Upload 4 files
Browse files- config.json +36 -0
- pytorch_model.pt +3 -0
- spiece.model +3 -0
- tokenizer.json +0 -0
config.json
ADDED
@@ -0,0 +1,36 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"vocab_size": 32128,
|
3 |
+
"dim_model": 1024,
|
4 |
+
"num_heads": 32,
|
5 |
+
"dim_head" : 128,
|
6 |
+
"dim_ff" : 16384,
|
7 |
+
"num_encoder_layers" : 24,
|
8 |
+
"num_decoder_layers" : 24,
|
9 |
+
"dropout_p" : 0.0,
|
10 |
+
"emb_init_mean": 0.0,
|
11 |
+
"emb_init_std": 1.0,
|
12 |
+
"pos_bias_type": "relative",
|
13 |
+
"position_bias_num_buckets" : 32,
|
14 |
+
"position_bias_max_distance" : 128,
|
15 |
+
"pos_init_mean" : 0.0,
|
16 |
+
"pos_init_std" : 1.0,
|
17 |
+
"norm_init_var" : 1.0,
|
18 |
+
"norm_bias": false,
|
19 |
+
"norm_eps" : 1e-6,
|
20 |
+
"att_init_mean" : 0.0,
|
21 |
+
"att_init_std" : 1.0,
|
22 |
+
"att_bias": false,
|
23 |
+
"att_mask_value": "-inf",
|
24 |
+
"ffn_init_mean" : 0.0,
|
25 |
+
"ffn_init_std" : 1.0,
|
26 |
+
"ffn_bias": false,
|
27 |
+
"ffn_activate_fn": "relu",
|
28 |
+
"proj_init_mean": 0.0,
|
29 |
+
"proj_init_std": 1.0,
|
30 |
+
"proj_bias": false,
|
31 |
+
"length_scale" : false,
|
32 |
+
"attn_scale": false,
|
33 |
+
"half" : true,
|
34 |
+
"int8" : false,
|
35 |
+
"tied" : true
|
36 |
+
}
|
pytorch_model.pt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:4b263e64460bf62d0182536de3995f75db26d0c40178dabf820587f9a28ed603
|
3 |
+
size 5703367479
|
spiece.model
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:d60acb128cf7b7f2536e8f38a5b18a05535c9e14c7a355904270e15b0945ea86
|
3 |
+
size 791656
|
tokenizer.json
ADDED
The diff for this file is too large to render.
See raw diff
|
|