jise commited on
Commit
2668a4c
·
verified ·
1 Parent(s): f7c1b03

Upload Qwen2_5_VLForConditionalGeneration

Browse files
config.json CHANGED
@@ -5,13 +5,9 @@
5
  "attention_dropout": 0.0,
6
  "bos_token_id": 151643,
7
  "eos_token_id": 151645,
8
- "vision_start_token_id": 151652,
9
- "vision_end_token_id": 151653,
10
- "vision_token_id": 151654,
11
- "image_token_id": 151655,
12
- "video_token_id": 151656,
13
  "hidden_act": "silu",
14
  "hidden_size": 3584,
 
15
  "initializer_range": 0.02,
16
  "intermediate_size": 18944,
17
  "max_position_embeddings": 128000,
@@ -20,42 +16,65 @@
20
  "num_attention_heads": 28,
21
  "num_hidden_layers": 28,
22
  "num_key_value_heads": 4,
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
23
  "rms_norm_eps": 1e-06,
 
 
 
 
 
 
 
 
 
24
  "rope_theta": 1000000.0,
25
  "sliding_window": 32768,
26
  "tie_word_embeddings": false,
27
- "torch_dtype": "bfloat16",
28
- "transformers_version": "4.41.2",
29
  "use_cache": true,
30
  "use_sliding_window": false,
 
31
  "vision_config": {
32
  "depth": 32,
 
 
 
 
 
 
33
  "hidden_act": "silu",
34
  "hidden_size": 1280,
 
 
35
  "intermediate_size": 3420,
 
36
  "num_heads": 16,
37
- "in_chans": 3,
38
  "out_hidden_size": 3584,
39
  "patch_size": 14,
40
  "spatial_merge_size": 2,
41
  "spatial_patch_size": 14,
42
- "window_size": 112,
43
- "fullatt_block_indexes": [
44
- 7,
45
- 15,
46
- 23,
47
- 31
48
- ],
49
  "tokens_per_second": 2,
50
- "temporal_patch_size": 2
51
- },
52
- "rope_scaling": {
53
- "type": "mrope",
54
- "mrope_section": [
55
- 16,
56
- 24,
57
- 24
58
- ]
59
  },
 
 
 
60
  "vocab_size": 152064
61
- }
 
5
  "attention_dropout": 0.0,
6
  "bos_token_id": 151643,
7
  "eos_token_id": 151645,
 
 
 
 
 
8
  "hidden_act": "silu",
9
  "hidden_size": 3584,
10
+ "image_token_id": 151655,
11
  "initializer_range": 0.02,
12
  "intermediate_size": 18944,
13
  "max_position_embeddings": 128000,
 
16
  "num_attention_heads": 28,
17
  "num_hidden_layers": 28,
18
  "num_key_value_heads": 4,
19
+ "quantization_config": {
20
+ "_load_in_4bit": true,
21
+ "_load_in_8bit": false,
22
+ "bnb_4bit_compute_dtype": "bfloat16",
23
+ "bnb_4bit_quant_storage": "uint8",
24
+ "bnb_4bit_quant_type": "nf4",
25
+ "bnb_4bit_use_double_quant": true,
26
+ "llm_int8_enable_fp32_cpu_offload": false,
27
+ "llm_int8_has_fp16_weight": false,
28
+ "llm_int8_skip_modules": null,
29
+ "llm_int8_threshold": 6.0,
30
+ "load_in_4bit": true,
31
+ "load_in_8bit": false,
32
+ "quant_method": "bitsandbytes"
33
+ },
34
  "rms_norm_eps": 1e-06,
35
+ "rope_scaling": {
36
+ "mrope_section": [
37
+ 16,
38
+ 24,
39
+ 24
40
+ ],
41
+ "rope_type": "default",
42
+ "type": "default"
43
+ },
44
  "rope_theta": 1000000.0,
45
  "sliding_window": 32768,
46
  "tie_word_embeddings": false,
47
+ "torch_dtype": "float32",
48
+ "transformers_version": "4.51.3",
49
  "use_cache": true,
50
  "use_sliding_window": false,
51
+ "video_token_id": 151656,
52
  "vision_config": {
53
  "depth": 32,
54
+ "fullatt_block_indexes": [
55
+ 7,
56
+ 15,
57
+ 23,
58
+ 31
59
+ ],
60
  "hidden_act": "silu",
61
  "hidden_size": 1280,
62
+ "in_channels": 3,
63
+ "in_chans": 3,
64
  "intermediate_size": 3420,
65
+ "model_type": "qwen2_5_vl",
66
  "num_heads": 16,
 
67
  "out_hidden_size": 3584,
68
  "patch_size": 14,
69
  "spatial_merge_size": 2,
70
  "spatial_patch_size": 14,
71
+ "temporal_patch_size": 2,
 
 
 
 
 
 
72
  "tokens_per_second": 2,
73
+ "torch_dtype": "bfloat16",
74
+ "window_size": 112
 
 
 
 
 
 
 
75
  },
76
+ "vision_end_token_id": 151653,
77
+ "vision_start_token_id": 151652,
78
+ "vision_token_id": 151654,
79
  "vocab_size": 152064
80
+ }
generation_config.json ADDED
@@ -0,0 +1,12 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token_id": 151643,
3
+ "do_sample": true,
4
+ "eos_token_id": [
5
+ 151645,
6
+ 151643
7
+ ],
8
+ "pad_token_id": 151643,
9
+ "repetition_penalty": 1.05,
10
+ "temperature": 1e-06,
11
+ "transformers_version": "4.51.3"
12
+ }
model-00001-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:420f332c55a85ef8943fb395eacc2e9e4afa0dc0fd5d219213e027825945c968
3
+ size 4814236822
model-00002-of-00002.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0c859795ad3a627a9b95bcb762e059d5b768a4a36fdd4affeff269d93fdecc67
3
+ size 1089994880
model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff