onkarpandit-g42
commited on
Commit
•
de07f29
1
Parent(s):
a4a238a
Upload config.json with huggingface_hub
Browse files- config.json +43 -0
config.json
ADDED
@@ -0,0 +1,43 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_name_or_path": "core42/jais-family-13b-chat",
|
3 |
+
"activation_function": "swiglu",
|
4 |
+
"alibi_scaling": null,
|
5 |
+
"architectures": [
|
6 |
+
"JAISLMHeadModel"
|
7 |
+
],
|
8 |
+
"attn_pdrop": 0.0,
|
9 |
+
"auto_map": {
|
10 |
+
"AutoConfig": "configuration_jais.JAISConfig",
|
11 |
+
"AutoModel": "modeling_jais.JAISModel",
|
12 |
+
"AutoModelForCausalLM": "modeling_jais.JAISLMHeadModel",
|
13 |
+
"AutoModelForQuestionAnswering": "modeling_jais.JAISForQuestionAnswering",
|
14 |
+
"AutoModelForSequenceClassification": "modeling_jais.JAISForSequenceClassification",
|
15 |
+
"AutoModelForTokenClassification": "modeling_jais.JAISForTokenClassification"
|
16 |
+
},
|
17 |
+
"bos_token_id": 0,
|
18 |
+
"embd_pdrop": 0.0,
|
19 |
+
"eos_token_id": 0,
|
20 |
+
"initializer_range": 0.02,
|
21 |
+
"layer_norm_epsilon": 1e-05,
|
22 |
+
"model_type": "jais",
|
23 |
+
"mup_embeddings_scale": 9.1705785388303,
|
24 |
+
"mup_output_alpha": 1.0951834981576898,
|
25 |
+
"mup_scale_qk_dot_by_d": true,
|
26 |
+
"mup_width_scale": 0.05,
|
27 |
+
"n_embd": 5120,
|
28 |
+
"n_head": 40,
|
29 |
+
"n_inner": 13664,
|
30 |
+
"n_layer": 40,
|
31 |
+
"n_positions": 2048,
|
32 |
+
"pad_token_id": 0,
|
33 |
+
"position_embedding_type": "alibi",
|
34 |
+
"reorder_and_upcast_attn": false,
|
35 |
+
"resid_pdrop": 0.0,
|
36 |
+
"rotary_dim": null,
|
37 |
+
"scale_attn_by_inverse_layer_idx": false,
|
38 |
+
"scale_attn_weights": true,
|
39 |
+
"torch_dtype": "float32",
|
40 |
+
"transformers_version": "4.40.1",
|
41 |
+
"use_cache": true,
|
42 |
+
"vocab_size": 84992
|
43 |
+
}
|