}]}
2.6
2
0
{"0": {"raw_counts": {"person": 2, "tie": 1, "total": 3}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 3}}, "1": {"raw_counts": {"person": 2, "tie": 1, "total": 3}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 3}}, "2": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "3": {"raw_counts": {"person": 2, "tie": 1, "total": 3}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 3}}, "4": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}}
0.017523
[-1 1 1 -1 -1 1 1 -1 -1 1 1 1 1 -1 1 -1 -1 -1 1 -1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=8RGjds-aK00
scene_001
1
1,441
1,561
00:01:00.102
00:01:05.107
60.101768
65.106773
5.005005
60.101768
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A man and a woman are featured in a medium shot that is close-up, they are seemingly in deep conversation, with dramatic lighting that casts shadows across their faces. The man is positioned slightly behind and to the left of the woman, his face is turned towards her as if he is listening intently or trying to offer support. He has his arm draped over her shoulders in a comforting gesture. The woman, in the meantime, appears to be expressing a range of emotions, including distress, exasperation, and resolve, as she speaks or reacts. Her facial expressions shift and twist, suggesting a complex inner state. The background is indistinct, adding to the focus on the pair and their interaction.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A man has his arm around a woman's shoulder as she makes a series of facial expressions. The scene is shot in black and white.
5.005005
1,280
540
120
1280x540
short_scene_001_short_Adam's_Rib_(1949)_-_Vive_Le_Difference_Scene_(10_10)_|_Movieclips
4.552497
{'frame_00:00.png': 4.440340042114258, 'frame_00:01.png': 4.6604838371276855, 'frame_00:02.png': 4.51479959487915, 'frame_00:03.png': 4.8054375648498535, 'frame_00:04.png': 4.3414225578308105}
short_Adam's_Rib_(1949)_-_Vive_Le_Difference_Scene_(10_10)_|_Movieclips
../assets/split_scenes/short_Adam's_Rib_(1949)_-_Vive_Le_Difference_Scene_(10_10)_|_Movieclips/short_scene_001.mp4
0.335676
0.40248
{'frame_00:00.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:01.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Low contrast, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:02.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:03.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Center'}], 'frame_00:04.png': [{'<COLOR>': 'Desaturated, Black and White'}, {'<LIGHTING>': 'Hard light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}]}
0.8
0.8
0
{"0": {"raw_counts": {"total": 0}, "bucketed_counts": {"humans": 0, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 0}}, "1": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "2": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "3": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "4": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}}
-0.025301
[-1 1 1 -1 -1 -1 1 -1 -1 -1 -1 -1 1 -1 1 -1 1 -1 1 -1 -1 -1 1 -1 -1 1]
https://youtube.com/watch?v=Hl9OzCY9fVE
scene_001
1
142
262
00:00:05.923
00:00:10.928
5.922589
10.927594
5.005005
5.922589
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A person, wearing pajamas, walks into a room. The room appears dimly lit, illuminated primarily by light entering from the window. The room features a white door with rectangular panels, decorative plates hanging on the wall, and a piece of furniture with a dark, intricate design. The individual seems concerned or alert, looking around the room with a focused expression.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A man in pajamas emerges from a doorway, looking alert and concerned. He is illuminated by light from a window, creating a sense of mystery and anticipation.
5.005005
1,280
540
120
1280x540
short_scene_001_short_Adam_Driver_Shoots_Dinosaurs_Scene_-_65_(2023)_|_Movieclips
4.665559
{'frame_00:00.png': 4.024568557739258, 'frame_00:01.png': 4.449066162109375, 'frame_00:02.png': 4.794315338134766, 'frame_00:03.png': 5.201606273651123, 'frame_00:04.png': 4.858241081237793}
short_Adam_Driver_Shoots_Dinosaurs_Scene_-_65_(2023)_|_Movieclips
../assets/split_scenes/short_Adam_Driver_Shoots_Dinosaurs_Scene_-_65_(2023)_|_Movieclips/short_scene_001.mp4
-0.028252
0.142927
{'frame_00:00.png': [{'<COLOR>': 'Warm, Desaturated'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:01.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Soft light, Low contrast, Backlight'}, {'<LIGHTING_TYPE>': 'Daylight, Overcast'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:02.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Soft light, Low contrast, Backlight'}, {'<LIGHTING_TYPE>': 'Daylight, Overcast'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:03.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Soft light, Low contrast, Backlight'}, {'<LIGHTING_TYPE>': 'Daylight, Overcast'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:04.png': [{'<COLOR>': 'Desaturated, Green'}, {'<LIGHTING>': 'Soft light, Low contrast, Backlight'}, {'<LIGHTING_TYPE>': 'Daylight, Overcast'}, {'<COMPOSITION>': 'Center'}]}
0.8
0.8
0
{"0": {"raw_counts": {"total": 0}, "bucketed_counts": {"humans": 0, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 0}}, "1": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "2": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "3": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "4": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}}
-0.10669
[-1 -1 -1 -1 -1 -1 1 -1 -1 -1 -1 -1 -1 -1 -1 1 1 -1 -1 -1 -1 -1 -1 -1 -1 1]
https://youtube.com/watch?v=285GSGx1-7Q
scene_001
1
692
793
00:00:28.862
00:00:33.075
28.862196
33.074741
4.212546
28.862196
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
In a forest setting shrouded in mist, a figure with long dark hair and a light brown shirt is seen from behind, holding a futuristic weapon. The figure moves through the dense foliage, eventually raising the weapon and firing a shot, the muzzle flash briefly illuminating the scene. The camera focuses on the weapon as it discharges, highlighting its sleek, modern design against the backdrop of the natural environment.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A person is seen firing a weapon in a forest with smoke and trees visible in the background. The scene captures the intensity of a firefight in a dense environment.
4.212546
1,920
778
101
1920x778
short_scene_001_short_Addams_Family_Values_(1993)_-_Bachelor_Party_Scene_(4_10)_|_Movieclips
5.36109
{'frame_00:00.png': 5.486785888671875, 'frame_00:01.png': 5.313824653625488, 'frame_00:02.png': 5.282659530639648}
short_Addams_Family_Values_(1993)_-_Bachelor_Party_Scene_(4_10)_|_Movieclips
../assets/split_scenes/short_Addams_Family_Values_(1993)_-_Bachelor_Party_Scene_(4_10)_|_Movieclips/short_scene_001.mp4
-0.120472
0.32549
{'frame_00:00.png': [{'<COLOR>': 'Mixed, Saturated'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light'}, {'<COMPOSITION>': 'Center'}], 'frame_00:01.png': [{'<COLOR>': 'Mixed, Saturated'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light'}, {'<COMPOSITION>': 'Center'}], 'frame_00:02.png': [{'<COLOR>': 'Mixed, Saturated'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light'}, {'<COMPOSITION>': 'Center'}]}
1
1
0
{"0": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "1": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "2": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}}
-0.019743
[-1 1 1 1 -1 1 1 -1 -1 1 -1 1 1 -1 1 -1 1 -1 -1 -1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=D1IshjFWDJk
scene_001
1
315
374
00:00:13.138
00:00:15.599
13.138138
15.598932
2.460794
13.138138
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A woman with dark, curly hair and a white headband smiles gently while standing indoors with wall sconces behind her. She wears a mauve-colored dress with a light purple wrap, and a large silver necklace with a portrait pendant. The camera captures her from the chest up. As she opens her mouth wide, beginning to laugh, her eyes narrow slightly, and the shadows on her face deepen. She looks directly at the camera. A small bit of a white garment is visible in the lower right corner of the shot.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A woman with dark hair is speaking and smiling while standing indoors; the room is dimly lit by wall sconces. She is wearing a light pink dress, a lavender scarf, and a large, multi-strand silver necklace with a cameo pendant.
2.460794
1,920
810
59
1920x810
short_scene_001_short_Addams_Family_Values_(1993)_-_Bombing_Fester_Scene_(9_10)_|_Movieclips
5.18124
{'frame_00:00.png': 4.7143659591674805, 'frame_00:01.png': 5.3490519523620605, 'frame_00:02.png': 5.510220527648926, 'frame_00:03.png': 5.326087951660156, 'frame_00:04.png': 5.006475448608398}
short_Addams_Family_Values_(1993)_-_Bombing_Fester_Scene_(9_10)_|_Movieclips
../assets/split_scenes/short_Addams_Family_Values_(1993)_-_Bombing_Fester_Scene_(9_10)_|_Movieclips/short_scene_001.mp4
-0.253902
0.119554
{'frame_00:00.png': [{'<COLOR>': 'Cool, Desaturated, Blue'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Daylight, Overcast'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:01.png': [{'<COLOR>': 'Cool, Desaturated, Blue'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Daylight, Overcast'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:02.png': [{'<COLOR>': 'Cool, Desaturated, Blue'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Daylight, Overcast'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:03.png': [{'<COLOR>': 'Cool, Desaturated, Blue'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Daylight, Overcast'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:04.png': [{'<COLOR>': 'Warm, Saturated'}, {'<LIGHTING>': 'Hard light, High contrast, Backlight, Edge light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Left heavy'}]}
1
1
0
{"0": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "1": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "2": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "3": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "4": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}}
-0.022193
[-1 -1 1 1 -1 1 1 -1 -1 1 1 1 1 -1 1 -1 1 -1 -1 -1 -1 -1 -1 -1 -1 1]
https://youtube.com/watch?v=yprYw3FQUpQ
scene_001
1
287
407
00:00:11.970
00:00:16.975
11.970304
16.975309
5.005005
11.970304
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A woman with blonde hair, styled in a short bob, is seen in the driver's seat of a car, with a cream-colored scarf elegantly draped over her head and shoulders. She is wearing a cream-colored ribbed top with a gold chain detail around the neckline, and a watch is visible on her left wrist. She looks intently ahead, her expression shifting subtly. Behind her, through the car window, is a large, imposing house with multiple windows. Initially, the house is softly lit, but the scene dramatically changes as the house is engulfed in bright orange flames, reflecting the intense light across the scene. The fire casts a vivid glow on the woman's face, highlighting her composed yet contemplative expression as she watches the inferno.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A woman wearing a scarf is seen driving a car, with her house visible in the background; suddenly, flames erupt from the building in a dramatic turn of events. The woman appears stoic as the house burns in the background.
5.005005
1,920
810
120
1920x810
short_scene_001_short_Addams_Family_Values_(1993)_-_It's_an_Addams!_Scene_(1_10)_|_Movieclips
4.342345
{'frame_00:00.png': 4.403697967529297, 'frame_00:01.png': 4.440762996673584, 'frame_00:02.png': 4.182574272155762}
short_Addams_Family_Values_(1993)_-_It's_an_Addams!_Scene_(1_10)_|_Movieclips
../assets/split_scenes/short_Addams_Family_Values_(1993)_-_It's_an_Addams!_Scene_(1_10)_|_Movieclips/short_scene_001.mp4
0.047624
0.760373
{'frame_00:00.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Soft light, Low contrast, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Center'}], 'frame_00:01.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Soft light, Low contrast, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Center'}], 'frame_00:02.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Soft light, Low contrast, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Center'}]}
1
1
0
{"0": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "1": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "2": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}}
-0.006515
[-1 1 1 -1 -1 1 -1 -1 -1 -1 -1 1 1 -1 1 -1 -1 -1 1 -1 -1 -1 1 -1 -1 1]
https://youtube.com/watch?v=7p0J9wVGwZ0
scene_001
1
569
619
00:00:23.732
00:00:25.817
23.732065
25.817484
2.085419
23.732065
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A tight close-up shows two hands clasped together. One hand rests on top of the other, fingers intertwined, with subtle skin texture visible. The hand on the bottom exposes a crisp white shirt cuff underneath the dark suit jacket. The lighting is low-key, casting soft shadows and emphasizing the warmth of the skin tones against the dark, blurred background.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
Two hands interlock in what seems to be a handshake or a symbolic joining of forces. The lighting is subdued, casting shadows that emphasize the texture and form of the hands.
2.085419
1,920
810
50
1920x810
short_scene_001_short_Addams_Family_Values_(1993)_-_Shock_Value_Scene_(10_10)_|_Movieclips
5.371722
{'frame_00:00.png': 5.343033790588379, 'frame_00:01.png': 5.371079444885254, 'frame_00:02.png': 5.401052474975586}
short_Addams_Family_Values_(1993)_-_Shock_Value_Scene_(10_10)_|_Movieclips
../assets/split_scenes/short_Addams_Family_Values_(1993)_-_Shock_Value_Scene_(10_10)_|_Movieclips/short_scene_001.mp4
-0.03716
0.003145
{'frame_00:00.png': [{'<COLOR>': 'Cool, Desaturated'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Center'}], 'frame_00:01.png': [{'<COLOR>': 'Cool, Desaturated'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Moonlight, Artificial light'}, {'<COMPOSITION>': 'Center'}], 'frame_00:02.png': [{'<COLOR>': 'Cool, Desaturated'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Moonlight, Artificial light'}, {'<COMPOSITION>': 'Center'}]}
1.333333
1
0
{"0": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "1": {"raw_counts": {"person": 1, "tie": 1, "total": 2}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 2}}, "2": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}}
-0.026201
[-1 1 1 -1 -1 1 1 -1 -1 1 -1 1 1 -1 1 -1 -1 -1 -1 -1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=PmYdvwAqwDg
scene_001
1
403
465
00:00:16.808
00:00:19.394
16.808475
19.394394
2.585919
16.808475
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
The scene presents a man with a slender mustache and a somewhat theatrical expression, framed by a backdrop of dark, mechanical-looking equipment and shadowed corners; he wears a dark fez-like hat and a patterned robe with red trim over a patterned shirt, which suggests a vintage or theatrical costume, positioning him in what appears to be a dungeon-like setting with visible metallic elements and caution signs in the background, all under what seems to be dramatic lighting, emphasizing the somber and mysterious atmosphere of the location.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A man with slicked-back hair and a patterned smoking jacket stands in a dimly lit, seemingly enclosed space. The scene has a slightly unsettling yet comical atmosphere.
2.585919
1,920
810
62
1920x810
short_scene_001_short_Addams_Family_Values_(1993)_-_Thanksgiving_Play_Scene_(8_10)_|_Movieclips
5.188786
{'frame_00:00.png': 5.084604263305664, 'frame_00:01.png': 5.208200454711914, 'frame_00:02.png': 5.27355432510376}
short_Addams_Family_Values_(1993)_-_Thanksgiving_Play_Scene_(8_10)_|_Movieclips
../assets/split_scenes/short_Addams_Family_Values_(1993)_-_Thanksgiving_Play_Scene_(8_10)_|_Movieclips/short_scene_001.mp4
0.020515
0.356373
{'frame_00:00.png': [{'<COLOR>': 'Saturated, Orange, Cyan'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Daylight, Overcast'}, {'<COMPOSITION>': 'Center'}], 'frame_00:01.png': [{'<COLOR>': 'Saturated, Orange, White'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Daylight, Overcast'}, {'<COMPOSITION>': 'Center'}], 'frame_00:02.png': [{'<COLOR>': 'Saturated, Orange, White'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Daylight, Overcast'}, {'<COMPOSITION>': 'Center'}]}
1.666667
1
0
{"0": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "1": {"raw_counts": {"person": 1, "tie": 2, "total": 3}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 2, "total": 3}}, "2": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}}
0.020001
[-1 1 1 1 -1 1 1 -1 -1 1 1 1 1 -1 1 -1 -1 -1 1 -1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=v9UIDDlnSgA
scene_001
1
246
303
00:00:10.260
00:00:12.638
10.26026
12.637638
2.377377
10.26026
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A young fair-skinned girl with blonde hair stands in an outdoor setting, dressed in a historical costume featuring a white bonnet, a burgundy robe with white shoulder drapes, and a small bow at her neck. She also wears a pendant necklace. Her expression suggests seriousness and perhaps a hint of concern. To the left of the frame, a display with orange feathers is partially visible, adding a pop of color. The backdrop includes a glimpse of a wooden structure and greenery, hinting at a village or historical settlement. The lighting is even, casting soft shadows and highlighting the details of her costume and facial features.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A blonde-haired young person dressed in period costume expresses concern as they react to something in the scene. The subject wears a white cap, a red and white cloak, and a bow tie necklace.
2.377377
1,920
962
57
1920x962
short_scene_001_short_Addams_Family_Values_(1993)_-_The_Happy_Hut_Scene_(6_10)_|_Movieclips
5.380733
{'frame_00:00.png': 5.324366092681885, 'frame_00:01.png': 5.373470306396484, 'frame_00:02.png': 5.444361686706543}
short_Addams_Family_Values_(1993)_-_The_Happy_Hut_Scene_(6_10)_|_Movieclips
../assets/split_scenes/short_Addams_Family_Values_(1993)_-_The_Happy_Hut_Scene_(6_10)_|_Movieclips/short_scene_001.mp4
0.346307
0.048813
{'frame_00:00.png': [{'<COLOR>': 'Blue'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Balanced, Symmetrical'}], 'frame_00:01.png': [{'<COLOR>': 'Blue'}, {'<LIGHTING>': 'Hard light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Balanced, Symmetrical'}], 'frame_00:02.png': [{'<COLOR>': 'Blue'}, {'<LIGHTING>': 'Hard light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Balanced, Symmetrical'}]}
5
3.333333
0
{"0": {"raw_counts": {"person": 4, "chair": 1, "total": 5}, "bucketed_counts": {"humans": 4, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 5}}, "1": {"raw_counts": {"person": 3, "chair": 1, "total": 4}, "bucketed_counts": {"humans": 3, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 4}}, "2": {"raw_counts": {"person": 3, "chair": 2, "tv": 1, "total": 6}, "bucketed_counts": {"humans": 3, "animals": 0, "vehicles": 0, "food": 0, "others": 3, "total": 6}}}
0.010046
[-1 1 1 1 -1 1 1 -1 -1 1 -1 1 1 -1 1 -1 -1 -1 1 -1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=w80ZSg7kNbE
scene_001
1
101
156
00:00:04.213
00:00:06.507
4.212546
6.506507
2.293961
4.212546
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
Centered in the frame, a pale, dark-haired girl with braids and a solemn expression sits on a wicker couch between two boys. The boy on the left wears a striped shirt and the boy on the right wears a plaid shirt and glasses. All three figures face forward, eyes looking directly at the viewer with serious expressions. Behind them, the wall is decorated with various framed pictures and children's items, including a white unicorn poster, images of cats, and a shelf filled with stuffed animals. The overall lighting is bright, highlighting the details of the scene.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
Three individuals sit on a wicker couch with neutral expressions, positioned closely together against a backdrop filled with playful decorations and posters, creating a stark contrast with their somber demeanor. The setting appears to be a child's play area with an assortment of toys and cheerful wall decor.
2.293961
1,920
810
55
1920x810
short_scene_001_short_Addams_Family_Values_(1993)_-_Which_One_Will_Bounce?_Scene_(2_10)_|_Movieclips
4.55865
{'frame_00:00.png': 4.727413654327393, 'frame_00:01.png': 4.373783111572266, 'frame_00:02.png': 4.07943868637085, 'frame_00:03.png': 4.5478363037109375, 'frame_00:04.png': 4.772960186004639, 'frame_00:05.png': 4.744719982147217, 'frame_00:06.png': 4.664400100708008}
short_Addams_Family_Values_(1993)_-_Which_One_Will_Bounce?_Scene_(2_10)_|_Movieclips
../assets/split_scenes/short_Addams_Family_Values_(1993)_-_Which_One_Will_Bounce?_Scene_(2_10)_|_Movieclips/short_scene_001.mp4
0.198187
1.074174
{'frame_00:00.png': [{'<COLOR>': 'Cool, Desaturated'}, {'<LIGHTING>': 'Soft light'}, {'<LIGHTING_TYPE>': 'Moonlight, Artificial light'}, {'<COMPOSITION>': 'Balanced, Symmetrical'}], 'frame_00:01.png': [{'<COLOR>': 'Cool, Desaturated'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Daylight, Overcast'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:02.png': [{'<COLOR>': 'Cool, Desaturated, Blue'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Center'}], 'frame_00:03.png': [{'<COLOR>': 'Cool, Desaturated, Blue'}, {'<LIGHTING>': 'Hard light, High contrast, Side light'}, {'<LIGHTING_TYPE>': 'Moonlight'}, {'<COMPOSITION>': 'Balanced, Symmetrical'}], 'frame_00:04.png': [{'<COLOR>': 'Cool, Desaturated, Blue'}, {'<LIGHTING>': 'Hard light, High contrast, Side light'}, {'<LIGHTING_TYPE>': 'Moonlight'}, {'<COMPOSITION>': 'Balanced, Symmetrical'}], 'frame_00:05.png': [{'<COLOR>': 'Cool, Desaturated, Blue'}, {'<LIGHTING>': 'Hard light, High contrast, Side light'}, {'<LIGHTING_TYPE>': 'Moonlight'}, {'<COMPOSITION>': 'Balanced, Symmetrical'}], 'frame_00:06.png': [{'<COLOR>': 'Cool, Desaturated, Blue'}, {'<LIGHTING>': 'Hard light, High contrast, Side light'}, {'<LIGHTING_TYPE>': 'Moonlight'}, {'<COMPOSITION>': 'Balanced, Symmetrical'}]}
1.142857
1.142857
0
{"0": {"raw_counts": {"total": 0}, "bucketed_counts": {"humans": 0, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 0}}, "1": {"raw_counts": {"total": 0}, "bucketed_counts": {"humans": 0, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 0}}, "2": {"raw_counts": {"total": 0}, "bucketed_counts": {"humans": 0, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 0}}, "3": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "4": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "5": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "6": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}}
-0.066911
[-1 -1 1 -1 -1 -1 1 -1 -1 -1 -1 -1 -1 -1 1 1 1 -1 -1 -1 -1 -1 1 -1 -1 1]
https://youtube.com/watch?v=Ke-MYW3XORg
scene_001
1
102
261
00:00:04.254
00:00:10.886
4.254254
10.885886
6.631632
4.254254
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
The scene opens with a close-up view of two large, arched windows, their wooden shutters partially closed, giving a glimpse of the light from within, followed by a slow pan upwards along the exterior of a dark, wood-paneled building until two figures are revealed standing on the roof. One is a fair-haired person wearing a striped shirt and holding a dark ball, while the other, with braided hair, holds a baby wrapped in a light-colored blanket, under dim, nighttime lighting.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A person with braids stands on a roof while holding a baby next to another person in a striped shirt. They are seen from below the roof and the scene is dark.
6.631632
1,920
810
159
1920x810
short_scene_001_short_Addams_Family_Values_(1993)_-_Would_You_Die_for_Me?_Scene_(5_10)_|_Movieclips
5.343519
{'frame_00:00.png': 5.311827659606934, 'frame_00:01.png': 5.406716346740723, 'frame_00:02.png': 5.315691947937012, 'frame_00:03.png': 5.328753471374512, 'frame_00:04.png': 5.35063362121582, 'frame_00:05.png': 5.347491264343262}
short_Addams_Family_Values_(1993)_-_Would_You_Die_for_Me?_Scene_(5_10)_|_Movieclips
../assets/split_scenes/short_Addams_Family_Values_(1993)_-_Would_You_Die_for_Me?_Scene_(5_10)_|_Movieclips/short_scene_001.mp4
-0.025494
0.016881
{'frame_00:00.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:01.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Soft light, Low contrast, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:02.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Soft light, Low contrast, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:03.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Soft light, Low contrast, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:04.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:05.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Left heavy'}]}
1
0.666667
0.333333
{"0": {"raw_counts": {"dog": 1, "total": 1}, "bucketed_counts": {"humans": 0, "animals": 1, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "1": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "2": {"raw_counts": {"dog": 1, "total": 1}, "bucketed_counts": {"humans": 0, "animals": 1, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "3": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "4": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "5": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}}
-0.023722
[-1 1 1 1 -1 1 1 -1 -1 1 -1 1 1 -1 1 -1 -1 -1 -1 -1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=oK1zfJausVM
scene_001
1
46
175
00:00:01.919
00:00:07.299
1.918585
7.298966
5.38038
1.918585
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
In a luxurious bathroom, a bald man is submerged up to his neck in a tub full of white, fluffy bubbles. A gleaming gold railing surrounds the edge of the tub, with a glimpse of a small dish containing items in the background. The man's face is visible, featuring dark circles under his eyes and a somewhat unsettling smile. He looks directly forward, his expression unchanging.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A bald man with dark eye makeup is submerged in a bathtub full of bubbles, his head and shoulders visible above the foamy water, with a mischievous expression. The bathroom appears luxurious, with a golden railing around the tub and a glimpse of ornate decor in the background.
5.38038
1,920
908
129
1920x908
short_scene_001_short_Adventureland_(10_12)_Movie_CLIP_-_You_Wanna_End_This?_(2009)_HD
4.888511
{'frame_00:00.png': 4.6871843338012695, 'frame_00:01.png': 4.945954322814941, 'frame_00:02.png': 4.965178966522217, 'frame_00:03.png': 4.985257625579834, 'frame_00:04.png': 4.9654221534729, 'frame_00:05.png': 4.700114727020264, 'frame_00:06.png': 4.970465183258057}
short_Adventureland_(10_12)_Movie_CLIP_-_You_Wanna_End_This?_(2009)_HD
../assets/split_scenes/short_Adventureland_(10_12)_Movie_CLIP_-_You_Wanna_End_This?_(2009)_HD/short_scene_001.mp4
-0.092803
1.376335
{'frame_00:00.png': [{'<COLOR>': 'Warm, Desaturated'}, {'<LIGHTING>': 'Soft light, Top light, Side light, Edge light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:01.png': [{'<COLOR>': 'Warm, Desaturated'}, {'<LIGHTING>': 'Soft light, Top light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:02.png': [{'<COLOR>': 'Warm, Saturated'}, {'<LIGHTING>': 'Soft light, Side light, Edge light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:03.png': [{'<COLOR>': 'Warm, Saturated'}, {'<LIGHTING>': 'Soft light, Side light, Edge light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:04.png': [{'<COLOR>': 'Warm, Saturated'}, {'<LIGHTING>': 'Soft light, Side light, Edge light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:05.png': [{'<COLOR>': 'Warm, Saturated'}, {'<LIGHTING>': 'Soft light, Side light, Edge light'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:06.png': [{'<COLOR>': 'Warm, Saturated'}, {'<LIGHTING>': 'Soft light, Side light, Edge light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}]}
2.285714
2.285714
0
{"0": {"raw_counts": {"person": 3, "total": 3}, "bucketed_counts": {"humans": 3, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 3}}, "1": {"raw_counts": {"person": 3, "total": 3}, "bucketed_counts": {"humans": 3, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 3}}, "2": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "3": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "4": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "5": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "6": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}}
0.055161
[-1 1 1 1 1 1 1 -1 -1 1 1 1 1 -1 1 -1 -1 1 1 1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=TvTdzkWNq28
scene_001
1
166
330
00:00:06.924
00:00:13.764
6.92359
13.763764
6.840174
6.92359
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A man stands in close proximity to a woman in a dimly lit interior, gently placing his hands on her head. The woman, with a concerned expression, looks up at the man, her hands also raised towards her head. The man, positioned to her left, gazes back at her with a soft expression. The warm, indoor light casts a gentle glow on their faces, highlighting their interaction. As the woman lowers her hands, her expression shifts slightly, conveying a sense of vulnerability and apprehension. The man's steady gaze and gentle touch suggest a moment of comfort or concern between the two figures. The background is soft with wood paneling and a framed picture.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A man places his hands on a woman's head, looking at her with concern as she appears disoriented. The woman then looks up at the man, her expression still reflecting some confusion.
6.840174
1,280
540
164
1280x540
short_scene_001_short_Adventureland_(11_12)_Movie_CLIP_-_I'm_Sorry_(2009)_HD
4.770799
{'frame_00:00.png': 5.0789666175842285, 'frame_00:01.png': 4.992055416107178, 'frame_00:02.png': 4.756015300750732, 'frame_00:03.png': 4.5953288078308105, 'frame_00:04.png': 4.4316277503967285}
short_Adventureland_(11_12)_Movie_CLIP_-_I'm_Sorry_(2009)_HD
../assets/split_scenes/short_Adventureland_(11_12)_Movie_CLIP_-_I'm_Sorry_(2009)_HD/short_scene_001.mp4
0.243281
2.44839
{'frame_00:00.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Hard light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Center'}], 'frame_00:01.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Hard light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Center'}], 'frame_00:02.png': [{'<COLOR>': 'Mixed, Desaturated, Blue'}, {'<LIGHTING>': 'Hard light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Center'}], 'frame_00:03.png': [{'<COLOR>': 'Warm, Desaturated'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:04.png': [{'<COLOR>': 'Warm, Desaturated'}, {'<LIGHTING>': 'Soft light, Low contrast, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}]}
1
1
0
{"0": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "1": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "2": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "3": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "4": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}}
-0.002944
[-1 -1 1 1 -1 -1 1 -1 -1 1 -1 -1 1 -1 1 1 1 -1 1 -1 -1 -1 -1 -1 1 1]
https://youtube.com/watch?v=ZlGg7QIlGp8
scene_001
1
109
225
00:00:04.546
00:00:09.384
4.546213
9.384384
4.838172
4.546213
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A young man with curly hair, wearing a hooded sweatshirt and backpack, is illuminated with dramatic lighting as he stands between graffiti-covered surfaces. He appears alert, looking around warily before stepping out into the rain. The camera follows him as the rainfall intensifies, capturing the wetness on his face and clothes. The scene unfolds with a sense of tension and vulnerability as he ventures further into the downpour.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A young man stands near a graffitied wall, looking nervously around. The scene transitions to show him walking outside in the rain.
4.838172
1,280
540
116
1280x540
short_scene_001_short_Adventureland_(2_12)_Movie_CLIP_-_I_Can_Give_You_a_Ride_(2009)_HD
4.407413
{'frame_00:00.png': 4.60399866104126, 'frame_00:01.png': 4.387227535247803, 'frame_00:02.png': 4.340104103088379, 'frame_00:03.png': 4.008901596069336, 'frame_00:04.png': 4.696834564208984}
short_Adventureland_(2_12)_Movie_CLIP_-_I_Can_Give_You_a_Ride_(2009)_HD
../assets/split_scenes/short_Adventureland_(2_12)_Movie_CLIP_-_I_Can_Give_You_a_Ride_(2009)_HD/short_scene_001.mp4
-0.796811
1.885423
{'frame_00:00.png': [{'<COLOR>': 'Warm, Saturated'}, {'<LIGHTING>': 'Soft light, Side light, Edge light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Center'}], 'frame_00:01.png': [{'<COLOR>': 'Blue'}, {'<LIGHTING>': 'Soft light, Side light, Edge light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:02.png': [{'<COLOR>': 'Warm, Desaturated'}, {'<LIGHTING>': 'High contrast, Side light, Edge light'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light'}, {'<COMPOSITION>': 'Center'}], 'frame_00:03.png': [{'<COLOR>': 'Warm, Saturated'}, {'<LIGHTING>': 'Soft light, Side light, Edge light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:04.png': [{'<COLOR>': 'Warm, Saturated'}, {'<LIGHTING>': 'Soft light, Side light, Edge light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Left heavy'}]}
1.4
1
0.4
{"0": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "1": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "2": {"raw_counts": {"bird": 1, "person": 1, "total": 2}, "bucketed_counts": {"humans": 1, "animals": 1, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "3": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "4": {"raw_counts": {"dog": 1, "person": 1, "total": 2}, "bucketed_counts": {"humans": 1, "animals": 1, "vehicles": 0, "food": 0, "others": 0, "total": 2}}}
0.012416
[-1 -1 1 1 -1 1 1 -1 -1 1 -1 -1 1 -1 1 1 1 -1 1 -1 -1 -1 1 -1 -1 1]
https://youtube.com/watch?v=-nX_jQxWFN4
scene_001
1
31
151
00:00:01.293
00:00:06.298
1.29296
6.297965
5.005005
1.29296
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A person with shoulder-length dark hair and glasses looks down, then shifts their gaze slightly upward, while another person with short, curly brown hair looks off to the side with a soft expression. The focus alternates between the two individuals, capturing their subtle shifts in expression and gaze. The background remains softly blurred, emphasizing the subjects' presence and connection. The scene unfolds with a sense of intimacy and introspection, as the two figures share a quiet moment of contemplation.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A man with dark hair and glasses looks down intently, followed by a view of another man's face in soft lighting. Both are wearing t-shirts and appear to be in a casual setting.
5.005005
1,280
540
120
1280x540
short_scene_001_short_Adventureland_(3_12)_Movie_CLIP_-_Boner!_(2009)_HD
4.862386
{'frame_00:00.png': 4.9712724685668945, 'frame_00:01.png': 4.862250804901123, 'frame_00:02.png': 4.757541656494141, 'frame_00:03.png': 4.8584794998168945}
short_Adventureland_(3_12)_Movie_CLIP_-_Boner!_(2009)_HD
../assets/split_scenes/short_Adventureland_(3_12)_Movie_CLIP_-_Boner!_(2009)_HD/short_scene_001.mp4
-0.625864
0.036913
{'frame_00:00.png': [{'<COLOR>': 'Mixed, Saturated, Cyan, Blue'}, {'<LIGHTING>': 'Soft light'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:01.png': [{'<COLOR>': 'Mixed, Saturated, Cyan, Blue'}, {'<LIGHTING>': 'Hard light'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:02.png': [{'<COLOR>': 'Mixed, Saturated, Cyan, Blue'}, {'<LIGHTING>': 'Hard light'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:03.png': [{'<COLOR>': 'Mixed, Saturated, Cyan, Blue'}, {'<LIGHTING>': 'Hard light'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light'}, {'<COMPOSITION>': 'Balanced'}]}
4
2.5
1.5
{"0": {"raw_counts": {"cow": 1, "horse": 1, "person": 1, "total": 3}, "bucketed_counts": {"humans": 1, "animals": 2, "vehicles": 0, "food": 0, "others": 0, "total": 3}}, "1": {"raw_counts": {"cow": 1, "horse": 1, "person": 1, "total": 3}, "bucketed_counts": {"humans": 1, "animals": 2, "vehicles": 0, "food": 0, "others": 0, "total": 3}}, "2": {"raw_counts": {"cow": 1, "person": 5, "total": 6}, "bucketed_counts": {"humans": 5, "animals": 1, "vehicles": 0, "food": 0, "others": 0, "total": 6}}, "3": {"raw_counts": {"cow": 1, "person": 3, "total": 4}, "bucketed_counts": {"humans": 3, "animals": 1, "vehicles": 0, "food": 0, "others": 0, "total": 4}}}
-0.023911
[-1 1 1 1 -1 1 1 -1 1 1 -1 -1 1 -1 1 -1 -1 -1 -1 -1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=rMWZUV287WA
scene_001
1
81
162
00:00:03.378
00:00:06.757
3.378378
6.756757
3.378378
3.378378
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
At night, a man stands on the edge of a backyard swimming pool removing his shoes and socks. Another man is swimming in the pool, facing away from the first man. The pool is lit with underwater lights and bordered by a concrete deck with lounge chairs and a table. Behind the pool, the back of a brick house is visible, lit by several lamps, with a striped awning hanging over sliding glass doors leading to the house. Tall, dark trees are in the background.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A person standing by a pool takes off their pants while another person swims in the water. The scene is set at night by the pool of a house with lit windows and a dark backyard.
3.378378
1,280
594
81
1280x594
short_scene_001_short_Adventureland_(4_12)_Movie_CLIP_-_Lisa_P's_Back!_(2009)_HD
5.202399
{'frame_00:00.png': 5.119435787200928, 'frame_00:01.png': 5.220577239990234, 'frame_00:02.png': 5.2939453125, 'frame_00:03.png': 5.181309700012207, 'frame_00:04.png': 5.196728229522705}
short_Adventureland_(4_12)_Movie_CLIP_-_Lisa_P's_Back!_(2009)_HD
../assets/split_scenes/short_Adventureland_(4_12)_Movie_CLIP_-_Lisa_P's_Back!_(2009)_HD/short_scene_001.mp4
2.184332
0.102254
{'frame_00:00.png': [{'<COLOR>': 'Warm, Saturated, Red, Yellow, Cyan, Blue'}, {'<LIGHTING>': 'Soft light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:01.png': [{'<COLOR>': 'Warm, Saturated, Red, Yellow, Green'}, {'<LIGHTING>': 'Soft light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:02.png': [{'<COLOR>': 'Warm, Saturated, Red, Yellow, Green'}, {'<LIGHTING>': 'Soft light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:03.png': [{'<COLOR>': 'Warm, Saturated, Red, Yellow, Green'}, {'<LIGHTING>': 'Soft light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:04.png': [{'<COLOR>': 'Warm, Saturated, Red, Yellow, Green'}, {'<LIGHTING>': 'Soft light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light'}, {'<COMPOSITION>': 'Balanced'}]}
2.4
2.2
0
{"0": {"raw_counts": {"person": 3, "total": 3}, "bucketed_counts": {"humans": 3, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 3}}, "1": {"raw_counts": {"person": 2, "train": 1, "total": 3}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 1, "food": 0, "others": 0, "total": 3}}, "2": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "3": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "4": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}}
0.084934
[-1 1 1 1 1 1 1 -1 -1 1 1 1 1 1 1 -1 -1 1 1 1 -1 1 1 1 1 1]
https://youtube.com/watch?v=kGViaTOfSow
scene_001
1
342
462
00:00:14.264
00:00:19.269
14.264264
19.269269
5.005005
14.264264
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A frame presents two men casually leaning in what appears to be a vintage games area. The first man, on the left, has curly brown hair, wearing a faded blue tee with the word "Games" printed multiple times. He gazes off to the side with a slight smirk. Beside him, another man with long, dark hair and round glasses points outward with a serious expression. In the background, a colorful arcade game stands next to a weathered white pillar, and a sign that reads "No Rebounds" hangs above. The scene conveys a sense of relaxed tension in a dated environment.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
Two men are seated in a booth at a fair or carnival, possibly employed at one of the games. One man looks off to the side while the other stares forward with a slightly surprised expression.
5.005005
1,280
684
120
1280x684
short_scene_001_short_Adventureland_(6_12)_Movie_CLIP_-_Lisa_P__Asks_James_Out_(2009)_HD
4.408041
{'frame_00:00.png': 4.416218280792236, 'frame_00:01.png': 4.015522480010986, 'frame_00:02.png': 4.156385898590088, 'frame_00:03.png': 4.806604862213135, 'frame_00:04.png': 4.6454758644104}
short_Adventureland_(6_12)_Movie_CLIP_-_Lisa_P__Asks_James_Out_(2009)_HD
../assets/split_scenes/short_Adventureland_(6_12)_Movie_CLIP_-_Lisa_P__Asks_James_Out_(2009)_HD/short_scene_001.mp4
-1.384762
1.268164
{'frame_00:00.png': [{'<COLOR>': 'Warm, Saturated, Yellow, Green'}, {'<LIGHTING>': 'Hard light, Backlight'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Center'}], 'frame_00:01.png': [{'<COLOR>': 'Warm, Saturated, Yellow'}, {'<LIGHTING>': 'Soft light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:02.png': [{'<COLOR>': 'Warm, Saturated, Yellow, Green'}, {'<LIGHTING>': 'Soft light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:03.png': [{'<COLOR>': 'Warm, Saturated, Yellow'}, {'<LIGHTING>': 'Hard light, Backlight, Edge light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:04.png': [{'<COLOR>': 'Warm, Saturated, Red'}, {'<LIGHTING>': 'Hard light, Backlight'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Center'}]}
1.6
1.2
0.2
{"0": {"raw_counts": {"person": 2, "umbrella": 1, "total": 3}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 3}}, "1": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "2": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "3": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "4": {"raw_counts": {"person": 1, "dog": 1, "total": 2}, "bucketed_counts": {"humans": 1, "animals": 1, "vehicles": 0, "food": 0, "others": 0, "total": 2}}}
0.055204
[ 1 1 1 1 1 1 1 -1 1 1 -1 -1 1 -1 1 1 1 1 1 -1 -1 1 1 1 1 1]
https://youtube.com/watch?v=JXEpQzze8Wo
scene_001
1
240
360
00:00:10.010
00:00:15.015
10.01001
15.015015
5.005005
10.01001
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A medium shot reveals a young man with curly brown hair, wearing a blue t-shirt, standing near a sign with the words "Mateys Must Be This Tall" and an illustration of a parrot on it. The camera pans from the man to a young woman with curly blonde hair, wearing a pink shirt and red pants, as she walks up the steps toward a carnival ride. The camera follows her as she walks away from the first man. The scene is set outdoors, presumably at a carnival or amusement park, with various tents and rides visible in the background.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A young man stands near a height requirement sign, while in the background, a woman in a pink shirt and red pants walks up a set of steps. The scene takes place outside and is well-lit.
5.005005
1,280
594
120
1280x594
short_scene_001_short_Adventureland_(7_12)_Movie_CLIP_-_A_Date_with_Lisa_P__(2009)_HD
5.160431
{'frame_00:00.png': 5.287395477294922, 'frame_00:01.png': 5.190548896789551, 'frame_00:02.png': 5.068190574645996, 'frame_00:03.png': 5.1320295333862305, 'frame_00:04.png': 5.123988151550293}
short_Adventureland_(7_12)_Movie_CLIP_-_A_Date_with_Lisa_P__(2009)_HD
../assets/split_scenes/short_Adventureland_(7_12)_Movie_CLIP_-_A_Date_with_Lisa_P__(2009)_HD/short_scene_001.mp4
0.46748
1.617736
{'frame_00:00.png': [{'<COLOR>': 'Mixed, Saturated, Blue'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:01.png': [{'<COLOR>': 'Mixed, Saturated, Blue'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:02.png': [{'<COLOR>': 'Warm, Saturated, Blue'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:03.png': [{'<COLOR>': 'Warm, Saturated, Blue'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:04.png': [{'<COLOR>': 'Warm, Saturated, Blue'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}]}
2.8
1.8
0
{"0": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "1": {"raw_counts": {"person": 1, "tie": 1, "total": 2}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 2}}, "2": {"raw_counts": {"person": 2, "wine glass": 1, "total": 3}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 3}}, "3": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "4": {"raw_counts": {"person": 2, "bottle": 1, "chair": 1, "wine glass": 1, "total": 5}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 3, "total": 5}}}
0.079715
[ 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 -1 1 1 1 1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=kYlPy24WJzU
scene_001
1
111
227
00:00:04.630
00:00:09.468
4.62963
9.467801
4.838172
4.62963
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
The scene presents a woman with curly hair and striking makeup, including silver eyeshadow and bright red lipstick. She is wearing a blue jacket adorned with studs. Initially, she looks down as she speaks, then her gaze shifts slightly upward as she continues to talk. The scene then shows her picking up a glass of water, revealing long, red fingernails. She glances at the glass, then directly at the camera, a slight smile gracing her lips.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A woman with curly brown hair and bright red lipstick sits at a dark table wearing a turquoise top. She picks up a glass of water and appears to be listening to someone out of frame.
4.838172
1,280
540
116
1280x540
short_scene_001_short_Adventureland_(8_12)_Movie_CLIP_-_Tokin'_on_the_Job_(2009)_HD
4.510887
{'frame_00:00.png': 4.42807149887085, 'frame_00:01.png': 4.266289710998535, 'frame_00:02.png': 4.8383002281188965}
short_Adventureland_(8_12)_Movie_CLIP_-_Tokin'_on_the_Job_(2009)_HD
../assets/split_scenes/short_Adventureland_(8_12)_Movie_CLIP_-_Tokin'_on_the_Job_(2009)_HD/short_scene_001.mp4
0.130205
3.525476
{'frame_00:00.png': [{'<COLOR>': 'Warm, Saturated'}, {'<LIGHTING>': 'Soft light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:01.png': [{'<COLOR>': 'Yellow, Blue'}, {'<LIGHTING>': 'Soft light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:02.png': [{'<COLOR>': 'Warm, Saturated, Blue'}, {'<LIGHTING>': 'Soft light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Balanced'}]}
3
3
0
{"0": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "1": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "2": {"raw_counts": {"person": 5, "total": 5}, "bucketed_counts": {"humans": 5, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 5}}}
-0.039757
[-1 -1 1 1 -1 1 1 -1 -1 1 -1 -1 1 -1 1 -1 1 -1 -1 -1 -1 -1 -1 -1 -1 1]
https://youtube.com/watch?v=r1S-yBBZsDI
scene_001
1
598
657
00:00:24.942
00:00:27.402
24.941608
27.402402
2.460794
24.941608
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A group of people are standing outside near light-covered trees at night. A man with a mustache and dark hair is looking off to the side with a concerned expression. To his left is a partially obscured person in a checkered shirt, and in front of the man is a woman with dark hair wearing a blue cardigan. Behind the man and the woman are other people, and the lights from the trees are blurred in the background.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
At an outdoor party, a moustached man watches as others converse. The setting is illuminated with string lights, lending a festive atmosphere to the social gathering.
2.460794
1,280
540
59
1280x540
short_scene_001_short_Adventureland_(9_12)_Movie_CLIP_-_People_Are_Trying_to_Kill_Me_(2009)_HD
5.041488
{'frame_00:00.png': 5.229292869567871, 'frame_00:01.png': 4.782280921936035, 'frame_00:02.png': 5.112890243530273}
short_Adventureland_(9_12)_Movie_CLIP_-_People_Are_Trying_to_Kill_Me_(2009)_HD
../assets/split_scenes/short_Adventureland_(9_12)_Movie_CLIP_-_People_Are_Trying_to_Kill_Me_(2009)_HD/short_scene_001.mp4
1.329174
2.801802
{'frame_00:00.png': [{'<COLOR>': 'Warm, Saturated, Cyan, Blue'}, {'<LIGHTING>': 'Hard light, Top light, Edge light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:01.png': [{'<COLOR>': 'Warm, Saturated'}, {'<LIGHTING>': 'Soft light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:02.png': [{'<COLOR>': 'Warm, Saturated, Cyan'}, {'<LIGHTING>': 'Soft light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Balanced'}]}
2
2
0
{"0": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "1": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "2": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}}
0.010046
[-1 1 1 1 -1 1 1 -1 -1 1 -1 1 1 -1 1 -1 -1 -1 1 -1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=TEmvPX06cJo
scene_001
1
201
253
00:00:08.383
00:00:10.552
8.383383
10.552219
2.168836
8.383383
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A man with curly hair and wearing a hoodie sits closely next to a woman with shoulder-length hair as they both smile, with blurred lights in the background. The man's expression is cheerful as he looks to his left, while the woman beside him squints her eyes slightly as she smiles, as if she is having a good time. The background features bokeh-style lights, suggesting an out-of-focus environment, potentially at night. As the sequence progresses, their expressions vary from bright smiles to softer, more contemplative looks, yet they remain close, suggesting a warm and intimate connection between them.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
Two people are sitting next to each other and smiling, illuminated by string lights. The moment is warm and joyful, with both individuals appearing content.
2.168836
1,280
540
52
1280x540
short_scene_001_short_Aeon_Flux_(2_10)_Movie_CLIP_-_Welcome_to_Bregna_(2005)_HD
4.626518
{'frame_00:00.png': 4.721594333648682, 'frame_00:01.png': 4.897350311279297, 'frame_00:02.png': 4.57906436920166, 'frame_00:03.png': 4.422745704650879, 'frame_00:04.png': 4.511836528778076}
short_Aeon_Flux_(2_10)_Movie_CLIP_-_Welcome_to_Bregna_(2005)_HD
../assets/split_scenes/short_Aeon_Flux_(2_10)_Movie_CLIP_-_Welcome_to_Bregna_(2005)_HD/short_scene_001.mp4
0.1798
2.365826
{'frame_00:00.png': [{'<COLOR>': 'White'}, {'<LIGHTING>': 'Soft light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:01.png': [{'<COLOR>': 'White'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Daylight, Overcast'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:02.png': [{'<COLOR>': 'Red'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Daylight, Overcast'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:03.png': [{'<COLOR>': 'Warm, Saturated'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:04.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Daylight, Overcast'}, {'<COMPOSITION>': 'Right heavy'}]}
5.4
5.2
0
{"0": {"raw_counts": {"person": 9, "motorcycle": 1, "total": 10}, "bucketed_counts": {"humans": 9, "animals": 0, "vehicles": 1, "food": 0, "others": 0, "total": 10}}, "1": {"raw_counts": {"person": 11, "total": 11}, "bucketed_counts": {"humans": 11, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 11}}, "2": {"raw_counts": {"person": 3, "total": 3}, "bucketed_counts": {"humans": 3, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 3}}, "3": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "4": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}}
0.029508
[-1 1 1 1 1 1 1 -1 -1 1 -1 -1 1 -1 1 1 1 1 1 -1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=rwk6Obqrj9M
scene_001
1
226
337
00:00:09.426
00:00:14.056
9.426093
14.055722
4.62963
9.426093
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
The scene unfolds on a stark, monumental staircase, where a crowd of people ascends under a pale, overcast sky. A futuristic bicycle with large, solid disc wheels rests against the steps, momentarily obstructing the path. Two men, one in a dark jacket facing away from the camera and the other in a grey jacket with a distinct red collar, converge on the steps. The camera focuses in close-up, capturing their faces as they meet.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
Two men have a serious conversation on a gray, concrete staircase surrounded by other people in a bleak setting. One man wears a dark jacket while the other wears a red collar jacket.
4.62963
1,280
538
111
1280x538
short_scene_001_short_Aeon_Flux_(5_10)_Movie_CLIP_-_Why_Do_I_Feel_This_Way?_(2005)_HD
5.243999
{'frame_00:00.png': 5.47736930847168, 'frame_00:01.png': 5.198086738586426, 'frame_00:02.png': 5.08303689956665, 'frame_00:03.png': 5.217501163482666}
short_Aeon_Flux_(5_10)_Movie_CLIP_-_Why_Do_I_Feel_This_Way?_(2005)_HD
../assets/split_scenes/short_Aeon_Flux_(5_10)_Movie_CLIP_-_Why_Do_I_Feel_This_Way?_(2005)_HD/short_scene_001.mp4
0.886131
0.450519
{'frame_00:00.png': [{'<COLOR>': 'Warm, Desaturated'}, {'<LIGHTING>': 'Hard light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:01.png': [{'<COLOR>': 'Warm, Saturated'}, {'<LIGHTING>': 'Hard light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:02.png': [{'<COLOR>': 'Warm, Saturated'}, {'<LIGHTING>': 'Hard light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:03.png': [{'<COLOR>': 'Warm, Desaturated'}, {'<LIGHTING>': 'Hard light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Left heavy'}]}
2.25
2.25
0
{"0": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "1": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "2": {"raw_counts": {"person": 3, "total": 3}, "bucketed_counts": {"humans": 3, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 3}}, "3": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}}
0.036509
[-1 1 1 1 1 1 1 -1 -1 1 1 1 1 -1 1 -1 1 -1 1 -1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=TiHM7F5EUaI
scene_001
1
109
194
00:00:04.546
00:00:08.091
4.546213
8.091425
3.545212
4.546213
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
The scene captures two individuals in a tense interaction, set against a backdrop of muted, patterned walls. A man with light skin and wavy dark hair is dressed in a dark jacket with a high collar, his face etched with a mix of concern and vulnerability. His gaze is fixed on the other individual, whose back is to the camera. The person facing away has short, black, neatly styled hair and is wearing a dark outfit. The hand of this second individual is gripping the man's collar, creating a sense of dominance and control. The lighting is soft but focused on the man's face, highlighting his expression, with the rest of the scene fading slightly into shadow. The composition emphasizes the power dynamic and impending confrontation between the two characters.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A man with dark hair and a shocked expression stands facing a woman with black hair, who is out of focus and holding onto the man's shirt. The scene has dramatic lighting, emphasizing the tension and emotion between the two figures.
3.545212
1,280
540
85
1280x540
short_scene_001_short_After_Earth_(2013)_-_Baboon_Attack_Scene_(4_10)_|_Movieclips
5.269119
{'frame_00:00.png': 4.988982200622559, 'frame_00:01.png': 5.377795696258545, 'frame_00:02.png': 5.440579414367676}
short_After_Earth_(2013)_-_Baboon_Attack_Scene_(4_10)_|_Movieclips
../assets/split_scenes/short_After_Earth_(2013)_-_Baboon_Attack_Scene_(4_10)_|_Movieclips/short_scene_001.mp4
-0.044888
0.041766
{'frame_00:00.png': [{'<COLOR>': 'Cool, Desaturated, Green'}, {'<LIGHTING>': 'Soft light, Low contrast, Top light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:01.png': [{'<COLOR>': 'Cool, Desaturated, Green'}, {'<LIGHTING>': 'Soft light, Low contrast, Top light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:02.png': [{'<COLOR>': 'Cool, Desaturated, Green'}, {'<LIGHTING>': 'Soft light, Low contrast, Top light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Balanced'}]}
0.666667
0.666667
0
{"0": {"raw_counts": {"total": 0}, "bucketed_counts": {"humans": 0, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 0}}, "1": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "2": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}}
-0.009466
[-1 1 1 1 -1 1 1 1 1 1 -1 -1 1 -1 1 -1 1 -1 -1 -1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=t-lIuwPGT9w
scene_001
1
105
171
00:00:04.379
00:00:07.132
4.379379
7.132132
2.752753
4.379379
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
The camera gazes down from an elevated position into a verdant forest scene, where an individual clad in dark attire stands amidst dense vegetation. Towering trees, some fallen and covered in moss, define the landscape, while various species of ferns create a lush undergrowth. A fallen log stretches across the scene, partially illuminated by dappled sunlight filtering through the canopy. The person, positioned near the center, appears stationary, their presence a contrast to the surrounding wilderness. The scene evokes a sense of depth and tranquility.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
In a dense, green forest, a dark-clothed figure stands among ferns and fallen logs near a stream, surrounded by lush foliage and dappled sunlight. The scene conveys a sense of mystery and natural beauty with a hint of isolation.
2.752753
1,920
810
66
1920x810
short_scene_001_short_After_Earth_(2013)_-_Blood_Contamination_Scene_(5_10)_|_Movieclips
5.004654
{'frame_00:00.png': 4.819650173187256, 'frame_00:01.png': 4.717767238616943, 'frame_00:02.png': 5.223093509674072, 'frame_00:03.png': 5.029790878295898, 'frame_00:04.png': 5.116952896118164, 'frame_00:05.png': 5.12066650390625}
short_After_Earth_(2013)_-_Blood_Contamination_Scene_(5_10)_|_Movieclips
../assets/split_scenes/short_After_Earth_(2013)_-_Blood_Contamination_Scene_(5_10)_|_Movieclips/short_scene_001.mp4
-0.978096
1.098979
{'frame_00:00.png': [{'<COLOR>': 'Warm, Desaturated, Green'}, {'<LIGHTING>': 'Hard light, High contrast, Backlight'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Center'}], 'frame_00:01.png': [{'<COLOR>': 'Warm, Green'}, {'<LIGHTING>': 'Hard light, High contrast, Backlight'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Center'}], 'frame_00:02.png': [{'<COLOR>': 'Warm, Desaturated, Green'}, {'<LIGHTING>': 'Hard light, High contrast, Backlight'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Center'}], 'frame_00:03.png': [{'<COLOR>': 'Warm, Green'}, {'<LIGHTING>': 'Hard light, High contrast, Backlight'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Center'}], 'frame_00:04.png': [{'<COLOR>': 'Warm, Green'}, {'<LIGHTING>': 'Hard light, High contrast, Top light'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Center'}], 'frame_00:05.png': [{'<COLOR>': 'Warm, Green'}, {'<LIGHTING>': 'Hard light, High contrast, Backlight'}, {'<LIGHTING_TYPE>': 'Daylight, Sunny'}, {'<COMPOSITION>': 'Center'}]}
1.333333
0.833333
0.5
{"0": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "1": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "2": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "3": {"raw_counts": {"dog": 1, "elephant": 1, "total": 2}, "bucketed_counts": {"humans": 0, "animals": 2, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "4": {"raw_counts": {"elephant": 1, "person": 1, "total": 2}, "bucketed_counts": {"humans": 1, "animals": 1, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "5": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}}
-0.005837
[-1 1 1 1 -1 1 1 1 1 1 -1 1 1 -1 1 -1 -1 -1 -1 -1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=eMgfTq1Z2n8
scene_001
1
143
278
00:00:05.964
00:00:11.595
5.964298
11.594928
5.630631
5.964298
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
Deep within a forest of towering trees, a young person wearing a full-body light-colored textured garment crouches amongst the ferns and greenery of the forest floor. Sunlight filters through the dense canopy above, casting an ethereal glow on the scene, with dramatic lighting highlighting the individual's posture and environment. The person seems to be interacting with the vegetation, perhaps examining or gathering the ferns, suggesting a connection with the natural world. The focus remains tightly on the figure, emphasizing their solitude and the tranquil atmosphere of the forest.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A person dressed in a beige outfit is bending down, apparently looking at plants in a lush green forest. Towering trees surround the scene, with dramatic lighting creating a mystical atmosphere.
5.630631
1,920
810
135
1920x810
short_scene_001_short_After_Earth_(2013)_-_Fear_is_a_Choice_Scene_(6_10)_|_Movieclips
4.537273
{'frame_00:00.png': 4.44990348815918, 'frame_00:01.png': 4.463808536529541, 'frame_00:02.png': 4.478339672088623, 'frame_00:03.png': 4.406519889831543, 'frame_00:04.png': 4.58352518081665, 'frame_00:05.png': 4.841541290283203}
short_After_Earth_(2013)_-_Fear_is_a_Choice_Scene_(6_10)_|_Movieclips
../assets/split_scenes/short_After_Earth_(2013)_-_Fear_is_a_Choice_Scene_(6_10)_|_Movieclips/short_scene_001.mp4
0.007633
0.016303
{'frame_00:00.png': [{'<COLOR>': 'Cool, Desaturated'}, {'<LIGHTING>': 'Soft light, High contrast, Side light'}, {'<LIGHTING_TYPE>': 'Moonlight'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:01.png': [{'<COLOR>': 'Cool, Desaturated'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Moonlight'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:02.png': [{'<COLOR>': 'Cool, Desaturated'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Moonlight'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:03.png': [{'<COLOR>': 'Cool, Desaturated'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Moonlight'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:04.png': [{'<COLOR>': 'Cool, Desaturated'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Moonlight'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:05.png': [{'<COLOR>': 'Cool, Desaturated'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Moonlight'}, {'<COMPOSITION>': 'Right heavy'}]}
1
1
0
{"0": {"raw_counts": {"total": 0}, "bucketed_counts": {"humans": 0, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 0}}, "1": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "2": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "3": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "4": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "5": {"raw_counts": {"total": 0}, "bucketed_counts": {"humans": 0, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 0}}}
0.004091
[-1 1 1 -1 -1 1 -1 -1 -1 -1 1 1 1 -1 1 -1 -1 -1 1 -1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=SV8jbzudYJ8
scene_001
1
1,459
1,602
00:01:00.853
00:01:06.817
60.852519
66.816817
5.964298
60.852519
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A young person is captured in a close-up shot, their face partially illuminated, as they sit in a dimly lit outdoor setting; they appear contemplative or concerned, glancing slightly to the side, with a dark, rocky or wooded background. The lighting enhances the sense of solitude and introspection, creating a somber atmosphere. The focus remains tightly on the individual's face, capturing subtle expressions and the surrounding darkness contributing to the scene's overall mood.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A young person sits in darkness, appearing thoughtful or concerned. The scene is dimly lit, creating a somber and intimate atmosphere.
5.964298
1,920
810
143
1920x810
short_scene_001_short_After_Earth_(2013)_-_I'm_Not_a_Coward!_Scene_(7_10)_|_Movieclips
5.551267
{'frame_00:00.png': 5.475214004516602, 'frame_00:01.png': 5.449059009552002, 'frame_00:02.png': 5.85355281829834, 'frame_00:03.png': 5.418549060821533, 'frame_00:04.png': 5.559960842132568}
short_After_Earth_(2013)_-_I'm_Not_a_Coward!_Scene_(7_10)_|_Movieclips
../assets/split_scenes/short_After_Earth_(2013)_-_I'm_Not_a_Coward!_Scene_(7_10)_|_Movieclips/short_scene_001.mp4
1.038832
0.540564
{'frame_00:00.png': [{'<COLOR>': 'Cool, Desaturated, White'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Daylight, Overcast'}, {'<COMPOSITION>': 'Center'}], 'frame_00:01.png': [{'<COLOR>': 'Cool, Desaturated, Cyan'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Daylight, Overcast'}, {'<COMPOSITION>': 'Center, Symmetrical'}], 'frame_00:02.png': [{'<COLOR>': 'Cool, Desaturated, White'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Daylight, Overcast'}, {'<COMPOSITION>': 'Center, Symmetrical'}], 'frame_00:03.png': [{'<COLOR>': 'Cool, Desaturated, White'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Daylight, Overcast'}, {'<COMPOSITION>': 'Center, Symmetrical'}], 'frame_00:04.png': [{'<COLOR>': 'Cool, Desaturated, White'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Daylight, Overcast'}, {'<COMPOSITION>': 'Center, Symmetrical'}]}
1
1
0
{"0": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "1": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "2": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "3": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "4": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}}
0.000832
[ 1 1 1 1 -1 1 1 1 1 1 -1 1 1 -1 1 -1 1 -1 -1 -1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=4KICpbB5YQY
scene_001
1
67
179
00:00:02.794
00:00:07.466
2.794461
7.465799
4.671338
2.794461
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A person, silhouetted against a vast sky filled with dramatic clouds, stands atop a rocky, moss-covered precipice. Initially facing away from the viewer with a long object strapped to their back, they slowly turn to face forward. Dressed in a dark, form-fitting suit, the figure's posture is upright as they survey the landscape before them. In the distance, a low-lying land stretches out, suggesting a mountainous terrain fading into the horizon. The figure then begins to walk forward.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A lone figure stands atop a mossy cliff, outlined against a backdrop of a dramatic sky filled with heavy clouds. The person slowly turns and walks towards the camera.
4.671338
1,920
810
112
1920x810
short_scene_001_short_After_Earth_(2013)_-_The_Asteroid_Storm_Scene_(2_10)_|_Movieclips
4.853035
{'frame_00:00.png': 4.849587440490723, 'frame_00:01.png': 4.903903961181641, 'frame_00:02.png': 4.909785747528076, 'frame_00:03.png': 4.748864650726318}
short_After_Earth_(2013)_-_The_Asteroid_Storm_Scene_(2_10)_|_Movieclips
../assets/split_scenes/short_After_Earth_(2013)_-_The_Asteroid_Storm_Scene_(2_10)_|_Movieclips/short_scene_001.mp4
-0.15354
0.402532
{'frame_00:00.png': [{'<COLOR>': 'Cool, Desaturated'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light'}, {'<COMPOSITION>': 'Center'}], 'frame_00:01.png': [{'<COLOR>': 'Cool, Desaturated'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:02.png': [{'<COLOR>': 'Cool, Desaturated'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:03.png': [{'<COLOR>': 'Cool, Desaturated'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light'}, {'<COMPOSITION>': 'Left heavy'}]}
1.25
1.25
0
{"0": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "1": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "2": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "3": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}}
0.021503
[-1 1 1 -1 -1 1 1 -1 -1 1 1 1 1 -1 1 -1 1 -1 1 -1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=1Zwl6vfqjNQ
scene_001
1
65
141
00:00:02.711
00:00:05.881
2.711044
5.880881
3.169837
2.711044
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A man in a dark blue uniform stands in a dimly lit, metallic interior space. He appears to be looking around with a slightly concerned expression. The lighting is subdued, creating a sense of mystery and tension. The man is framed by the surrounding metal structures, with various lights and technical details visible in the background. The overall atmosphere is one of anticipation or uncertainty as the man seems to be assessing his surroundings.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A person in a dark uniform stands in a dimly lit, industrial-looking interior space, looking concerned or surprised. The setting is enclosed and somewhat claustrophobic, with metallic structures and dim lights visible in the background.
3.169837
1,920
810
76
1920x810
short_scene_001_short_After_Hours_(1985)_-_Art_Is_Forever_Scene_(9_9)_|_Movieclips
4.089624
{'frame_00:00.png': 3.9451241493225098, 'frame_00:01.png': 4.029567241668701, 'frame_00:02.png': 4.212866306304932, 'frame_00:03.png': 4.170936584472656}
short_After_Hours_(1985)_-_Art_Is_Forever_Scene_(9_9)_|_Movieclips
../assets/split_scenes/short_After_Hours_(1985)_-_Art_Is_Forever_Scene_(9_9)_|_Movieclips/short_scene_001.mp4
0.287645
0.914282
{'frame_00:00.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Hard light, High contrast, Backlight'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:01.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Hard light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:02.png': [{'<COLOR>': 'Cool, Desaturated'}, {'<LIGHTING>': 'Hard light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:03.png': [{'<COLOR>': 'Cool, Desaturated'}, {'<LIGHTING>': 'Hard light, High contrast, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light'}, {'<COMPOSITION>': 'Balanced'}]}
2.25
2
0
{"0": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "1": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "2": {"raw_counts": {"person": 2, "tie": 1, "total": 3}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 3}}, "3": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}}
-0.070824
[-1 -1 1 -1 -1 -1 1 -1 -1 -1 -1 -1 1 -1 1 -1 1 -1 -1 -1 -1 -1 -1 -1 -1 1]
https://youtube.com/watch?v=Hp93d2bsfQc
scene_001
1
439
534
00:00:18.310
00:00:22.272
18.309977
22.272272
3.962296
18.309977
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
Two men stand in a dimly lit room, one facing the camera and the other with his back to the camera and facing right; the man facing the camera wears a patterned jacket and a hat, and he has a mustache, while the man facing away from the camera wears a light-colored coat and holds a red jar in his right hand, looking down at it; a lamp hangs from the ceiling, casting a soft light on the scene; there are some plants on top of a cabinet, and the overall atmosphere is somewhat mysterious and serious.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
Two men converse in a dimly lit interior, one holding a jar and the other wearing a camo jacket and a hat. The scene is set with a lamp providing soft light, creating a casual, yet potentially tense atmosphere.
3.962296
1,920
810
95
1920x810
short_scene_001_short_After_Hours_(1985)_-_Just_Let_Me_In_Scene_(7_9)_|_Movieclips
4.679017
{'frame_00:00.png': 4.727360248565674, 'frame_00:01.png': 4.560733318328857, 'frame_00:02.png': 4.718967437744141, 'frame_00:03.png': 4.488397598266602, 'frame_00:04.png': 4.754596710205078, 'frame_00:05.png': 4.732428550720215, 'frame_00:06.png': 4.770631790161133}
short_After_Hours_(1985)_-_Just_Let_Me_In_Scene_(7_9)_|_Movieclips
../assets/split_scenes/short_After_Hours_(1985)_-_Just_Let_Me_In_Scene_(7_9)_|_Movieclips/short_scene_001.mp4
0.03761
0.325598
{'frame_00:00.png': [{'<COLOR>': 'Cool, Saturated, Purple, Magenta'}, {'<LIGHTING>': 'Hard light, High contrast, Backlight'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light, Mixed light'}, {'<COMPOSITION>': 'Center'}], 'frame_00:01.png': [{'<COLOR>': 'Mixed, Saturated, Purple, Magenta'}, {'<LIGHTING>': 'Hard light, High contrast, Backlight'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Center'}], 'frame_00:02.png': [{'<COLOR>': 'Cool, Saturated, Purple, Magenta'}, {'<LIGHTING>': 'Hard light, High contrast, Backlight'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:03.png': [{'<COLOR>': 'Cool, Saturated, Purple, Magenta'}, {'<LIGHTING>': 'Hard light, High contrast, Backlight'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:04.png': [{'<COLOR>': 'Cool, Saturated, Purple, Magenta'}, {'<LIGHTING>': 'Hard light, High contrast, Backlight'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light, Mixed light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:05.png': [{'<COLOR>': 'Cool, Saturated, Purple, Magenta'}, {'<LIGHTING>': 'Hard light, High contrast, Backlight'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:06.png': [{'<COLOR>': 'Cool, Saturated, Purple, Magenta'}, {'<LIGHTING>': 'Hard light, High contrast, Backlight'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light, Mixed light'}, {'<COMPOSITION>': 'Right heavy'}]}
2
2
0
{"0": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "1": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "2": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "3": {"raw_counts": {"person": 3, "total": 3}, "bucketed_counts": {"humans": 3, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 3}}, "4": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "5": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "6": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}}
-0.008617
[-1 1 1 1 -1 1 1 -1 -1 1 -1 1 1 -1 1 -1 1 -1 -1 -1 -1 1 1 -1 1 1]
https://youtube.com/watch?v=A-rs-kWL5-s
scene_001
1
37
193
00:00:01.543
00:00:08.050
1.54321
8.049716
6.506507
1.54321
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A man with arms crossed stands framed by a chain barrier and a painted "W" logo on the wall, under a purple circular light. The scene progresses as another person approaches him, stepping into the frame, illuminated against the darkness, with their back to the viewer, engaging the first man in conversation.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A person stands in front of a large "W" logo, arms crossed, looking serious as someone approaches them in a dimly lit room. The encounter is set against a backdrop of dark walls and a purple light.
6.506507
1,606
810
156
1606x810
short_scene_001_short_After_Hours_(1985)_-_Phone_Call_Scene_(8_9)_|_Movieclips
4.536286
{'frame_00:00.png': 4.5181121826171875, 'frame_00:01.png': 4.586216449737549, 'frame_00:02.png': 4.504528522491455}
short_After_Hours_(1985)_-_Phone_Call_Scene_(8_9)_|_Movieclips
../assets/split_scenes/short_After_Hours_(1985)_-_Phone_Call_Scene_(8_9)_|_Movieclips/short_scene_001.mp4
-0.009878
0.344623
{'frame_00:00.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Soft light, Low contrast, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:01.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Soft light, Low contrast, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:02.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Soft light, Low contrast, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}]}
1.333333
1
0.333333
{"0": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "1": {"raw_counts": {"person": 1, "dog": 1, "total": 2}, "bucketed_counts": {"humans": 1, "animals": 1, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "2": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}}
0.034188
[-1 1 1 1 -1 1 1 -1 -1 1 1 1 1 1 1 -1 -1 -1 1 -1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=MsKaN_QrVT8
scene_001
1
138
206
00:00:05.756
00:00:08.592
5.755756
8.591925
2.83617
5.755756
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A woman with blonde hair and dark eyeliner is captured in a medium shot as she stands partially obscured by a door frame. She is wearing a yellow shirt. She appears to be indoors, with a shelf visible behind her to the left. The lighting is soft, casting shadows that define the contours of her face and hair. Her expression shifts subtly, starting with a faint smile that fades to a more neutral, thoughtful look.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A woman with blonde hair looks through an open doorway with a knowing expression. The scene is bathed in natural light, giving it a realistic and slightly unsettling atmosphere.
2.83617
1,920
908
68
1920x908
short_scene_001_short_After_Hours_(1985)_-_You_Have_a_Great_Body_Scene_(2_9)_|_Movieclips
4.376064
{'frame_00:00.png': 4.379262924194336, 'frame_00:01.png': 4.306840896606445, 'frame_00:02.png': 4.60925817489624, 'frame_00:03.png': 4.5017619132995605, 'frame_00:04.png': 4.24881649017334, 'frame_00:05.png': 4.210444927215576}
short_After_Hours_(1985)_-_You_Have_a_Great_Body_Scene_(2_9)_|_Movieclips
../assets/split_scenes/short_After_Hours_(1985)_-_You_Have_a_Great_Body_Scene_(2_9)_|_Movieclips/short_scene_001.mp4
-1.229804
0.861427
{'frame_00:00.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Hard light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:01.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Hard light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:02.png': [{'<COLOR>': 'Desaturated'}, {'<LIGHTING>': 'Hard light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:03.png': [{'<COLOR>': 'Mixed, Saturated, Orange, Blue'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:04.png': [{'<COLOR>': 'Mixed, Saturated, Orange, Blue'}, {'<LIGHTING>': 'Soft light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:05.png': [{'<COLOR>': 'Mixed, Saturated, Orange, Blue'}, {'<LIGHTING>': 'Soft light, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light'}, {'<COMPOSITION>': 'Balanced'}]}
4.833333
3.166667
0
{"0": {"raw_counts": {"person": 1, "couch": 1, "total": 2}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 2}}, "1": {"raw_counts": {"person": 2, "couch": 1, "total": 3}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 3}}, "2": {"raw_counts": {"person": 1, "couch": 1, "total": 2}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 2}}, "3": {"raw_counts": {"person": 5, "tv": 1, "clock": 1, "bottle": 1, "total": 8}, "bucketed_counts": {"humans": 5, "animals": 0, "vehicles": 0, "food": 0, "others": 3, "total": 8}}, "4": {"raw_counts": {"person": 5, "tv": 1, "clock": 1, "total": 7}, "bucketed_counts": {"humans": 5, "animals": 0, "vehicles": 0, "food": 0, "others": 2, "total": 7}}, "5": {"raw_counts": {"person": 5, "tv": 1, "clock": 1, "total": 7}, "bucketed_counts": {"humans": 5, "animals": 0, "vehicles": 0, "food": 0, "others": 2, "total": 7}}}
0.040704
[ 1 1 1 1 -1 1 1 -1 1 1 -1 -1 1 -1 1 1 1 -1 1 -1 -1 1 1 1 1 1]
https://youtube.com/watch?v=zt4ek_5zQgY
scene_001
1
3,457
3,584
00:02:24.186
00:02:29.483
144.185853
149.482816
5.296964
144.185853
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
On a dark couch, a man leans over a woman lying down. The woman wears a black bra and patterned shorts, and she is covered in white powder. A black-and-white checkered pillow sits beside her, along with a white shirt and a red tie. The man, in a striped shirt, looks down at her with a serious expression.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A man leans over a woman lying on a couch, while objects like a checkered pillow and clothing are nearby. The scene has a dark and somewhat intimate atmosphere.
5.296964
1,920
810
127
1920x810
short_scene_001_short_Agent_Cody_Banks_(2_10)_Movie_CLIP_-_Spy_Gadgets_(2003)_HD
4.349549
{'frame_00:00.png': 4.417807102203369, 'frame_00:01.png': 4.3425774574279785, 'frame_00:02.png': 4.30987548828125, 'frame_00:03.png': 4.327935695648193}
short_Agent_Cody_Banks_(2_10)_Movie_CLIP_-_Spy_Gadgets_(2003)_HD
../assets/split_scenes/short_Agent_Cody_Banks_(2_10)_Movie_CLIP_-_Spy_Gadgets_(2003)_HD/short_scene_001.mp4
1.508562
5.565894
{'frame_00:00.png': [{'<COLOR>': 'Cool, Desaturated, Blue'}, {'<LIGHTING>': 'Hard light, High contrast, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light, Fluorescent'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:01.png': [{'<COLOR>': 'Blue'}, {'<LIGHTING>': 'Hard light, High contrast, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light, Fluorescent'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:02.png': [{'<COLOR>': 'Blue'}, {'<LIGHTING>': 'Hard light, High contrast, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light, Fluorescent'}, {'<COMPOSITION>': 'Right heavy'}], 'frame_00:03.png': [{'<COLOR>': 'Blue'}, {'<LIGHTING>': 'Hard light, High contrast, Top light'}, {'<LIGHTING_TYPE>': 'Artificial light, Practical light, Fluorescent'}, {'<COMPOSITION>': 'Balanced'}]}
4.25
4.25
0
{"0": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "1": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "2": {"raw_counts": {"person": 5, "total": 5}, "bucketed_counts": {"humans": 5, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 5}}, "3": {"raw_counts": {"person": 8, "total": 8}, "bucketed_counts": {"humans": 8, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 8}}}
-0.062232
[-1 -1 1 -1 -1 -1 1 -1 -1 -1 -1 -1 1 -1 1 1 1 -1 -1 -1 -1 -1 -1 -1 -1 1]
https://youtube.com/watch?v=XWXmnyflZtw
scene_001
1
36
126
00:00:01.502
00:00:05.255
1.501502
5.255255
3.753754
1.501502
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
Inside a bright, futuristic laboratory, a man in a white coat walks with papers in hand, surrounded by scientists and equipment. Several people are working at computer stations while another person is in a motion capture suit, jumping and gesturing. The lab is filled with monitors, metallic surfaces, and overhead lighting, creating a sterile and technologically advanced environment. Scientists are looking at data on screens, while others seem to be observing the motion capture performance, conveying a sense of focused activity and experimentation.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
The scene is in a futuristic lab where several scientists are working on experiments; one scientist in a white coat is walking and holding papers. A person wearing virtual reality gear is standing, while a woman in a lab coat is exercising on a machine.
3.753754
1,280
540
90
1280x540
short_scene_001_short_Agent_Cody_Banks_(3_10)_Movie_CLIP_-_The_New_Kid_(2003)_HD
4.829426
{'frame_00:00.png': 4.586249828338623, 'frame_00:01.png': 4.883735656738281, 'frame_00:02.png': 5.018291473388672}
short_Agent_Cody_Banks_(3_10)_Movie_CLIP_-_The_New_Kid_(2003)_HD
../assets/split_scenes/short_Agent_Cody_Banks_(3_10)_Movie_CLIP_-_The_New_Kid_(2003)_HD/short_scene_001.mp4
0.630707
0.784851
{'frame_00:00.png': [{'<COLOR>': 'Cool, Desaturated'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:01.png': [{'<COLOR>': 'Cool, Desaturated'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:02.png': [{'<COLOR>': 'Blue'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Left heavy'}]}
2
1.333333
0
{"0": {"raw_counts": {"person": 1, "tie": 1, "total": 2}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 2}}, "1": {"raw_counts": {"person": 1, "tie": 1, "total": 2}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 2}}, "2": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}}
0.069348
[-1 1 1 1 1 1 1 -1 -1 1 1 1 1 1 1 -1 -1 1 1 1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=pN7yX45G6WA
scene_001
1
134
195
00:00:05.589
00:00:08.133
5.588922
8.133133
2.544211
5.588922
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A young man dressed in a dark suit and tie, with light skin and blue eyes, is the central focus of the scene. He stands indoors, his expression shifting from surprise to a slight smile. Behind him, there's a glimpse of an office setting with shelves, files, and various items. To his right, a dark wooden door adds depth to the background. The lighting is bright, highlighting his features and the details of his suit, creating a clear and focused image.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A young man in a suit looks towards the camera with wide eyes and a slight smile. The scene takes place indoors near a door and office setting.
2.544211
1,280
526
61
1280x526
short_scene_001_short_Agent_Cody_Banks_(4_10)_Movie_CLIP_-_How_to_Talk_to_Girls_(2003)_HD
4.456937
{'frame_00:00.png': 4.2937541007995605, 'frame_00:01.png': 4.725071907043457, 'frame_00:02.png': 4.7305498123168945, 'frame_00:03.png': 4.444718360900879, 'frame_00:04.png': 4.360906600952148, 'frame_00:05.png': 4.431175231933594, 'frame_00:06.png': 4.212383270263672}
short_Agent_Cody_Banks_(4_10)_Movie_CLIP_-_How_to_Talk_to_Girls_(2003)_HD
../assets/split_scenes/short_Agent_Cody_Banks_(4_10)_Movie_CLIP_-_How_to_Talk_to_Girls_(2003)_HD/short_scene_001.mp4
0.037377
0.187995
{'frame_00:00.png': [{'<COLOR>': 'Green'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:01.png': [{'<COLOR>': 'Saturated, Green'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:02.png': [{'<COLOR>': 'Saturated, Green, Magenta'}, {'<LIGHTING>': 'Soft light, Low contrast'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Balanced'}], 'frame_00:03.png': [{'<COLOR>': 'Saturated, Green, Magenta'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:04.png': [{'<COLOR>': 'Mixed, Saturated, Green, Magenta'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:05.png': [{'<COLOR>': 'Saturated, Green, Magenta'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Left heavy'}], 'frame_00:06.png': [{'<COLOR>': 'Mixed, Saturated, Green, Magenta'}, {'<LIGHTING>': 'Soft light, Side light'}, {'<LIGHTING_TYPE>': 'Artificial light'}, {'<COMPOSITION>': 'Left heavy'}]}
1.857143
1.571429
0
{"0": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}, "1": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "2": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "3": {"raw_counts": {"person": 2, "total": 2}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 2}}, "4": {"raw_counts": {"person": 2, "tv": 1, "total": 3}, "bucketed_counts": {"humans": 2, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 3}}, "5": {"raw_counts": {"person": 1, "chair": 1, "total": 2}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 1, "total": 2}}, "6": {"raw_counts": {"person": 1, "total": 1}, "bucketed_counts": {"humans": 1, "animals": 0, "vehicles": 0, "food": 0, "others": 0, "total": 1}}}
-0.027352
[-1 1 1 1 -1 1 1 -1 -1 1 -1 -1 1 -1 1 -1 1 -1 -1 -1 -1 1 1 -1 -1 1]
https://youtube.com/watch?v=ENjz0-Ut-nQ
scene_001
1
29
174
00:00:01.210
00:00:07.257
1.209543
7.257257
6.047714
1.209543
You're a helpful assistant who is a part of a team working on video generation. As an assistant your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. Keeping that in mind, provide a caption for the input video. Please focus on the articulate details of the scenes presented in the video. Here are some guidelines: * Describe the composition of the scene, how it's progressing with all the components involved. * First describe the main subjects of the video and then how they are connected with one another. * DO NOT start the caption with "In this video,". * When appropriate for the scene, incorporate cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" IMPORTANT: Use these example phrases only when they naturally fit the scene you're describing. ## REQUIRED OUTPUT FORMAT: - Focus exclusively on what is visually happening in the frames - Describe actions, emotions, settings, and atmosphere - Maintain an objective, descriptive tone - Narrate as if describing the scene to someone who cannot see it - DO NOT mention character names, actor names, or film titles - Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A young person is sitting on a black bench while a person in a suit stands behind them. The person in the suit leans in with an angry expression and gestures with their hands. The young person looks up, appearing surprised and worried, and looks to their left with an open mouth. The setting appears to be in an airport or similar public space, with a display screen visible in the background. The scene is well-lit with modern architectural elements visible.
You're a helpful assistant who is a part of a team working on video generation. As an assistant, your task is to provide a meaningful caption for a provided video. These captions will be used to fine-tune other video generation models. **IMPORTANT: Your caption must be no more than one to two sentences.** Keep it short, high-level, and descriptive—focusing on the overall action, mood, or atmosphere in the scene. Here are some guidelines: * Identify the main subjects and describe the general visual activity or setting. * Avoid going into fine-grained or overly specific details—this is a summary, not a full narration. * Use cinematic phrases such as: * "with dramatic lighting" * "in slow motion" * "captured in close-up" only when they naturally fit the visual tone of the video. ## REQUIRED OUTPUT FORMAT: * Focus exclusively on what is visually happening in the frames * Describe actions, emotions, settings, and atmosphere * Maintain an objective, descriptive tone * **Limit your response to a maximum of two sentences** * DO NOT mention character names, actor names, or film titles * DO NOT start the caption with "In this video," * Write in a continuous, cohesive paragraph that reads like a natural description of the entire scene, rather than listing or breaking it down frame by frame
A young man sits in a modern lobby while an authority figure in a suit aggressively leans over him and yells. The young man is startled, as he looks up with a surprised expression.
6.047714
1,280
526
145
1280x526