Upload folder using huggingface_hub
Browse files- best_model.pt +2 -2
- config.json +13 -6
- tokenizer_src.json +0 -0
- tokenizer_tgt.json +0 -0
best_model.pt
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e823e872870368411ef59dbccce0327bebb3ffea203188b65d67c2bcacf4050a
|
| 3 |
+
size 138081340
|
config.json
CHANGED
|
@@ -1,9 +1,4 @@
|
|
| 1 |
{
|
| 2 |
-
"run_name": "c2",
|
| 3 |
-
"positional_encoding": "rope",
|
| 4 |
-
"attention_type": "mha",
|
| 5 |
-
"norm_type": "layernorm",
|
| 6 |
-
"tokenization": "subword",
|
| 7 |
"d_model": 256,
|
| 8 |
"num_heads": 8,
|
| 9 |
"num_encoder_layers": 4,
|
|
@@ -14,9 +9,21 @@
|
|
| 14 |
"learning_rate": 0.0003,
|
| 15 |
"warmup_steps": 2000,
|
| 16 |
"batch_size": 64,
|
|
|
|
| 17 |
"max_seq_len": 512,
|
| 18 |
"seed": 42,
|
| 19 |
"epochs": 50,
|
| 20 |
-
"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 21 |
"device": "cuda"
|
| 22 |
}
|
|
|
|
| 1 |
{
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2 |
"d_model": 256,
|
| 3 |
"num_heads": 8,
|
| 4 |
"num_encoder_layers": 4,
|
|
|
|
| 9 |
"learning_rate": 0.0003,
|
| 10 |
"warmup_steps": 2000,
|
| 11 |
"batch_size": 64,
|
| 12 |
+
"grad_clip_norm": 1.0,
|
| 13 |
"max_seq_len": 512,
|
| 14 |
"seed": 42,
|
| 15 |
"epochs": 50,
|
| 16 |
+
"patience": 5,
|
| 17 |
+
"bpe_vocab_size": 8000,
|
| 18 |
+
"positional_encoding": "rope",
|
| 19 |
+
"attention_type": "mha",
|
| 20 |
+
"norm_type": "layernorm",
|
| 21 |
+
"tokenization": "subword",
|
| 22 |
+
"gqa_kv_heads": 2,
|
| 23 |
+
"blt_patch_size": 9,
|
| 24 |
+
"d_local": 256,
|
| 25 |
+
"num_local_layers": 4,
|
| 26 |
+
"run_name": "c2",
|
| 27 |
+
"num_params": 11477824,
|
| 28 |
"device": "cuda"
|
| 29 |
}
|
tokenizer_src.json
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_tgt.json
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|