The dataset is currently empty. Upload or create new data files. Then, you will be able to explore them in the Dataset Viewer.
Description
100,000 Sets - Video Dialogue Dataset, which includes 100000 videos of multiple themes, with corresponding dialogue text and audio. The dialogue audio simulates human-machine conversations based on video content. This dataset can be used for tasks such as video realtime understanding and embodied AI.
For more details, please refer to the link: https://www.nexdata.ai/datasets/computervision?source=Huggingface
Specifications
Data content：
100,000 sets of real-time video dialogue data, simulating human-machine conversations based on video content. Each set includes: ①video files (mp4/avi/mov); ②dialogue transcripts in Chinese (json); ③dialogue audio in Chinese(wav); ④dialogue transcripts in English(json)
Data diversity：
covers ①video themes(people, plant, animal, food, item, etc); ②dialogue topics(factual QAs, extended suggestions, etc); ③voices(1,123 distinct questioner voices, 80 distinct responder voices)
Data features：
incorporates interruptions in selected Chinese dialogue transcripts and audio to better simulate real-world application scenarios
Data quality：
①video resolution >= 1080p; ②dialogue turns >= 3 rounds per conversation; ③audio’s sampling rate is 24kHz and depth is 16-bit
Licensing Information
Commercial License
- Downloads last month
- 1