xieli123 commited on
Commit
36d8894
·
verified ·
1 Parent(s): b4e91cc

Upload config.json with huggingface_hub

Browse files
Files changed (1) hide show
  1. config.json +59 -0
config.json ADDED
@@ -0,0 +1,59 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "architectures": [
3
+ "Step1ForCausalLM"
4
+ ],
5
+ "auto_map": {
6
+ "AutoConfig": "configuration_step1.Step1Config"
7
+ },
8
+ "bos_token_id": 1,
9
+ "eos_token_id": 3,
10
+ "hidden_size": 3072,
11
+ "intermediate_size": 8192,
12
+ "max_seq_len": 32768,
13
+ "model_type": "step1",
14
+ "num_attention_groups": 4,
15
+ "num_attention_heads": 48,
16
+ "num_hidden_layers": 32,
17
+ "pad_token_id": 0,
18
+ "quantization_config": {
19
+ "config_groups": {
20
+ "group_0": {
21
+ "format": "pack-quantized",
22
+ "input_activations": null,
23
+ "output_activations": null,
24
+ "targets": [
25
+ "Linear"
26
+ ],
27
+ "weights": {
28
+ "actorder": null,
29
+ "block_structure": null,
30
+ "dynamic": false,
31
+ "group_size": 128,
32
+ "num_bits": 4,
33
+ "observer": "minmax",
34
+ "observer_kwargs": {},
35
+ "strategy": "group",
36
+ "symmetric": true,
37
+ "type": "int"
38
+ }
39
+ }
40
+ },
41
+ "format": "pack-quantized",
42
+ "global_compression_ratio": null,
43
+ "ignore": [
44
+ "lm_head"
45
+ ],
46
+ "kv_cache_scheme": null,
47
+ "quant_method": "compressed-tensors",
48
+ "quantization_status": "compressed",
49
+ "sparsity_config": {},
50
+ "transform_config": {},
51
+ "version": "0.11.0"
52
+ },
53
+ "rms_norm_eps": 1e-05,
54
+ "tie_word_embeddings": false,
55
+ "torch_dtype": "bfloat16",
56
+ "transformers_version": "4.53.3",
57
+ "use_cache": true,
58
+ "vocab_size": 74752
59
+ }