Commit
·
35cea74
1
Parent(s):
b83af74
Upload 11 files
Browse files- added_tokens.json +7 -0
- config (6).json +38 -0
- optimizer.pt +3 -0
- pytorch_model (2).bin +3 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- special_tokens_map.json +7 -0
- tokenizer_config.json +54 -0
- trainer_state.json +1293 -0
- training_args.bin +3 -0
- vocab.txt +33 -0
added_tokens.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<cls>": 0,
|
| 3 |
+
"<eos>": 2,
|
| 4 |
+
"<mask>": 32,
|
| 5 |
+
"<pad>": 1,
|
| 6 |
+
"<unk>": 3
|
| 7 |
+
}
|
config (6).json
ADDED
|
@@ -0,0 +1,38 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "facebook/esm2_t6_8M_UR50D",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"EsmForTokenClassification"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.5,
|
| 7 |
+
"classifier_dropout": null,
|
| 8 |
+
"emb_layer_norm_before": false,
|
| 9 |
+
"esmfold_config": null,
|
| 10 |
+
"hidden_act": "gelu",
|
| 11 |
+
"hidden_dropout_prob": 0.5,
|
| 12 |
+
"hidden_size": 320,
|
| 13 |
+
"id2label": {
|
| 14 |
+
"0": "No binding site",
|
| 15 |
+
"1": "Binding site"
|
| 16 |
+
},
|
| 17 |
+
"initializer_range": 0.02,
|
| 18 |
+
"intermediate_size": 1280,
|
| 19 |
+
"is_folding_model": false,
|
| 20 |
+
"label2id": {
|
| 21 |
+
"Binding site": 1,
|
| 22 |
+
"No binding site": 0
|
| 23 |
+
},
|
| 24 |
+
"layer_norm_eps": 1e-05,
|
| 25 |
+
"mask_token_id": 32,
|
| 26 |
+
"max_position_embeddings": 1026,
|
| 27 |
+
"model_type": "esm",
|
| 28 |
+
"num_attention_heads": 20,
|
| 29 |
+
"num_hidden_layers": 6,
|
| 30 |
+
"pad_token_id": 1,
|
| 31 |
+
"position_embedding_type": "rotary",
|
| 32 |
+
"token_dropout": true,
|
| 33 |
+
"torch_dtype": "float32",
|
| 34 |
+
"transformers_version": "4.34.0.dev0",
|
| 35 |
+
"use_cache": true,
|
| 36 |
+
"vocab_list": null,
|
| 37 |
+
"vocab_size": 33
|
| 38 |
+
}
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ed01130202ab25bb28e408aff3d24b21e5262e7b557d9e0cd96d55cac4e0692a
|
| 3 |
+
size 59361093
|
pytorch_model (2).bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0a0e3863ebd00a136b483eb80d58882f03a47f08d7767b8a5158682fd7cc8c5c
|
| 3 |
+
size 30989061
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1a85c8f82e6ff59890e9b6b487c977f3ce1fcedec53eae378d17f83e2ecbb1a6
|
| 3 |
+
size 14575
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:73eb32d30209f5b72ceb1037fc68aabd56fd23c02d2f88302d7a2c36d742bf16
|
| 3 |
+
size 627
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"cls_token": "<cls>",
|
| 3 |
+
"eos_token": "<eos>",
|
| 4 |
+
"mask_token": "<mask>",
|
| 5 |
+
"pad_token": "<pad>",
|
| 6 |
+
"unk_token": "<unk>"
|
| 7 |
+
}
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,54 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<cls>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "<eos>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"32": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"additional_special_tokens": [],
|
| 45 |
+
"clean_up_tokenization_spaces": true,
|
| 46 |
+
"cls_token": "<cls>",
|
| 47 |
+
"eos_token": "<eos>",
|
| 48 |
+
"mask_token": "<mask>",
|
| 49 |
+
"model_max_length": 1024,
|
| 50 |
+
"pad_token": "<pad>",
|
| 51 |
+
"tokenizer_class": "EsmTokenizer",
|
| 52 |
+
"tokenizer_file": null,
|
| 53 |
+
"unk_token": "<unk>"
|
| 54 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1293 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.23276936632271397,
|
| 3 |
+
"best_model_checkpoint": "esm2_t6_8M_finetune_2023-10-08_00-58-24/checkpoint-42015",
|
| 4 |
+
"epoch": 1.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 42015,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.0,
|
| 13 |
+
"learning_rate": 0.0005701249285950224,
|
| 14 |
+
"loss": 0.65,
|
| 15 |
+
"step": 200
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.01,
|
| 19 |
+
"learning_rate": 0.0005700293047710233,
|
| 20 |
+
"loss": 0.5381,
|
| 21 |
+
"step": 400
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"epoch": 0.01,
|
| 25 |
+
"learning_rate": 0.0005698699554923001,
|
| 26 |
+
"loss": 0.5093,
|
| 27 |
+
"step": 600
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 0.02,
|
| 31 |
+
"learning_rate": 0.0005696469163951857,
|
| 32 |
+
"loss": 0.4747,
|
| 33 |
+
"step": 800
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 0.02,
|
| 37 |
+
"learning_rate": 0.0005693602373593884,
|
| 38 |
+
"loss": 0.4731,
|
| 39 |
+
"step": 1000
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.03,
|
| 43 |
+
"learning_rate": 0.0005690099824968371,
|
| 44 |
+
"loss": 0.454,
|
| 45 |
+
"step": 1200
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.03,
|
| 49 |
+
"learning_rate": 0.000568596230137343,
|
| 50 |
+
"loss": 0.4384,
|
| 51 |
+
"step": 1400
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.04,
|
| 55 |
+
"learning_rate": 0.0005681190728110825,
|
| 56 |
+
"loss": 0.4432,
|
| 57 |
+
"step": 1600
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.04,
|
| 61 |
+
"learning_rate": 0.0005675786172279044,
|
| 62 |
+
"loss": 0.4288,
|
| 63 |
+
"step": 1800
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 0.05,
|
| 67 |
+
"learning_rate": 0.0005669749842534648,
|
| 68 |
+
"loss": 0.4278,
|
| 69 |
+
"step": 2000
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.05,
|
| 73 |
+
"learning_rate": 0.000566308308882198,
|
| 74 |
+
"loss": 0.4289,
|
| 75 |
+
"step": 2200
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 0.06,
|
| 79 |
+
"learning_rate": 0.0005655787402071262,
|
| 80 |
+
"loss": 0.4171,
|
| 81 |
+
"step": 2400
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 0.06,
|
| 85 |
+
"learning_rate": 0.0005647864413865174,
|
| 86 |
+
"loss": 0.4262,
|
| 87 |
+
"step": 2600
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.07,
|
| 91 |
+
"learning_rate": 0.0005639315896073965,
|
| 92 |
+
"loss": 0.4155,
|
| 93 |
+
"step": 2800
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.07,
|
| 97 |
+
"learning_rate": 0.0005630143760459205,
|
| 98 |
+
"loss": 0.4042,
|
| 99 |
+
"step": 3000
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.08,
|
| 103 |
+
"learning_rate": 0.000562035005824624,
|
| 104 |
+
"loss": 0.4102,
|
| 105 |
+
"step": 3200
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.08,
|
| 109 |
+
"learning_rate": 0.0005609936979665467,
|
| 110 |
+
"loss": 0.403,
|
| 111 |
+
"step": 3400
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 0.09,
|
| 115 |
+
"learning_rate": 0.0005598906853462516,
|
| 116 |
+
"loss": 0.3957,
|
| 117 |
+
"step": 3600
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 0.09,
|
| 121 |
+
"learning_rate": 0.0005587262146377457,
|
| 122 |
+
"loss": 0.3939,
|
| 123 |
+
"step": 3800
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 0.1,
|
| 127 |
+
"learning_rate": 0.0005575005462593147,
|
| 128 |
+
"loss": 0.4017,
|
| 129 |
+
"step": 4000
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.1,
|
| 133 |
+
"learning_rate": 0.0005562139543152841,
|
| 134 |
+
"loss": 0.3957,
|
| 135 |
+
"step": 4200
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.1,
|
| 139 |
+
"learning_rate": 0.0005548736130124954,
|
| 140 |
+
"loss": 0.3983,
|
| 141 |
+
"step": 4400
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 0.11,
|
| 145 |
+
"learning_rate": 0.0005534663515857835,
|
| 146 |
+
"loss": 0.3824,
|
| 147 |
+
"step": 4600
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 0.11,
|
| 151 |
+
"learning_rate": 0.0005519990687871066,
|
| 152 |
+
"loss": 0.3809,
|
| 153 |
+
"step": 4800
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 0.12,
|
| 157 |
+
"learning_rate": 0.0005504720927546211,
|
| 158 |
+
"loss": 0.3862,
|
| 159 |
+
"step": 5000
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.12,
|
| 163 |
+
"learning_rate": 0.0005488857649760757,
|
| 164 |
+
"loss": 0.3723,
|
| 165 |
+
"step": 5200
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 0.13,
|
| 169 |
+
"learning_rate": 0.0005472404402124425,
|
| 170 |
+
"loss": 0.3708,
|
| 171 |
+
"step": 5400
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.13,
|
| 175 |
+
"learning_rate": 0.0005455364864185798,
|
| 176 |
+
"loss": 0.372,
|
| 177 |
+
"step": 5600
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.14,
|
| 181 |
+
"learning_rate": 0.0005437742846609435,
|
| 182 |
+
"loss": 0.3652,
|
| 183 |
+
"step": 5800
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 0.14,
|
| 187 |
+
"learning_rate": 0.000541954229032367,
|
| 188 |
+
"loss": 0.3534,
|
| 189 |
+
"step": 6000
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 0.15,
|
| 193 |
+
"learning_rate": 0.0005400767265639273,
|
| 194 |
+
"loss": 0.3557,
|
| 195 |
+
"step": 6200
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 0.15,
|
| 199 |
+
"learning_rate": 0.0005381421971339184,
|
| 200 |
+
"loss": 0.3563,
|
| 201 |
+
"step": 6400
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 0.16,
|
| 205 |
+
"learning_rate": 0.0005361510733739505,
|
| 206 |
+
"loss": 0.3641,
|
| 207 |
+
"step": 6600
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 0.16,
|
| 211 |
+
"learning_rate": 0.0005341141758547804,
|
| 212 |
+
"loss": 0.365,
|
| 213 |
+
"step": 6800
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 0.17,
|
| 217 |
+
"learning_rate": 0.0005320114891528482,
|
| 218 |
+
"loss": 0.3507,
|
| 219 |
+
"step": 7000
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.17,
|
| 223 |
+
"learning_rate": 0.000529853579171728,
|
| 224 |
+
"loss": 0.347,
|
| 225 |
+
"step": 7200
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 0.18,
|
| 229 |
+
"learning_rate": 0.0005276409284991003,
|
| 230 |
+
"loss": 0.349,
|
| 231 |
+
"step": 7400
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 0.18,
|
| 235 |
+
"learning_rate": 0.0005253740319646683,
|
| 236 |
+
"loss": 0.3518,
|
| 237 |
+
"step": 7600
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 0.19,
|
| 241 |
+
"learning_rate": 0.0005230533965294961,
|
| 242 |
+
"loss": 0.3329,
|
| 243 |
+
"step": 7800
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 0.19,
|
| 247 |
+
"learning_rate": 0.0005206795411726335,
|
| 248 |
+
"loss": 0.3401,
|
| 249 |
+
"step": 8000
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 0.2,
|
| 253 |
+
"learning_rate": 0.0005182529967750532,
|
| 254 |
+
"loss": 0.3342,
|
| 255 |
+
"step": 8200
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 0.2,
|
| 259 |
+
"learning_rate": 0.000515774306000927,
|
| 260 |
+
"loss": 0.3409,
|
| 261 |
+
"step": 8400
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.2,
|
| 265 |
+
"learning_rate": 0.0005132440231762652,
|
| 266 |
+
"loss": 0.331,
|
| 267 |
+
"step": 8600
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 0.21,
|
| 271 |
+
"learning_rate": 0.00051066271416495,
|
| 272 |
+
"loss": 0.3356,
|
| 273 |
+
"step": 8800
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 0.21,
|
| 277 |
+
"learning_rate": 0.0005080309562421869,
|
| 278 |
+
"loss": 0.3235,
|
| 279 |
+
"step": 9000
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 0.22,
|
| 283 |
+
"learning_rate": 0.0005053493379654053,
|
| 284 |
+
"loss": 0.3292,
|
| 285 |
+
"step": 9200
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 0.22,
|
| 289 |
+
"learning_rate": 0.0005026184590426352,
|
| 290 |
+
"loss": 0.3198,
|
| 291 |
+
"step": 9400
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 0.23,
|
| 295 |
+
"learning_rate": 0.0004998389301983908,
|
| 296 |
+
"loss": 0.3262,
|
| 297 |
+
"step": 9600
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.23,
|
| 301 |
+
"learning_rate": 0.0004970113730370897,
|
| 302 |
+
"loss": 0.3239,
|
| 303 |
+
"step": 9800
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.24,
|
| 307 |
+
"learning_rate": 0.0004941364199040399,
|
| 308 |
+
"loss": 0.3177,
|
| 309 |
+
"step": 10000
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 0.24,
|
| 313 |
+
"learning_rate": 0.0004912147137440239,
|
| 314 |
+
"loss": 0.3103,
|
| 315 |
+
"step": 10200
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"epoch": 0.25,
|
| 319 |
+
"learning_rate": 0.000488246907957513,
|
| 320 |
+
"loss": 0.3169,
|
| 321 |
+
"step": 10400
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 0.25,
|
| 325 |
+
"learning_rate": 0.00048524884437437616,
|
| 326 |
+
"loss": 0.3071,
|
| 327 |
+
"step": 10600
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 0.26,
|
| 331 |
+
"learning_rate": 0.0004822064618785885,
|
| 332 |
+
"loss": 0.3091,
|
| 333 |
+
"step": 10800
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 0.26,
|
| 337 |
+
"learning_rate": 0.00047910481733056954,
|
| 338 |
+
"loss": 0.305,
|
| 339 |
+
"step": 11000
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.27,
|
| 343 |
+
"learning_rate": 0.000475959781372279,
|
| 344 |
+
"loss": 0.3086,
|
| 345 |
+
"step": 11200
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.27,
|
| 349 |
+
"learning_rate": 0.0004727720573489078,
|
| 350 |
+
"loss": 0.3034,
|
| 351 |
+
"step": 11400
|
| 352 |
+
},
|
| 353 |
+
{
|
| 354 |
+
"epoch": 0.28,
|
| 355 |
+
"learning_rate": 0.00046954235815226125,
|
| 356 |
+
"loss": 0.2921,
|
| 357 |
+
"step": 11600
|
| 358 |
+
},
|
| 359 |
+
{
|
| 360 |
+
"epoch": 0.28,
|
| 361 |
+
"learning_rate": 0.00046627140606133027,
|
| 362 |
+
"loss": 0.3043,
|
| 363 |
+
"step": 11800
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 0.29,
|
| 367 |
+
"learning_rate": 0.00046295993258076314,
|
| 368 |
+
"loss": 0.2905,
|
| 369 |
+
"step": 12000
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 0.29,
|
| 373 |
+
"learning_rate": 0.00045960867827727475,
|
| 374 |
+
"loss": 0.2969,
|
| 375 |
+
"step": 12200
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 0.3,
|
| 379 |
+
"learning_rate": 0.00045621839261402846,
|
| 380 |
+
"loss": 0.2938,
|
| 381 |
+
"step": 12400
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 0.3,
|
| 385 |
+
"learning_rate": 0.00045278983378302856,
|
| 386 |
+
"loss": 0.2934,
|
| 387 |
+
"step": 12600
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.3,
|
| 391 |
+
"learning_rate": 0.0004493237685355611,
|
| 392 |
+
"loss": 0.2844,
|
| 393 |
+
"step": 12800
|
| 394 |
+
},
|
| 395 |
+
{
|
| 396 |
+
"epoch": 0.31,
|
| 397 |
+
"learning_rate": 0.00044582097201072027,
|
| 398 |
+
"loss": 0.2882,
|
| 399 |
+
"step": 13000
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 0.31,
|
| 403 |
+
"learning_rate": 0.00044228222756205834,
|
| 404 |
+
"loss": 0.2936,
|
| 405 |
+
"step": 13200
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 0.32,
|
| 409 |
+
"learning_rate": 0.00043870832658239985,
|
| 410 |
+
"loss": 0.2776,
|
| 411 |
+
"step": 13400
|
| 412 |
+
},
|
| 413 |
+
{
|
| 414 |
+
"epoch": 0.32,
|
| 415 |
+
"learning_rate": 0.00043510006832685687,
|
| 416 |
+
"loss": 0.2811,
|
| 417 |
+
"step": 13600
|
| 418 |
+
},
|
| 419 |
+
{
|
| 420 |
+
"epoch": 0.33,
|
| 421 |
+
"learning_rate": 0.0004314582597340859,
|
| 422 |
+
"loss": 0.2821,
|
| 423 |
+
"step": 13800
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 0.33,
|
| 427 |
+
"learning_rate": 0.00042778371524582784,
|
| 428 |
+
"loss": 0.287,
|
| 429 |
+
"step": 14000
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.34,
|
| 433 |
+
"learning_rate": 0.00042407725662476785,
|
| 434 |
+
"loss": 0.2789,
|
| 435 |
+
"step": 14200
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 0.34,
|
| 439 |
+
"learning_rate": 0.00042033971277076005,
|
| 440 |
+
"loss": 0.2763,
|
| 441 |
+
"step": 14400
|
| 442 |
+
},
|
| 443 |
+
{
|
| 444 |
+
"epoch": 0.35,
|
| 445 |
+
"learning_rate": 0.0004165719195354547,
|
| 446 |
+
"loss": 0.2695,
|
| 447 |
+
"step": 14600
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 0.35,
|
| 451 |
+
"learning_rate": 0.00041277471953537104,
|
| 452 |
+
"loss": 0.2635,
|
| 453 |
+
"step": 14800
|
| 454 |
+
},
|
| 455 |
+
{
|
| 456 |
+
"epoch": 0.36,
|
| 457 |
+
"learning_rate": 0.0004089489619634578,
|
| 458 |
+
"loss": 0.2636,
|
| 459 |
+
"step": 15000
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 0.36,
|
| 463 |
+
"learning_rate": 0.0004050955023991823,
|
| 464 |
+
"loss": 0.2632,
|
| 465 |
+
"step": 15200
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 0.37,
|
| 469 |
+
"learning_rate": 0.0004012152026171918,
|
| 470 |
+
"loss": 0.2542,
|
| 471 |
+
"step": 15400
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.37,
|
| 475 |
+
"learning_rate": 0.000397328524919439,
|
| 476 |
+
"loss": 0.2705,
|
| 477 |
+
"step": 15600
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"epoch": 0.38,
|
| 481 |
+
"learning_rate": 0.0003933972771535795,
|
| 482 |
+
"loss": 0.2513,
|
| 483 |
+
"step": 15800
|
| 484 |
+
},
|
| 485 |
+
{
|
| 486 |
+
"epoch": 0.38,
|
| 487 |
+
"learning_rate": 0.0003894418053214825,
|
| 488 |
+
"loss": 0.245,
|
| 489 |
+
"step": 16000
|
| 490 |
+
},
|
| 491 |
+
{
|
| 492 |
+
"epoch": 0.39,
|
| 493 |
+
"learning_rate": 0.00038546299401147295,
|
| 494 |
+
"loss": 0.2578,
|
| 495 |
+
"step": 16200
|
| 496 |
+
},
|
| 497 |
+
{
|
| 498 |
+
"epoch": 0.39,
|
| 499 |
+
"learning_rate": 0.0003814617330314382,
|
| 500 |
+
"loss": 0.2577,
|
| 501 |
+
"step": 16400
|
| 502 |
+
},
|
| 503 |
+
{
|
| 504 |
+
"epoch": 0.4,
|
| 505 |
+
"learning_rate": 0.0003774389172098332,
|
| 506 |
+
"loss": 0.2491,
|
| 507 |
+
"step": 16600
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 0.4,
|
| 511 |
+
"learning_rate": 0.0003733954461955649,
|
| 512 |
+
"loss": 0.2563,
|
| 513 |
+
"step": 16800
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.4,
|
| 517 |
+
"learning_rate": 0.00036933222425679775,
|
| 518 |
+
"loss": 0.2552,
|
| 519 |
+
"step": 17000
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 0.41,
|
| 523 |
+
"learning_rate": 0.00036525016007872573,
|
| 524 |
+
"loss": 0.236,
|
| 525 |
+
"step": 17200
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 0.41,
|
| 529 |
+
"learning_rate": 0.00036115016656035776,
|
| 530 |
+
"loss": 0.2482,
|
| 531 |
+
"step": 17400
|
| 532 |
+
},
|
| 533 |
+
{
|
| 534 |
+
"epoch": 0.42,
|
| 535 |
+
"learning_rate": 0.00035703316061035936,
|
| 536 |
+
"loss": 0.2412,
|
| 537 |
+
"step": 17600
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 0.42,
|
| 541 |
+
"learning_rate": 0.0003529000629419989,
|
| 542 |
+
"loss": 0.2402,
|
| 543 |
+
"step": 17800
|
| 544 |
+
},
|
| 545 |
+
{
|
| 546 |
+
"epoch": 0.43,
|
| 547 |
+
"learning_rate": 0.00034877257538830126,
|
| 548 |
+
"loss": 0.2383,
|
| 549 |
+
"step": 18000
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 0.43,
|
| 553 |
+
"learning_rate": 0.00034461013949659567,
|
| 554 |
+
"loss": 0.2303,
|
| 555 |
+
"step": 18200
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.44,
|
| 559 |
+
"learning_rate": 0.00034043439012889894,
|
| 560 |
+
"loss": 0.2368,
|
| 561 |
+
"step": 18400
|
| 562 |
+
},
|
| 563 |
+
{
|
| 564 |
+
"epoch": 0.44,
|
| 565 |
+
"learning_rate": 0.00033624626113565846,
|
| 566 |
+
"loss": 0.227,
|
| 567 |
+
"step": 18600
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 0.45,
|
| 571 |
+
"learning_rate": 0.0003320466891358589,
|
| 572 |
+
"loss": 0.2321,
|
| 573 |
+
"step": 18800
|
| 574 |
+
},
|
| 575 |
+
{
|
| 576 |
+
"epoch": 0.45,
|
| 577 |
+
"learning_rate": 0.00032783661330756047,
|
| 578 |
+
"loss": 0.2307,
|
| 579 |
+
"step": 19000
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"epoch": 0.46,
|
| 583 |
+
"learning_rate": 0.00032361697517786374,
|
| 584 |
+
"loss": 0.2209,
|
| 585 |
+
"step": 19200
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 0.46,
|
| 589 |
+
"learning_rate": 0.0003193887184123502,
|
| 590 |
+
"loss": 0.2283,
|
| 591 |
+
"step": 19400
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 0.47,
|
| 595 |
+
"learning_rate": 0.00031515278860404374,
|
| 596 |
+
"loss": 0.2231,
|
| 597 |
+
"step": 19600
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.47,
|
| 601 |
+
"learning_rate": 0.00031091013306194153,
|
| 602 |
+
"loss": 0.213,
|
| 603 |
+
"step": 19800
|
| 604 |
+
},
|
| 605 |
+
{
|
| 606 |
+
"epoch": 0.48,
|
| 607 |
+
"learning_rate": 0.0003066617005991611,
|
| 608 |
+
"loss": 0.2227,
|
| 609 |
+
"step": 20000
|
| 610 |
+
},
|
| 611 |
+
{
|
| 612 |
+
"epoch": 0.48,
|
| 613 |
+
"learning_rate": 0.0003024084413207512,
|
| 614 |
+
"loss": 0.2106,
|
| 615 |
+
"step": 20200
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 0.49,
|
| 619 |
+
"learning_rate": 0.0002981726001518289,
|
| 620 |
+
"loss": 0.2156,
|
| 621 |
+
"step": 20400
|
| 622 |
+
},
|
| 623 |
+
{
|
| 624 |
+
"epoch": 0.49,
|
| 625 |
+
"learning_rate": 0.00029391255391089834,
|
| 626 |
+
"loss": 0.2185,
|
| 627 |
+
"step": 20600
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 0.5,
|
| 631 |
+
"learning_rate": 0.0002896505320303282,
|
| 632 |
+
"loss": 0.2133,
|
| 633 |
+
"step": 20800
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 0.5,
|
| 637 |
+
"learning_rate": 0.00028538748765425824,
|
| 638 |
+
"loss": 0.2169,
|
| 639 |
+
"step": 21000
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.5,
|
| 643 |
+
"learning_rate": 0.0002811243741554959,
|
| 644 |
+
"loss": 0.2082,
|
| 645 |
+
"step": 21200
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 0.51,
|
| 649 |
+
"learning_rate": 0.00027686214492230696,
|
| 650 |
+
"loss": 0.2052,
|
| 651 |
+
"step": 21400
|
| 652 |
+
},
|
| 653 |
+
{
|
| 654 |
+
"epoch": 0.51,
|
| 655 |
+
"learning_rate": 0.00027262304895711635,
|
| 656 |
+
"loss": 0.2166,
|
| 657 |
+
"step": 21600
|
| 658 |
+
},
|
| 659 |
+
{
|
| 660 |
+
"epoch": 0.52,
|
| 661 |
+
"learning_rate": 0.00026836543109541225,
|
| 662 |
+
"loss": 0.211,
|
| 663 |
+
"step": 21800
|
| 664 |
+
},
|
| 665 |
+
{
|
| 666 |
+
"epoch": 0.52,
|
| 667 |
+
"learning_rate": 0.00026411155086609875,
|
| 668 |
+
"loss": 0.2014,
|
| 669 |
+
"step": 22000
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"epoch": 0.53,
|
| 673 |
+
"learning_rate": 0.0002598623595925447,
|
| 674 |
+
"loss": 0.2029,
|
| 675 |
+
"step": 22200
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 0.53,
|
| 679 |
+
"learning_rate": 0.0002556188075494962,
|
| 680 |
+
"loss": 0.1988,
|
| 681 |
+
"step": 22400
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.54,
|
| 685 |
+
"learning_rate": 0.00025138184375056156,
|
| 686 |
+
"loss": 0.2003,
|
| 687 |
+
"step": 22600
|
| 688 |
+
},
|
| 689 |
+
{
|
| 690 |
+
"epoch": 0.54,
|
| 691 |
+
"learning_rate": 0.0002471524157359758,
|
| 692 |
+
"loss": 0.197,
|
| 693 |
+
"step": 22800
|
| 694 |
+
},
|
| 695 |
+
{
|
| 696 |
+
"epoch": 0.55,
|
| 697 |
+
"learning_rate": 0.0002429314693606973,
|
| 698 |
+
"loss": 0.1958,
|
| 699 |
+
"step": 23000
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 0.55,
|
| 703 |
+
"learning_rate": 0.00023871994858287913,
|
| 704 |
+
"loss": 0.1848,
|
| 705 |
+
"step": 23200
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"epoch": 0.56,
|
| 709 |
+
"learning_rate": 0.00023451879525276522,
|
| 710 |
+
"loss": 0.1925,
|
| 711 |
+
"step": 23400
|
| 712 |
+
},
|
| 713 |
+
{
|
| 714 |
+
"epoch": 0.56,
|
| 715 |
+
"learning_rate": 0.0002303289489020594,
|
| 716 |
+
"loss": 0.1923,
|
| 717 |
+
"step": 23600
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 0.57,
|
| 721 |
+
"learning_rate": 0.00022615134653381023,
|
| 722 |
+
"loss": 0.191,
|
| 723 |
+
"step": 23800
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.57,
|
| 727 |
+
"learning_rate": 0.00022200771021072758,
|
| 728 |
+
"loss": 0.1906,
|
| 729 |
+
"step": 24000
|
| 730 |
+
},
|
| 731 |
+
{
|
| 732 |
+
"epoch": 0.58,
|
| 733 |
+
"learning_rate": 0.00021785732279544682,
|
| 734 |
+
"loss": 0.1905,
|
| 735 |
+
"step": 24200
|
| 736 |
+
},
|
| 737 |
+
{
|
| 738 |
+
"epoch": 0.58,
|
| 739 |
+
"learning_rate": 0.0002137219684748498,
|
| 740 |
+
"loss": 0.1884,
|
| 741 |
+
"step": 24400
|
| 742 |
+
},
|
| 743 |
+
{
|
| 744 |
+
"epoch": 0.59,
|
| 745 |
+
"learning_rate": 0.00020960257206557216,
|
| 746 |
+
"loss": 0.1844,
|
| 747 |
+
"step": 24600
|
| 748 |
+
},
|
| 749 |
+
{
|
| 750 |
+
"epoch": 0.59,
|
| 751 |
+
"learning_rate": 0.0002055000548154766,
|
| 752 |
+
"loss": 0.1865,
|
| 753 |
+
"step": 24800
|
| 754 |
+
},
|
| 755 |
+
{
|
| 756 |
+
"epoch": 0.6,
|
| 757 |
+
"learning_rate": 0.00020141533419762732,
|
| 758 |
+
"loss": 0.1729,
|
| 759 |
+
"step": 25000
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"epoch": 0.6,
|
| 763 |
+
"learning_rate": 0.00019734932370511055,
|
| 764 |
+
"loss": 0.1717,
|
| 765 |
+
"step": 25200
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.6,
|
| 769 |
+
"learning_rate": 0.00019330293264674297,
|
| 770 |
+
"loss": 0.1797,
|
| 771 |
+
"step": 25400
|
| 772 |
+
},
|
| 773 |
+
{
|
| 774 |
+
"epoch": 0.61,
|
| 775 |
+
"learning_rate": 0.0001892770659437178,
|
| 776 |
+
"loss": 0.1756,
|
| 777 |
+
"step": 25600
|
| 778 |
+
},
|
| 779 |
+
{
|
| 780 |
+
"epoch": 0.61,
|
| 781 |
+
"learning_rate": 0.00018527262392723065,
|
| 782 |
+
"loss": 0.184,
|
| 783 |
+
"step": 25800
|
| 784 |
+
},
|
| 785 |
+
{
|
| 786 |
+
"epoch": 0.62,
|
| 787 |
+
"learning_rate": 0.00018129050213713306,
|
| 788 |
+
"loss": 0.1806,
|
| 789 |
+
"step": 26000
|
| 790 |
+
},
|
| 791 |
+
{
|
| 792 |
+
"epoch": 0.62,
|
| 793 |
+
"learning_rate": 0.00017733159112165648,
|
| 794 |
+
"loss": 0.175,
|
| 795 |
+
"step": 26200
|
| 796 |
+
},
|
| 797 |
+
{
|
| 798 |
+
"epoch": 0.63,
|
| 799 |
+
"learning_rate": 0.00017339677623825358,
|
| 800 |
+
"loss": 0.1746,
|
| 801 |
+
"step": 26400
|
| 802 |
+
},
|
| 803 |
+
{
|
| 804 |
+
"epoch": 0.63,
|
| 805 |
+
"learning_rate": 0.00016948693745559923,
|
| 806 |
+
"loss": 0.1667,
|
| 807 |
+
"step": 26600
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.64,
|
| 811 |
+
"learning_rate": 0.00016560294915679837,
|
| 812 |
+
"loss": 0.1737,
|
| 813 |
+
"step": 26800
|
| 814 |
+
},
|
| 815 |
+
{
|
| 816 |
+
"epoch": 0.64,
|
| 817 |
+
"learning_rate": 0.00016174567994384052,
|
| 818 |
+
"loss": 0.1624,
|
| 819 |
+
"step": 27000
|
| 820 |
+
},
|
| 821 |
+
{
|
| 822 |
+
"epoch": 0.65,
|
| 823 |
+
"learning_rate": 0.00015793507085198864,
|
| 824 |
+
"loss": 0.1667,
|
| 825 |
+
"step": 27200
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 0.65,
|
| 829 |
+
"learning_rate": 0.00015413367721145552,
|
| 830 |
+
"loss": 0.1679,
|
| 831 |
+
"step": 27400
|
| 832 |
+
},
|
| 833 |
+
{
|
| 834 |
+
"epoch": 0.66,
|
| 835 |
+
"learning_rate": 0.00015036156760590242,
|
| 836 |
+
"loss": 0.1626,
|
| 837 |
+
"step": 27600
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 0.66,
|
| 841 |
+
"learning_rate": 0.00014661958561714702,
|
| 842 |
+
"loss": 0.1614,
|
| 843 |
+
"step": 27800
|
| 844 |
+
},
|
| 845 |
+
{
|
| 846 |
+
"epoch": 0.67,
|
| 847 |
+
"learning_rate": 0.00014290856808936924,
|
| 848 |
+
"loss": 0.1603,
|
| 849 |
+
"step": 28000
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.67,
|
| 853 |
+
"learning_rate": 0.0001392293449419613,
|
| 854 |
+
"loss": 0.1611,
|
| 855 |
+
"step": 28200
|
| 856 |
+
},
|
| 857 |
+
{
|
| 858 |
+
"epoch": 0.68,
|
| 859 |
+
"learning_rate": 0.0001355827389839281,
|
| 860 |
+
"loss": 0.156,
|
| 861 |
+
"step": 28400
|
| 862 |
+
},
|
| 863 |
+
{
|
| 864 |
+
"epoch": 0.68,
|
| 865 |
+
"learning_rate": 0.0001319695657298769,
|
| 866 |
+
"loss": 0.1618,
|
| 867 |
+
"step": 28600
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"epoch": 0.69,
|
| 871 |
+
"learning_rate": 0.000128390633217638,
|
| 872 |
+
"loss": 0.1653,
|
| 873 |
+
"step": 28800
|
| 874 |
+
},
|
| 875 |
+
{
|
| 876 |
+
"epoch": 0.69,
|
| 877 |
+
"learning_rate": 0.00012484674182755833,
|
| 878 |
+
"loss": 0.1573,
|
| 879 |
+
"step": 29000
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"epoch": 0.69,
|
| 883 |
+
"learning_rate": 0.00012133868410350671,
|
| 884 |
+
"loss": 0.1629,
|
| 885 |
+
"step": 29200
|
| 886 |
+
},
|
| 887 |
+
{
|
| 888 |
+
"epoch": 0.7,
|
| 889 |
+
"learning_rate": 0.00011786724457563275,
|
| 890 |
+
"loss": 0.1531,
|
| 891 |
+
"step": 29400
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.7,
|
| 895 |
+
"learning_rate": 0.00011443319958491592,
|
| 896 |
+
"loss": 0.1491,
|
| 897 |
+
"step": 29600
|
| 898 |
+
},
|
| 899 |
+
{
|
| 900 |
+
"epoch": 0.71,
|
| 901 |
+
"learning_rate": 0.00011103731710954869,
|
| 902 |
+
"loss": 0.1443,
|
| 903 |
+
"step": 29800
|
| 904 |
+
},
|
| 905 |
+
{
|
| 906 |
+
"epoch": 0.71,
|
| 907 |
+
"learning_rate": 0.00010768035659318755,
|
| 908 |
+
"loss": 0.1454,
|
| 909 |
+
"step": 30000
|
| 910 |
+
},
|
| 911 |
+
{
|
| 912 |
+
"epoch": 0.72,
|
| 913 |
+
"learning_rate": 0.00010436306877511392,
|
| 914 |
+
"loss": 0.1524,
|
| 915 |
+
"step": 30200
|
| 916 |
+
},
|
| 917 |
+
{
|
| 918 |
+
"epoch": 0.72,
|
| 919 |
+
"learning_rate": 0.00010108619552234125,
|
| 920 |
+
"loss": 0.1476,
|
| 921 |
+
"step": 30400
|
| 922 |
+
},
|
| 923 |
+
{
|
| 924 |
+
"epoch": 0.73,
|
| 925 |
+
"learning_rate": 9.78665447379777e-05,
|
| 926 |
+
"loss": 0.1519,
|
| 927 |
+
"step": 30600
|
| 928 |
+
},
|
| 929 |
+
{
|
| 930 |
+
"epoch": 0.73,
|
| 931 |
+
"learning_rate": 9.467247876055098e-05,
|
| 932 |
+
"loss": 0.1421,
|
| 933 |
+
"step": 30800
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.74,
|
| 937 |
+
"learning_rate": 9.152099451915894e-05,
|
| 938 |
+
"loss": 0.1445,
|
| 939 |
+
"step": 31000
|
| 940 |
+
},
|
| 941 |
+
{
|
| 942 |
+
"epoch": 0.74,
|
| 943 |
+
"learning_rate": 8.841279680106426e-05,
|
| 944 |
+
"loss": 0.1378,
|
| 945 |
+
"step": 31200
|
| 946 |
+
},
|
| 947 |
+
{
|
| 948 |
+
"epoch": 0.75,
|
| 949 |
+
"learning_rate": 8.534858071307888e-05,
|
| 950 |
+
"loss": 0.1484,
|
| 951 |
+
"step": 31400
|
| 952 |
+
},
|
| 953 |
+
{
|
| 954 |
+
"epoch": 0.75,
|
| 955 |
+
"learning_rate": 8.232903152611165e-05,
|
| 956 |
+
"loss": 0.1434,
|
| 957 |
+
"step": 31600
|
| 958 |
+
},
|
| 959 |
+
{
|
| 960 |
+
"epoch": 0.76,
|
| 961 |
+
"learning_rate": 7.935482452191686e-05,
|
| 962 |
+
"loss": 0.1342,
|
| 963 |
+
"step": 31800
|
| 964 |
+
},
|
| 965 |
+
{
|
| 966 |
+
"epoch": 0.76,
|
| 967 |
+
"learning_rate": 7.642662484207666e-05,
|
| 968 |
+
"loss": 0.139,
|
| 969 |
+
"step": 32000
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 0.77,
|
| 973 |
+
"learning_rate": 7.354508733925122e-05,
|
| 974 |
+
"loss": 0.1411,
|
| 975 |
+
"step": 32200
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.77,
|
| 979 |
+
"learning_rate": 7.071085643072893e-05,
|
| 980 |
+
"loss": 0.1319,
|
| 981 |
+
"step": 32400
|
| 982 |
+
},
|
| 983 |
+
{
|
| 984 |
+
"epoch": 0.78,
|
| 985 |
+
"learning_rate": 6.792456595431225e-05,
|
| 986 |
+
"loss": 0.1287,
|
| 987 |
+
"step": 32600
|
| 988 |
+
},
|
| 989 |
+
{
|
| 990 |
+
"epoch": 0.78,
|
| 991 |
+
"learning_rate": 6.518683902656733e-05,
|
| 992 |
+
"loss": 0.1273,
|
| 993 |
+
"step": 32800
|
| 994 |
+
},
|
| 995 |
+
{
|
| 996 |
+
"epoch": 0.79,
|
| 997 |
+
"learning_rate": 6.249828790347295e-05,
|
| 998 |
+
"loss": 0.129,
|
| 999 |
+
"step": 33000
|
| 1000 |
+
},
|
| 1001 |
+
{
|
| 1002 |
+
"epoch": 0.79,
|
| 1003 |
+
"learning_rate": 5.9859513843497324e-05,
|
| 1004 |
+
"loss": 0.1326,
|
| 1005 |
+
"step": 33200
|
| 1006 |
+
},
|
| 1007 |
+
{
|
| 1008 |
+
"epoch": 0.79,
|
| 1009 |
+
"learning_rate": 5.727110697313506e-05,
|
| 1010 |
+
"loss": 0.1302,
|
| 1011 |
+
"step": 33400
|
| 1012 |
+
},
|
| 1013 |
+
{
|
| 1014 |
+
"epoch": 0.8,
|
| 1015 |
+
"learning_rate": 5.47336461549329e-05,
|
| 1016 |
+
"loss": 0.1487,
|
| 1017 |
+
"step": 33600
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.8,
|
| 1021 |
+
"learning_rate": 5.224769885803498e-05,
|
| 1022 |
+
"loss": 0.1282,
|
| 1023 |
+
"step": 33800
|
| 1024 |
+
},
|
| 1025 |
+
{
|
| 1026 |
+
"epoch": 0.81,
|
| 1027 |
+
"learning_rate": 4.981382103127605e-05,
|
| 1028 |
+
"loss": 0.1278,
|
| 1029 |
+
"step": 34000
|
| 1030 |
+
},
|
| 1031 |
+
{
|
| 1032 |
+
"epoch": 0.81,
|
| 1033 |
+
"learning_rate": 4.7432556978851023e-05,
|
| 1034 |
+
"loss": 0.13,
|
| 1035 |
+
"step": 34200
|
| 1036 |
+
},
|
| 1037 |
+
{
|
| 1038 |
+
"epoch": 0.82,
|
| 1039 |
+
"learning_rate": 4.51044392385881e-05,
|
| 1040 |
+
"loss": 0.1281,
|
| 1041 |
+
"step": 34400
|
| 1042 |
+
},
|
| 1043 |
+
{
|
| 1044 |
+
"epoch": 0.82,
|
| 1045 |
+
"learning_rate": 4.28412263737391e-05,
|
| 1046 |
+
"loss": 0.1285,
|
| 1047 |
+
"step": 34600
|
| 1048 |
+
},
|
| 1049 |
+
{
|
| 1050 |
+
"epoch": 0.83,
|
| 1051 |
+
"learning_rate": 4.063164624396037e-05,
|
| 1052 |
+
"loss": 0.1288,
|
| 1053 |
+
"step": 34800
|
| 1054 |
+
},
|
| 1055 |
+
{
|
| 1056 |
+
"epoch": 0.83,
|
| 1057 |
+
"learning_rate": 3.846549409375182e-05,
|
| 1058 |
+
"loss": 0.1275,
|
| 1059 |
+
"step": 35000
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.84,
|
| 1063 |
+
"learning_rate": 3.635449361928196e-05,
|
| 1064 |
+
"loss": 0.1249,
|
| 1065 |
+
"step": 35200
|
| 1066 |
+
},
|
| 1067 |
+
{
|
| 1068 |
+
"epoch": 0.84,
|
| 1069 |
+
"learning_rate": 3.42991169175479e-05,
|
| 1070 |
+
"loss": 0.1258,
|
| 1071 |
+
"step": 35400
|
| 1072 |
+
},
|
| 1073 |
+
{
|
| 1074 |
+
"epoch": 0.85,
|
| 1075 |
+
"learning_rate": 3.229982364603461e-05,
|
| 1076 |
+
"loss": 0.1285,
|
| 1077 |
+
"step": 35600
|
| 1078 |
+
},
|
| 1079 |
+
{
|
| 1080 |
+
"epoch": 0.85,
|
| 1081 |
+
"learning_rate": 3.0357060919918606e-05,
|
| 1082 |
+
"loss": 0.1303,
|
| 1083 |
+
"step": 35800
|
| 1084 |
+
},
|
| 1085 |
+
{
|
| 1086 |
+
"epoch": 0.86,
|
| 1087 |
+
"learning_rate": 2.8471263212076872e-05,
|
| 1088 |
+
"loss": 0.1306,
|
| 1089 |
+
"step": 36000
|
| 1090 |
+
},
|
| 1091 |
+
{
|
| 1092 |
+
"epoch": 0.86,
|
| 1093 |
+
"learning_rate": 2.6642852255922315e-05,
|
| 1094 |
+
"loss": 0.1253,
|
| 1095 |
+
"step": 36200
|
| 1096 |
+
},
|
| 1097 |
+
{
|
| 1098 |
+
"epoch": 0.87,
|
| 1099 |
+
"learning_rate": 2.487223695108933e-05,
|
| 1100 |
+
"loss": 0.1175,
|
| 1101 |
+
"step": 36400
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 0.87,
|
| 1105 |
+
"learning_rate": 2.3159813271988392e-05,
|
| 1106 |
+
"loss": 0.1273,
|
| 1107 |
+
"step": 36600
|
| 1108 |
+
},
|
| 1109 |
+
{
|
| 1110 |
+
"epoch": 0.88,
|
| 1111 |
+
"learning_rate": 2.151408710314706e-05,
|
| 1112 |
+
"loss": 0.1316,
|
| 1113 |
+
"step": 36800
|
| 1114 |
+
},
|
| 1115 |
+
{
|
| 1116 |
+
"epoch": 0.88,
|
| 1117 |
+
"learning_rate": 1.9918886837489768e-05,
|
| 1118 |
+
"loss": 0.121,
|
| 1119 |
+
"step": 37000
|
| 1120 |
+
},
|
| 1121 |
+
{
|
| 1122 |
+
"epoch": 0.89,
|
| 1123 |
+
"learning_rate": 1.838298594800929e-05,
|
| 1124 |
+
"loss": 0.1307,
|
| 1125 |
+
"step": 37200
|
| 1126 |
+
},
|
| 1127 |
+
{
|
| 1128 |
+
"epoch": 0.89,
|
| 1129 |
+
"learning_rate": 1.69067279183791e-05,
|
| 1130 |
+
"loss": 0.1169,
|
| 1131 |
+
"step": 37400
|
| 1132 |
+
},
|
| 1133 |
+
{
|
| 1134 |
+
"epoch": 0.89,
|
| 1135 |
+
"learning_rate": 1.549044289394571e-05,
|
| 1136 |
+
"loss": 0.1202,
|
| 1137 |
+
"step": 37600
|
| 1138 |
+
},
|
| 1139 |
+
{
|
| 1140 |
+
"epoch": 0.9,
|
| 1141 |
+
"learning_rate": 1.4134447607895461e-05,
|
| 1142 |
+
"loss": 0.1139,
|
| 1143 |
+
"step": 37800
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 0.9,
|
| 1147 |
+
"learning_rate": 1.2839045310422171e-05,
|
| 1148 |
+
"loss": 0.1289,
|
| 1149 |
+
"step": 38000
|
| 1150 |
+
},
|
| 1151 |
+
{
|
| 1152 |
+
"epoch": 0.91,
|
| 1153 |
+
"learning_rate": 1.1604525700908673e-05,
|
| 1154 |
+
"loss": 0.1333,
|
| 1155 |
+
"step": 38200
|
| 1156 |
+
},
|
| 1157 |
+
{
|
| 1158 |
+
"epoch": 0.91,
|
| 1159 |
+
"learning_rate": 1.0431164863139752e-05,
|
| 1160 |
+
"loss": 0.1165,
|
| 1161 |
+
"step": 38400
|
| 1162 |
+
},
|
| 1163 |
+
{
|
| 1164 |
+
"epoch": 0.92,
|
| 1165 |
+
"learning_rate": 9.319225203559418e-06,
|
| 1166 |
+
"loss": 0.121,
|
| 1167 |
+
"step": 38600
|
| 1168 |
+
},
|
| 1169 |
+
{
|
| 1170 |
+
"epoch": 0.92,
|
| 1171 |
+
"learning_rate": 8.268955392587713e-06,
|
| 1172 |
+
"loss": 0.1265,
|
| 1173 |
+
"step": 38800
|
| 1174 |
+
},
|
| 1175 |
+
{
|
| 1176 |
+
"epoch": 0.93,
|
| 1177 |
+
"learning_rate": 7.2805903090082e-06,
|
| 1178 |
+
"loss": 0.114,
|
| 1179 |
+
"step": 39000
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 0.93,
|
| 1183 |
+
"learning_rate": 6.3543509874411355e-06,
|
| 1184 |
+
"loss": 0.1191,
|
| 1185 |
+
"step": 39200
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 0.94,
|
| 1189 |
+
"learning_rate": 5.4904445689116835e-06,
|
| 1190 |
+
"loss": 0.1085,
|
| 1191 |
+
"step": 39400
|
| 1192 |
+
},
|
| 1193 |
+
{
|
| 1194 |
+
"epoch": 0.94,
|
| 1195 |
+
"learning_rate": 4.689064254525586e-06,
|
| 1196 |
+
"loss": 0.1171,
|
| 1197 |
+
"step": 39600
|
| 1198 |
+
},
|
| 1199 |
+
{
|
| 1200 |
+
"epoch": 0.95,
|
| 1201 |
+
"learning_rate": 3.9503892622624656e-06,
|
| 1202 |
+
"loss": 0.1127,
|
| 1203 |
+
"step": 39800
|
| 1204 |
+
},
|
| 1205 |
+
{
|
| 1206 |
+
"epoch": 0.95,
|
| 1207 |
+
"learning_rate": 3.2745847868960813e-06,
|
| 1208 |
+
"loss": 0.1206,
|
| 1209 |
+
"step": 40000
|
| 1210 |
+
},
|
| 1211 |
+
{
|
| 1212 |
+
"epoch": 0.96,
|
| 1213 |
+
"learning_rate": 2.6618019630506124e-06,
|
| 1214 |
+
"loss": 0.1167,
|
| 1215 |
+
"step": 40200
|
| 1216 |
+
},
|
| 1217 |
+
{
|
| 1218 |
+
"epoch": 0.96,
|
| 1219 |
+
"learning_rate": 2.1121778314016173e-06,
|
| 1220 |
+
"loss": 0.116,
|
| 1221 |
+
"step": 40400
|
| 1222 |
+
},
|
| 1223 |
+
{
|
| 1224 |
+
"epoch": 0.97,
|
| 1225 |
+
"learning_rate": 1.6258353080285549e-06,
|
| 1226 |
+
"loss": 0.1155,
|
| 1227 |
+
"step": 40600
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 0.97,
|
| 1231 |
+
"learning_rate": 1.202883156926388e-06,
|
| 1232 |
+
"loss": 0.1112,
|
| 1233 |
+
"step": 40800
|
| 1234 |
+
},
|
| 1235 |
+
{
|
| 1236 |
+
"epoch": 0.98,
|
| 1237 |
+
"learning_rate": 8.450552439147354e-07,
|
| 1238 |
+
"loss": 0.124,
|
| 1239 |
+
"step": 41000
|
| 1240 |
+
},
|
| 1241 |
+
{
|
| 1242 |
+
"epoch": 0.98,
|
| 1243 |
+
"learning_rate": 5.488353993419233e-07,
|
| 1244 |
+
"loss": 0.1186,
|
| 1245 |
+
"step": 41200
|
| 1246 |
+
},
|
| 1247 |
+
{
|
| 1248 |
+
"epoch": 0.99,
|
| 1249 |
+
"learning_rate": 3.1624678365224445e-07,
|
| 1250 |
+
"loss": 0.1128,
|
| 1251 |
+
"step": 41400
|
| 1252 |
+
},
|
| 1253 |
+
{
|
| 1254 |
+
"epoch": 0.99,
|
| 1255 |
+
"learning_rate": 1.473414121760922e-07,
|
| 1256 |
+
"loss": 0.1269,
|
| 1257 |
+
"step": 41600
|
| 1258 |
+
},
|
| 1259 |
+
{
|
| 1260 |
+
"epoch": 0.99,
|
| 1261 |
+
"learning_rate": 4.215705833867654e-08,
|
| 1262 |
+
"loss": 0.1186,
|
| 1263 |
+
"step": 41800
|
| 1264 |
+
},
|
| 1265 |
+
{
|
| 1266 |
+
"epoch": 1.0,
|
| 1267 |
+
"learning_rate": 7.172452126860464e-10,
|
| 1268 |
+
"loss": 0.1156,
|
| 1269 |
+
"step": 42000
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 1.0,
|
| 1273 |
+
"eval_accuracy": 0.9232681404353718,
|
| 1274 |
+
"eval_auc": 0.7327222270892191,
|
| 1275 |
+
"eval_f1": 0.23276936632271397,
|
| 1276 |
+
"eval_loss": 1.4070135354995728,
|
| 1277 |
+
"eval_mcc": 0.2532717443913893,
|
| 1278 |
+
"eval_precision": 0.14886006058928025,
|
| 1279 |
+
"eval_recall": 0.5334820519181167,
|
| 1280 |
+
"eval_runtime": 1911.9976,
|
| 1281 |
+
"eval_samples_per_second": 66.763,
|
| 1282 |
+
"eval_steps_per_second": 5.564,
|
| 1283 |
+
"step": 42015
|
| 1284 |
+
}
|
| 1285 |
+
],
|
| 1286 |
+
"logging_steps": 200,
|
| 1287 |
+
"max_steps": 42015,
|
| 1288 |
+
"num_train_epochs": 1,
|
| 1289 |
+
"save_steps": 500,
|
| 1290 |
+
"total_flos": 2.0189159728820676e+16,
|
| 1291 |
+
"trial_name": null,
|
| 1292 |
+
"trial_params": null
|
| 1293 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:30d2e37dd745b136deee1518b887f496e20443336e529bbeb49feb47e366d3ff
|
| 3 |
+
size 4091
|
vocab.txt
ADDED
|
@@ -0,0 +1,33 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
<cls>
|
| 2 |
+
<pad>
|
| 3 |
+
<eos>
|
| 4 |
+
<unk>
|
| 5 |
+
L
|
| 6 |
+
A
|
| 7 |
+
G
|
| 8 |
+
V
|
| 9 |
+
S
|
| 10 |
+
E
|
| 11 |
+
R
|
| 12 |
+
T
|
| 13 |
+
I
|
| 14 |
+
D
|
| 15 |
+
P
|
| 16 |
+
K
|
| 17 |
+
Q
|
| 18 |
+
N
|
| 19 |
+
F
|
| 20 |
+
Y
|
| 21 |
+
M
|
| 22 |
+
H
|
| 23 |
+
W
|
| 24 |
+
C
|
| 25 |
+
X
|
| 26 |
+
B
|
| 27 |
+
U
|
| 28 |
+
Z
|
| 29 |
+
O
|
| 30 |
+
.
|
| 31 |
+
-
|
| 32 |
+
<null_1>
|
| 33 |
+
<mask>
|