deepseekrzz commited on
Commit
732399b
1 Parent(s): 57df132

Update config.json

Browse files
Files changed (1) hide show
  1. config.json +29 -50
config.json CHANGED
@@ -1,52 +1,31 @@
1
  {
2
- "architectures": [
3
- "DeepseekForCausalLM"
4
- ],
5
- "attention_bias": false,
6
- "attention_dropout": 0.0,
7
- "auto_map": {
8
- "AutoConfig": "configuration_deepseek.DeepseekConfig",
9
- "AutoModel": "modeling_deepseek.DeepseekModel",
10
- "AutoModelForCausalLM": "modeling_deepseek.DeepseekForCausalLM"
11
- },
12
- "aux_loss_alpha": 0.001,
13
- "bos_token_id": 100000,
14
- "eos_token_id": 100001,
15
- "ep_size": 1,
16
- "first_k_dense_replace": null,
17
- "hidden_act": "silu",
18
- "hidden_size": 4096,
19
- "initializer_range": 0.02,
20
- "intermediate_size": 11008,
21
- "kv_lora_rank": null,
22
- "max_position_embeddings": 4096,
23
- "model_type": "deepseek",
24
- "moe_intermediate_size": 11008,
25
- "moe_layer_freq": null,
26
- "n_group": 1,
27
- "n_routed_experts": null,
28
- "n_shared_experts": null,
29
- "norm_topk_prob": null,
30
- "num_attention_heads": 32,
31
- "num_experts_per_tok": null,
32
- "num_hidden_layers": 30,
33
- "num_key_value_heads": 32,
34
- "pretraining_tp": 1,
35
- "q_lora_rank": null,
36
- "qk_nope_head_dim": 0,
37
- "qk_rope_head_dim": null,
38
- "rms_norm_eps": 1e-06,
39
- "rope_scaling": null,
40
- "rope_theta": 10000,
41
- "routed_scaling_factor": 1.0,
42
- "scoring_func": null,
43
- "seq_aux": true,
44
- "tie_word_embeddings": false,
45
- "topk_group": 1,
46
- "topk_method": "group_limited_greedy",
47
- "torch_dtype": "bfloat16",
48
- "transformers_version": "4.33.1",
49
- "use_cache": true,
50
- "v_head_dim": null,
51
- "vocab_size": 102400
52
  }
 
1
  {
2
+ "architectures": [
3
+ "LlamaForCausalLM"
4
+ ],
5
+ "attention_bias": false,
6
+ "attention_dropout": 0.0,
7
+ "bos_token_id": 100000,
8
+ "eos_token_id": 100001,
9
+ "hidden_act": "silu",
10
+ "hidden_size": 4096,
11
+ "initializer_range": 0.02,
12
+ "intermediate_size": 11008,
13
+ "max_position_embeddings": 4096,
14
+ "model_type": "llama",
15
+ "num_attention_heads": 32,
16
+ "num_hidden_layers": 30,
17
+ "num_key_value_heads": 32,
18
+ "pretraining_tp": 1,
19
+ "quantization": {
20
+ "group_size": 64,
21
+ "bits": 8
22
+ },
23
+ "rms_norm_eps": 1e-06,
24
+ "rope_scaling": null,
25
+ "rope_theta": 10000,
26
+ "tie_word_embeddings": false,
27
+ "torch_dtype": "bfloat16",
28
+ "transformers_version": "4.33.1",
29
+ "use_cache": true,
30
+ "vocab_size": 102400
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
31
  }