nalazhar commited on
Commit
35269fd
·
verified ·
1 Parent(s): 599881d

Add files using upload-large-folder tool

Browse files
Files changed (1) hide show
  1. config.json +84 -0
config.json ADDED
@@ -0,0 +1,84 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "UmansCoderForConditionalGeneration"
4
+ ],
5
+ "model_type": "deepseek_v32",
6
+ "attention_bias": false,
7
+ "attention_dropout": 0.0,
8
+ "bos_token_id": 0,
9
+ "eos_token_id": 1,
10
+ "ep_size": 1,
11
+ "first_k_dense_replace": 3,
12
+ "hidden_act": "silu",
13
+ "hidden_size": 7168,
14
+ "index_head_dim": 128,
15
+ "index_n_heads": 64,
16
+ "index_topk": 2048,
17
+ "initializer_range": 0.02,
18
+ "intermediate_size": 18432,
19
+ "kv_lora_rank": 512,
20
+ "max_position_embeddings": 163840,
21
+ "moe_intermediate_size": 2048,
22
+ "moe_layer_freq": 1,
23
+ "n_group": 8,
24
+ "n_routed_experts": 256,
25
+ "n_shared_experts": 1,
26
+ "norm_topk_prob": true,
27
+ "num_attention_heads": 128,
28
+ "num_experts_per_tok": 8,
29
+ "num_hidden_layers": 61,
30
+ "num_key_value_heads": 128,
31
+ "num_nextn_predict_layers": 1,
32
+ "q_lora_rank": 1536,
33
+ "qk_nope_head_dim": 128,
34
+ "qk_rope_head_dim": 64,
35
+ "quantization_config": {
36
+ "activation_scheme": "dynamic",
37
+ "fmt": "e4m3",
38
+ "quant_method": "fp8",
39
+ "scale_fmt": "ue8m0",
40
+ "weight_block_size": [
41
+ 128,
42
+ 128
43
+ ]
44
+ },
45
+ "rms_norm_eps": 1e-06,
46
+ "rope_scaling": {
47
+ "beta_fast": 32,
48
+ "beta_slow": 1,
49
+ "factor": 40,
50
+ "mscale": 1.0,
51
+ "mscale_all_dim": 1.0,
52
+ "original_max_position_embeddings": 4096,
53
+ "type": "yarn"
54
+ },
55
+ "rope_theta": 10000.0,
56
+ "routed_scaling_factor": 2.5,
57
+ "scoring_func": "sigmoid",
58
+ "tie_word_embeddings": false,
59
+ "topk_group": 4,
60
+ "topk_method": "noaux_tc",
61
+ "torch_dtype": "bfloat16",
62
+ "use_cache": true,
63
+ "v_head_dim": 128,
64
+ "vocab_size": 129280,
65
+ "vision_config": {
66
+ "model_type": "pixtral",
67
+ "adapter_bias": false,
68
+ "add_pre_mm_projector_layer_norm": true,
69
+ "hidden_size": 1664,
70
+ "image_break_token_id": 128816,
71
+ "image_end_token_id": 128817,
72
+ "image_size": 1540,
73
+ "image_token_id": 128815,
74
+ "intermediate_size": 8192,
75
+ "max_image_size": 1540,
76
+ "mm_projector_id": "patch_merge",
77
+ "num_attention_heads": 16,
78
+ "num_channels": 3,
79
+ "num_hidden_layers": 48,
80
+ "patch_size": 14,
81
+ "rope_theta": 10000.0,
82
+ "spatial_merge_size": 2
83
+ }
84
+ }