Cheshire-AI
commited on
Commit
•
c1faefd
1
Parent(s):
fa35084
Add config.json
Browse files- config.json +42 -0
config.json
ADDED
@@ -0,0 +1,42 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "CohereForAI/c4ai-command-r-plus",
|
3 |
+
"architectures": [
|
4 |
+
"CohereForCausalLM"
|
5 |
+
],
|
6 |
+
"attention_bias": false,
|
7 |
+
"attention_dropout": 0.0,
|
8 |
+
"bos_token_id": 5,
|
9 |
+
"eos_token_id": 255001,
|
10 |
+
"hidden_act": "silu",
|
11 |
+
"hidden_size": 12288,
|
12 |
+
"initializer_range": 0.02,
|
13 |
+
"intermediate_size": 33792,
|
14 |
+
"layer_norm_eps": 1e-05,
|
15 |
+
"logit_scale": 0.8333333333333334,
|
16 |
+
"max_position_embeddings": 8192,
|
17 |
+
"model_max_length": 131072,
|
18 |
+
"model_type": "cohere",
|
19 |
+
"num_attention_heads": 96,
|
20 |
+
"num_hidden_layers": 64,
|
21 |
+
"num_key_value_heads": 8,
|
22 |
+
"pad_token_id": 0,
|
23 |
+
"quantization_config": {
|
24 |
+
"bits": 4,
|
25 |
+
"checkpoint_format": "gptq",
|
26 |
+
"damp_percent": 0.1,
|
27 |
+
"desc_act": true,
|
28 |
+
"group_size": 128,
|
29 |
+
"model_file_base_name": null,
|
30 |
+
"model_name_or_path": null,
|
31 |
+
"quant_method": "gptq",
|
32 |
+
"static_groups": false,
|
33 |
+
"sym": true,
|
34 |
+
"true_sequential": true
|
35 |
+
},
|
36 |
+
"rope_theta": 75000000.0,
|
37 |
+
"torch_dtype": "float16",
|
38 |
+
"transformers_version": "4.42.3",
|
39 |
+
"use_cache": true,
|
40 |
+
"use_qk_norm": true,
|
41 |
+
"vocab_size": 256000
|
42 |
+
}
|