Add config.json (Qwen3, 20 layers, 1024 hidden, GQA 16q/8kv, tied embeddings)

#1
Files changed (1) hide show
  1. config.json +55 -0
config.json ADDED
@@ -0,0 +1,55 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "Qwen3ForCausalLM"
4
+ ],
5
+ "attention_bias": false,
6
+ "attention_dropout": 0.0,
7
+ "bos_token_id": 1,
8
+ "dtype": "float32",
9
+ "eos_token_id": 2,
10
+ "head_dim": 64,
11
+ "hidden_act": "silu",
12
+ "hidden_size": 1024,
13
+ "initializer_range": 0.02,
14
+ "intermediate_size": 2944,
15
+ "layer_types": [
16
+ "full_attention",
17
+ "full_attention",
18
+ "full_attention",
19
+ "full_attention",
20
+ "full_attention",
21
+ "full_attention",
22
+ "full_attention",
23
+ "full_attention",
24
+ "full_attention",
25
+ "full_attention",
26
+ "full_attention",
27
+ "full_attention",
28
+ "full_attention",
29
+ "full_attention",
30
+ "full_attention",
31
+ "full_attention",
32
+ "full_attention",
33
+ "full_attention",
34
+ "full_attention",
35
+ "full_attention"
36
+ ],
37
+ "max_position_embeddings": 2048,
38
+ "max_window_layers": 20,
39
+ "model_type": "qwen3",
40
+ "num_attention_heads": 16,
41
+ "num_hidden_layers": 20,
42
+ "num_key_value_heads": 8,
43
+ "pad_token_id": 3,
44
+ "rms_norm_eps": 1e-06,
45
+ "rope_parameters": {
46
+ "rope_theta": 10000.0,
47
+ "rope_type": "default"
48
+ },
49
+ "sliding_window": null,
50
+ "tie_word_embeddings": true,
51
+ "transformers_version": "5.8.0",
52
+ "use_cache": false,
53
+ "use_sliding_window": false,
54
+ "vocab_size": 49664
55
+ }