add weights, config for 450k model
Browse files- config.json +27 -0
- weights.ckpt +3 -0
config.json
ADDED
|
@@ -0,0 +1,27 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"d_model": 256,
|
| 3 |
+
"n_layer": 8,
|
| 4 |
+
"d_inner": 1024,
|
| 5 |
+
"vocab_size": 12,
|
| 6 |
+
"resid_dropout": 0.0,
|
| 7 |
+
"embed_dropout": 0.1,
|
| 8 |
+
"fused_mlp": false,
|
| 9 |
+
"fused_dropout_add_ln": true,
|
| 10 |
+
"checkpoint_mixer": true,
|
| 11 |
+
"checkpoint_mlp": true,
|
| 12 |
+
"residual_in_fp32": true,
|
| 13 |
+
"pad_vocab_size_multiple": 8,
|
| 14 |
+
"return_hidden_state": true,
|
| 15 |
+
"layer": {
|
| 16 |
+
"_name_": "hyena",
|
| 17 |
+
"emb_dim": 5,
|
| 18 |
+
"filter_order": 64,
|
| 19 |
+
"local_order": 3,
|
| 20 |
+
"l_max": 450002,
|
| 21 |
+
"modulate": true,
|
| 22 |
+
"w": 10,
|
| 23 |
+
"lr": 6e-4,
|
| 24 |
+
"wd": 0.0,
|
| 25 |
+
"lr_pos_emb": 0.0
|
| 26 |
+
}
|
| 27 |
+
}
|
weights.ckpt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3e315ab560f99e26db939267ed40224f3be771791833475c0794928a4cdbca1b
|
| 3 |
+
size 165322737
|