zhao1iang commited on
Commit
6f38345
1 Parent(s): 452b672

Update config.json

Browse files
Files changed (1) hide show
  1. config.json +39 -1
config.json CHANGED
@@ -1 +1,39 @@
1
- {"architectures": ["SkyworkForCausalLM"], "auto_map": {"AutoConfig": "configuration_skywork_moe.SkyworkMoeConfig", "AutoModelForCausalLM": "modeling_skywork_moe.SkyworkForCausalLM"}, "model_type": "skywork", "vocab_size": 65532, "bos_token_id": 1, "eos_token_id": 2, "pad_token_id": 0, "hidden_act": "silu", "hidden_size": 4608, "initializer_range": 0.01, "intermediate_size": 12288, "max_position_embeddings": 8192, "num_attention_heads": 36, "num_key_value_heads": 36, "num_hidden_layers": 52, "num_experts": [16], "moe_use_skywork_gating": false, "moe_2layer_gate": false, "moe_use_logits_norm": true, "moe_gate_norm_std": 1.0, "moe_feature_no_mul_topk": true, "sliding_window": null, "moe_expert_interval": 1, "rms_norm_eps": 1e-06, "rotary_percent": 1.0, "tie_word_embeddings": false, "torch_dtype": "bfloat16", "use_cache": true, "transformers_version": "4.40.1", "rope_theta": 10000}
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "SkyworkForCausalLM"
4
+ ],
5
+ "auto_map": {
6
+ "AutoConfig": "configuration_skywork_moe.SkyworkMoeConfig",
7
+ "AutoModelForCausalLM": "modeling_skywork_moe.SkyworkForCausalLM"
8
+ },
9
+ "model_type": "skywork_moe",
10
+ "vocab_size": 65532,
11
+ "bos_token_id": 1,
12
+ "eos_token_id": 2,
13
+ "pad_token_id": 0,
14
+ "hidden_act": "silu",
15
+ "hidden_size": 4608,
16
+ "initializer_range": 0.01,
17
+ "intermediate_size": 12288,
18
+ "max_position_embeddings": 8192,
19
+ "num_attention_heads": 36,
20
+ "num_key_value_heads": 36,
21
+ "num_hidden_layers": 52,
22
+ "num_experts": [
23
+ 16
24
+ ],
25
+ "moe_use_skywork_gating": false,
26
+ "moe_2layer_gate": false,
27
+ "moe_use_logits_norm": true,
28
+ "moe_gate_norm_std": 1.0,
29
+ "moe_feature_no_mul_topk": true,
30
+ "sliding_window": null,
31
+ "moe_expert_interval": 1,
32
+ "rms_norm_eps": 1e-06,
33
+ "rotary_percent": 1.0,
34
+ "tie_word_embeddings": false,
35
+ "torch_dtype": "bfloat16",
36
+ "use_cache": true,
37
+ "transformers_version": "4.40.1",
38
+ "rope_theta": 10000
39
+ }