Karroyan commited on
Commit
4639414
·
verified ·
1 Parent(s): 929f69b

Upload config.json with huggingface_hub

Browse files
Files changed (1) hide show
  1. config.json +73 -0
config.json ADDED
@@ -0,0 +1,73 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "KeyeForConditionalGeneration"
4
+ ],
5
+ "attention_bias": false,
6
+ "attention_dropout": 0.0,
7
+ "auto_map": {
8
+ "AutoConfig": "configuration_keye.KeyeConfig",
9
+ "AutoModel": "modeling_keye.KeyeForConditionalGeneration",
10
+ "AutoModelForCausalLM": "modeling_keye.KeyeForConditionalGeneration"
11
+ },
12
+ "bos_token_id": 151643,
13
+ "eos_token_id": 151645,
14
+ "head_dim": 128,
15
+ "hidden_act": "silu",
16
+ "hidden_size": 4096,
17
+ "image_token_id": 151655,
18
+ "initializer_factor": 1.0,
19
+ "initializer_range": 0.02,
20
+ "intermediate_size": 12288,
21
+ "max_position_embeddings": 40960,
22
+ "max_window_layers": 36,
23
+ "model_type": "Keye",
24
+ "num_attention_heads": 32,
25
+ "num_hidden_layers": 36,
26
+ "num_key_value_heads": 8,
27
+ "rms_norm_eps": 1e-06,
28
+ "rope_scaling": {
29
+ "mrope_section": [
30
+ 16,
31
+ 24,
32
+ 24
33
+ ],
34
+ "rope_type": "default",
35
+ "type": "default"
36
+ },
37
+ "rope_theta": 1000000,
38
+ "sliding_window": null,
39
+ "tie_word_embeddings": false,
40
+ "torch_dtype": "bfloat16",
41
+ "transformers_version": "4.52.1",
42
+ "use_cache": true,
43
+ "use_sliding_window": false,
44
+ "video_token_id": 151656,
45
+ "vision_config": {
46
+ "architectures": [
47
+ "SiglipVisionModel"
48
+ ],
49
+ "attention_dropout": 0.0,
50
+ "auto_map": {
51
+ "AutoConfig": "configuration_keye.KeyeVisionConfig",
52
+ "AutoModel": "modeling_keye.SiglipVisionModel"
53
+ },
54
+ "hidden_act": "gelu_pytorch_tanh",
55
+ "hidden_size": 1152,
56
+ "image_size": 384,
57
+ "intermediate_size": 4304,
58
+ "layer_norm_eps": 1e-06,
59
+ "model_type": "Keye",
60
+ "num_attention_heads": 16,
61
+ "num_channels": 3,
62
+ "num_hidden_layers": 27,
63
+ "patch_size": 14,
64
+ "spatial_merge_size": 2,
65
+ "temporal_patch_size": 2,
66
+ "tokens_per_second": 2,
67
+ "torch_dtype": "bfloat16"
68
+ },
69
+ "vision_end_token_id": 151653,
70
+ "vision_start_token_id": 151652,
71
+ "vision_token_id": 151654,
72
+ "vocab_size": 151936
73
+ }