kgrabko commited on
Commit
7d96e78
·
verified ·
1 Parent(s): 3923ae7

Create config.json

Browse files
Files changed (1) hide show
  1. config.json +37 -0
config.json ADDED
@@ -0,0 +1,37 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "Qwen3_5ForCausalLM"
4
+ ],
5
+ "model_type": "qwen35",
6
+ "vocab_size": 248320,
7
+ "hidden_size": 5120,
8
+ "intermediate_size": 17408,
9
+ "num_hidden_layers": 65,
10
+ "num_attention_heads": 24,
11
+ "num_key_value_heads": 4,
12
+ "head_dim": 256,
13
+ "max_position_embeddings": 262144,
14
+ "rope_theta": 10000000,
15
+ "rms_norm_eps": 1e-6,
16
+ "hidden_act": "silu",
17
+ "tie_word_embeddings": false,
18
+ "use_cache": true,
19
+
20
+ "linear_attn_config": {
21
+ "conv_kernel_size": 4,
22
+ "state_size": 128,
23
+ "num_groups": 16,
24
+ "time_step_rank": 48,
25
+ "inner_size": 6144
26
+ },
27
+
28
+ "bos_token_id": 248044,
29
+ "pad_token_id": 248044,
30
+ "eos_token_id": [
31
+ 248046,
32
+ 248044
33
+ ],
34
+
35
+ "torch_dtype": "bfloat16",
36
+ "transformers_version": "4.51.0"
37
+ }