RichardErkhov
commited on
Commit
•
605a4b7
1
Parent(s):
bc02817
uploaded model
Browse files- config.json +39 -0
config.json
ADDED
@@ -0,0 +1,39 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "mini-llama",
|
3 |
+
"architectures": [
|
4 |
+
"LlamaForCausalLM"
|
5 |
+
],
|
6 |
+
"attention_bias": false,
|
7 |
+
"attention_dropout": 0.0,
|
8 |
+
"bos_token_id": 2,
|
9 |
+
"eos_token_id": 3,
|
10 |
+
"head_dim": 64,
|
11 |
+
"hidden_act": "silu",
|
12 |
+
"hidden_size": 2048,
|
13 |
+
"initializer_range": 0.02,
|
14 |
+
"intermediate_size": 8192,
|
15 |
+
"max_position_embeddings": 2048,
|
16 |
+
"mlp_bias": false,
|
17 |
+
"model_type": "llama",
|
18 |
+
"num_attention_heads": 32,
|
19 |
+
"num_hidden_layers": 24,
|
20 |
+
"num_key_value_heads": 32,
|
21 |
+
"pretraining_tp": 1,
|
22 |
+
"quantization_config": {
|
23 |
+
"bits": 4,
|
24 |
+
"group_size": 128,
|
25 |
+
"modules_to_not_convert": null,
|
26 |
+
"quant_method": "awq",
|
27 |
+
"version": "gemm",
|
28 |
+
"zero_point": true
|
29 |
+
},
|
30 |
+
"rms_norm_eps": 1e-05,
|
31 |
+
"rope_scaling": null,
|
32 |
+
"rope_theta": 10000.0,
|
33 |
+
"tie_word_embeddings": true,
|
34 |
+
"tokenizer_class": "GPTSw3Tokenizer",
|
35 |
+
"torch_dtype": "float16",
|
36 |
+
"transformers_version": "4.46.3",
|
37 |
+
"use_cache": false,
|
38 |
+
"vocab_size": 64000
|
39 |
+
}
|