vinhnx90 commited on
Commit
e4d785b
·
verified ·
1 Parent(s): 48198aa

Upload config.json with huggingface_hub

Browse files
Files changed (1) hide show
  1. config.json +44 -0
config.json ADDED
@@ -0,0 +1,44 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "Gemma3ForCausalLM"
4
+ ],
5
+ "attention_bias": false,
6
+ "attention_dropout": 0.0,
7
+ "attn_logit_softcapping": null,
8
+ "bos_token_id": 2,
9
+ "cache_implementation": "hybrid",
10
+ "eos_token_id": 106,
11
+ "final_logit_softcapping": null,
12
+ "head_dim": 256,
13
+ "hidden_activation": "gelu_pytorch_tanh",
14
+ "hidden_size": 1152,
15
+ "initializer_range": 0.02,
16
+ "intermediate_size": 6912,
17
+ "max_position_embeddings": 32768,
18
+ "model_type": "gemma3_text",
19
+ "num_attention_heads": 4,
20
+ "num_hidden_layers": 26,
21
+ "num_key_value_heads": 1,
22
+ "pad_token_id": 0,
23
+ "quantization": {
24
+ "group_size": 64,
25
+ "bits": 6
26
+ },
27
+ "quantization_config": {
28
+ "group_size": 64,
29
+ "bits": 6
30
+ },
31
+ "query_pre_attn_scalar": 256,
32
+ "rms_norm_eps": 1e-06,
33
+ "rope_local_base_freq": 10000,
34
+ "rope_scaling": null,
35
+ "rope_theta": 1000000,
36
+ "sliding_window": 512,
37
+ "sliding_window_pattern": 6,
38
+ "torch_dtype": "float16",
39
+ "transformers_version": "4.50.0.dev0",
40
+ "unsloth_fixed": true,
41
+ "unsloth_version": "2025.3.18",
42
+ "use_cache": true,
43
+ "vocab_size": 262144
44
+ }