VarunGumma
commited on
Commit
•
242d57e
1
Parent(s):
f0e061d
Upload config.json with huggingface_hub
Browse files- config.json +41 -0
config.json
ADDED
@@ -0,0 +1,41 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"activation_dropout": 0.0,
|
3 |
+
"activation_function": "gelu",
|
4 |
+
"architectures": [
|
5 |
+
"RotaryIndicTransForConditionalGeneration"
|
6 |
+
],
|
7 |
+
"attention_dropout": 0.0,
|
8 |
+
"attn_implementation": "eager",
|
9 |
+
"bos_token_id": 0,
|
10 |
+
"decoder_attention_heads": 16,
|
11 |
+
"decoder_embed_dim": 1024,
|
12 |
+
"decoder_ffn_dim": 8192,
|
13 |
+
"decoder_layerdrop": 0,
|
14 |
+
"decoder_layers": 18,
|
15 |
+
"decoder_normalize_before": true,
|
16 |
+
"decoder_start_token_id": 2,
|
17 |
+
"decoder_vocab_size": 32296,
|
18 |
+
"dropout": 0.2,
|
19 |
+
"encoder_attention_heads": 16,
|
20 |
+
"encoder_embed_dim": 1024,
|
21 |
+
"encoder_ffn_dim": 8192,
|
22 |
+
"encoder_layerdrop": 0,
|
23 |
+
"encoder_layers": 18,
|
24 |
+
"encoder_normalize_before": true,
|
25 |
+
"encoder_vocab_size": 122706,
|
26 |
+
"eos_token_id": 2,
|
27 |
+
"init_std": 0.02,
|
28 |
+
"is_encoder_decoder": true,
|
29 |
+
"layernorm_embedding": false,
|
30 |
+
"model_type": "RotaryIndicTrans",
|
31 |
+
"num_hidden_layers": 18,
|
32 |
+
"pad_token_id": 1,
|
33 |
+
"rope_args": {
|
34 |
+
"theta": 10000
|
35 |
+
},
|
36 |
+
"scale_embedding": true,
|
37 |
+
"share_decoder_input_output_embed": false,
|
38 |
+
"torch_dtype": "float32",
|
39 |
+
"transformers_version": "4.44.0",
|
40 |
+
"use_cache": true
|
41 |
+
}
|