kingsley01 commited on
Commit
8ec0711
·
verified ·
1 Parent(s): cc86997

Upload config

Browse files
Files changed (1) hide show
  1. config.json +70 -0
config.json ADDED
@@ -0,0 +1,70 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "InternVLForConditionalGeneration"
4
+ ],
5
+ "downsample_ratio": 0.5,
6
+ "image_seq_length": 256,
7
+ "image_token_index": 151667,
8
+ "model_type": "internvl",
9
+ "projector_hidden_act": "gelu",
10
+ "text_config": {
11
+ "architectures": [
12
+ "Qwen2ForCausalLM"
13
+ ],
14
+ "attention_dropout": 0.0,
15
+ "bos_token_id": 151643,
16
+ "eos_token_id": 151645,
17
+ "hidden_act": "silu",
18
+ "hidden_size": 896,
19
+ "initializer_range": 0.02,
20
+ "intermediate_size": 4864,
21
+ "max_position_embeddings": 32768,
22
+ "max_window_layers": 21,
23
+ "model_type": "qwen2",
24
+ "num_attention_heads": 14,
25
+ "num_hidden_layers": 24,
26
+ "num_key_value_heads": 2,
27
+ "rms_norm_eps": 1e-06,
28
+ "rope_scaling": null,
29
+ "rope_theta": 1000000.0,
30
+ "sliding_window": 32768,
31
+ "torch_dtype": "bfloat16",
32
+ "use_cache": true,
33
+ "use_sliding_window": false,
34
+ "vocab_size": 151674
35
+ },
36
+ "transformers_version": "4.52.0.dev0",
37
+ "vision_config": {
38
+ "architectures": [
39
+ "InternVisionModel"
40
+ ],
41
+ "attention_dropout": 0.0,
42
+ "attention_probs_dropout_prob": 0.0,
43
+ "dropout": 0.0,
44
+ "hidden_act": "gelu",
45
+ "hidden_dropout_prob": 0.0,
46
+ "hidden_size": 1024,
47
+ "image_size": [
48
+ 448,
49
+ 448
50
+ ],
51
+ "initializer_factor": 1.0,
52
+ "initializer_range": 0.02,
53
+ "intermediate_size": 4096,
54
+ "layer_norm_eps": 1e-06,
55
+ "layer_scale_init_value": 0.1,
56
+ "model_type": "internvl_vision",
57
+ "num_attention_heads": 16,
58
+ "num_channels": 3,
59
+ "num_hidden_layers": 24,
60
+ "patch_size": [
61
+ 14,
62
+ 14
63
+ ],
64
+ "torch_dtype": "bfloat16",
65
+ "use_absolute_position_embeddings": true,
66
+ "use_mask_token": false,
67
+ "use_mean_pooling": true
68
+ },
69
+ "vision_feature_layer": -1
70
+ }