[email protected] commited on
Commit
1da3a4b
·
1 Parent(s): dbdf6e3

Change config.json

Browse files
Files changed (1) hide show
  1. config.json +25 -69
config.json CHANGED
@@ -5,9 +5,13 @@
5
  "attention_dropout": 0.0,
6
  "bos_token_id": 151643,
7
  "eos_token_id": 151645,
 
 
 
 
 
8
  "hidden_act": "silu",
9
  "hidden_size": 2048,
10
- "image_token_id": 151655,
11
  "initializer_range": 0.02,
12
  "intermediate_size": 11008,
13
  "max_position_embeddings": 128000,
@@ -17,89 +21,41 @@
17
  "num_hidden_layers": 36,
18
  "num_key_value_heads": 2,
19
  "rms_norm_eps": 1e-06,
20
- "rope_scaling": {
21
- "mrope_section": [
22
- 16,
23
- 24,
24
- 24
25
- ],
26
- "rope_type": "default",
27
- "type": "default"
28
- },
29
  "rope_theta": 1000000.0,
30
  "sliding_window": 32768,
31
- "text_config": {
32
- "architectures": [
33
- "Qwen2_5_VLForConditionalGeneration"
34
- ],
35
- "attention_dropout": 0.0,
36
- "bos_token_id": 151643,
37
- "eos_token_id": 151645,
38
- "hidden_act": "silu",
39
- "hidden_size": 2048,
40
- "image_token_id": null,
41
- "initializer_range": 0.02,
42
- "intermediate_size": 11008,
43
- "max_position_embeddings": 128000,
44
- "max_window_layers": 70,
45
- "model_type": "qwen2_5_vl_text",
46
- "num_attention_heads": 16,
47
- "num_hidden_layers": 36,
48
- "num_key_value_heads": 2,
49
- "rms_norm_eps": 1e-06,
50
- "rope_scaling": {
51
- "mrope_section": [
52
- 16,
53
- 24,
54
- 24
55
- ],
56
- "rope_type": "default",
57
- "type": "default"
58
- },
59
- "rope_theta": 1000000.0,
60
- "sliding_window": 32768,
61
- "tie_word_embeddings": true,
62
- "torch_dtype": "bfloat16",
63
- "use_cache": true,
64
- "use_sliding_window": false,
65
- "video_token_id": null,
66
- "vision_end_token_id": 151653,
67
- "vision_start_token_id": 151652,
68
- "vision_token_id": 151654,
69
- "vocab_size": 151936
70
- },
71
  "torch_dtype": "bfloat16",
72
- "transformers_version": "4.52.3",
73
  "use_cache": true,
74
  "use_sliding_window": false,
75
- "video_token_id": 151656,
76
  "vision_config": {
77
  "depth": 32,
78
- "fullatt_block_indexes": [
79
- 7,
80
- 15,
81
- 23,
82
- 31
83
- ],
84
  "hidden_act": "silu",
85
  "hidden_size": 1280,
86
- "in_channels": 3,
87
- "in_chans": 3,
88
- "initializer_range": 0.02,
89
  "intermediate_size": 3420,
90
- "model_type": "qwen2_5_vl",
91
  "num_heads": 16,
 
92
  "out_hidden_size": 2048,
93
  "patch_size": 14,
94
  "spatial_merge_size": 2,
95
  "spatial_patch_size": 14,
96
- "temporal_patch_size": 2,
 
 
 
 
 
 
97
  "tokens_per_second": 2,
98
- "torch_dtype": "bfloat16",
99
- "window_size": 112
 
 
 
 
 
 
 
100
  },
101
- "vision_end_token_id": 151653,
102
- "vision_start_token_id": 151652,
103
- "vision_token_id": 151654,
104
  "vocab_size": 151936
105
- }
 
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": 2048,
 
15
  "initializer_range": 0.02,
16
  "intermediate_size": 11008,
17
  "max_position_embeddings": 128000,
 
21
  "num_hidden_layers": 36,
22
  "num_key_value_heads": 2,
23
  "rms_norm_eps": 1e-06,
 
 
 
 
 
 
 
 
 
24
  "rope_theta": 1000000.0,
25
  "sliding_window": 32768,
26
+ "tie_word_embeddings": true,
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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": 2048,
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": 151936
61
+ }