llaa33219 commited on
Commit
6bd161d
·
verified ·
1 Parent(s): 300d8a0

Upload config.json with huggingface_hub

Browse files
Files changed (1) hide show
  1. config.json +42 -0
config.json ADDED
@@ -0,0 +1,42 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "model_type": "solar_open",
3
+ "architectures": [
4
+ "SolarOpenForCausalLM"
5
+ ],
6
+ "auto_map": {
7
+ "AutoConfig": "configuration_solar_open.SolarOpenConfig",
8
+ "AutoModel": "modeling_solar_open.SolarOpenModel",
9
+ "AutoModelForCausalLM": "modeling_solar_open.SolarOpenForCausalLM"
10
+ },
11
+ "pad_token_id": 2,
12
+ "bos_token_id": 1,
13
+ "eos_token_id": 2,
14
+ "partial_rotary_factor": 1.0,
15
+ "hidden_size": 4096,
16
+ "num_hidden_layers": 20,
17
+ "num_attention_heads": 64,
18
+ "head_dim": 128,
19
+ "num_key_value_heads": 8,
20
+ "vocab_size": 196608,
21
+ "intermediate_size": 10240,
22
+ "moe_intermediate_size": 1280,
23
+ "rms_norm_eps": 1e-05,
24
+ "rope_theta": 1000000,
25
+ "max_position_embeddings": 131072,
26
+ "n_routed_experts": 12,
27
+ "n_shared_experts": 1,
28
+ "norm_topk_prob": true,
29
+ "routed_scaling_factor": 1.0,
30
+ "num_experts_per_tok": 8,
31
+ "tie_word_embeddings": false,
32
+ "torch_dtype": "bfloat16",
33
+ "first_k_dense_replace": 0,
34
+ "use_cache": true,
35
+ "transformers_version": "4.57.1",
36
+ "rope_scaling": {
37
+ "type": "yarn",
38
+ "factor": 2.0,
39
+ "original_max_position_embeddings": 65536
40
+ },
41
+ "num_local_experts": 12
42
+ }