apepkuss79 commited on
Commit
e620559
1 Parent(s): 7d0d798

Upload config.json with huggingface_hub

Browse files
Files changed (1) hide show
  1. config.json +59 -0
config.json ADDED
@@ -0,0 +1,59 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "DeepseekV2ForCausalLM"
4
+ ],
5
+ "attention_bias": false,
6
+ "attention_dropout": 0.0,
7
+ "auto_map": {
8
+ "AutoConfig": "configuration_deepseek.DeepseekV2Config",
9
+ "AutoModel": "modeling_deepseek.DeepseekV2Model",
10
+ "AutoModelForCausalLM": "modeling_deepseek.DeepseekV2ForCausalLM"
11
+ },
12
+ "aux_loss_alpha": 0.001,
13
+ "bos_token_id": 100000,
14
+ "eos_token_id": 100001,
15
+ "first_k_dense_replace": 1,
16
+ "hidden_act": "silu",
17
+ "hidden_size": 2048,
18
+ "initializer_range": 0.02,
19
+ "intermediate_size": 10944,
20
+ "kv_lora_rank": 512,
21
+ "max_position_embeddings": 163840,
22
+ "model_type": "deepseek_v2",
23
+ "moe_intermediate_size": 1408,
24
+ "moe_layer_freq": 1,
25
+ "n_group": 1,
26
+ "n_routed_experts": 64,
27
+ "n_shared_experts": 2,
28
+ "norm_topk_prob": false,
29
+ "num_attention_heads": 16,
30
+ "num_experts_per_tok": 6,
31
+ "num_hidden_layers": 27,
32
+ "num_key_value_heads": 16,
33
+ "pretraining_tp": 1,
34
+ "q_lora_rank": null,
35
+ "qk_nope_head_dim": 128,
36
+ "qk_rope_head_dim": 64,
37
+ "rms_norm_eps": 1e-06,
38
+ "rope_scaling": {
39
+ "beta_fast": 32,
40
+ "beta_slow": 1,
41
+ "factor": 40,
42
+ "mscale": 0.707,
43
+ "mscale_all_dim": 0.707,
44
+ "original_max_position_embeddings": 4096,
45
+ "type": "yarn"
46
+ },
47
+ "rope_theta": 10000,
48
+ "routed_scaling_factor": 1.0,
49
+ "scoring_func": "softmax",
50
+ "seq_aux": true,
51
+ "tie_word_embeddings": false,
52
+ "topk_group": 1,
53
+ "topk_method": "greedy",
54
+ "torch_dtype": "bfloat16",
55
+ "transformers_version": "4.39.3",
56
+ "use_cache": true,
57
+ "v_head_dim": 128,
58
+ "vocab_size": 102400
59
+ }