Add files using upload-large-folder tool
Browse files- ood-vifactcheck-phobert-topic-domain-00-s42-20260814T103527Z-00ca7096/trainer/checkpoint-2805/added_tokens.json +3 -0
- ood-vifactcheck-phobert-topic-domain-00-s42-20260814T103527Z-00ca7096/trainer/checkpoint-2805/bpe.codes +0 -0
- ood-vifactcheck-phobert-topic-domain-00-s42-20260814T103527Z-00ca7096/trainer/checkpoint-2805/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-topic-domain-00-s42-20260814T103527Z-00ca7096/trainer/checkpoint-2805/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-topic-domain-00-s42-20260814T103527Z-00ca7096/trainer/checkpoint-2805/trainer_state.json +992 -0
- ood-vifactcheck-phobert-topic-domain-00-s42-20260814T103527Z-00ca7096/trainer/checkpoint-2805/vocab.txt +0 -0
- ood-vifactcheck-phobert-topic-domain-00-s42-20260814T103527Z-00ca7096/trainer/checkpoint-3366/added_tokens.json +3 -0
- ood-vifactcheck-phobert-topic-domain-00-s42-20260814T103527Z-00ca7096/trainer/checkpoint-3366/bpe.codes +0 -0
- ood-vifactcheck-phobert-topic-domain-00-s42-20260814T103527Z-00ca7096/trainer/checkpoint-3366/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-topic-domain-00-s42-20260814T103527Z-00ca7096/trainer/checkpoint-3366/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-topic-domain-00-s42-20260814T103527Z-00ca7096/trainer/checkpoint-3366/trainer_state.json +1179 -0
- ood-vifactcheck-phobert-topic-domain-00-s42-20260814T103527Z-00ca7096/trainer/checkpoint-3366/vocab.txt +0 -0
- ood-vifactcheck-phobert-topic-domain-01-s42-20260814T105726Z-00ca7096/trainer/checkpoint-1146/added_tokens.json +3 -0
- ood-vifactcheck-phobert-topic-domain-01-s42-20260814T105726Z-00ca7096/trainer/checkpoint-1146/bpe.codes +0 -0
- ood-vifactcheck-phobert-topic-domain-01-s42-20260814T105726Z-00ca7096/trainer/checkpoint-1146/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-topic-domain-01-s42-20260814T105726Z-00ca7096/trainer/checkpoint-1146/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-topic-domain-01-s42-20260814T105726Z-00ca7096/trainer/checkpoint-1146/trainer_state.json +424 -0
- ood-vifactcheck-phobert-topic-domain-01-s42-20260814T105726Z-00ca7096/trainer/checkpoint-1146/vocab.txt +0 -0
- ood-vifactcheck-phobert-topic-domain-01-s42-20260814T105726Z-00ca7096/trainer/checkpoint-573/added_tokens.json +3 -0
- ood-vifactcheck-phobert-topic-domain-01-s42-20260814T105726Z-00ca7096/trainer/checkpoint-573/bpe.codes +0 -0
- ood-vifactcheck-phobert-topic-domain-01-s42-20260814T105726Z-00ca7096/trainer/checkpoint-573/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-topic-domain-01-s42-20260814T105726Z-00ca7096/trainer/checkpoint-573/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-topic-domain-01-s42-20260814T105726Z-00ca7096/trainer/checkpoint-573/trainer_state.json +230 -0
- ood-vifactcheck-phobert-topic-domain-01-s42-20260814T105726Z-00ca7096/trainer/checkpoint-573/vocab.txt +0 -0
- ood-vifactcheck-phobert-topic-domain-02-s42-20260814T111300Z-00ca7096/trainer/checkpoint-2002/added_tokens.json +3 -0
- ood-vifactcheck-phobert-topic-domain-02-s42-20260814T111300Z-00ca7096/trainer/checkpoint-2002/bpe.codes +0 -0
- ood-vifactcheck-phobert-topic-domain-02-s42-20260814T111300Z-00ca7096/trainer/checkpoint-2002/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-topic-domain-02-s42-20260814T111300Z-00ca7096/trainer/checkpoint-2002/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-topic-domain-02-s42-20260814T111300Z-00ca7096/trainer/checkpoint-2002/trainer_state.json +724 -0
- ood-vifactcheck-phobert-topic-domain-02-s42-20260814T111300Z-00ca7096/trainer/checkpoint-2002/vocab.txt +0 -0
- ood-vifactcheck-phobert-topic-domain-02-s42-20260814T111300Z-00ca7096/trainer/checkpoint-2548/added_tokens.json +3 -0
- ood-vifactcheck-phobert-topic-domain-02-s42-20260814T111300Z-00ca7096/trainer/checkpoint-2548/bpe.codes +0 -0
- ood-vifactcheck-phobert-topic-domain-02-s42-20260814T111300Z-00ca7096/trainer/checkpoint-2548/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-topic-domain-02-s42-20260814T111300Z-00ca7096/trainer/checkpoint-2548/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-topic-domain-02-s42-20260814T111300Z-00ca7096/trainer/checkpoint-2548/trainer_state.json +904 -0
- ood-vifactcheck-phobert-topic-domain-02-s42-20260814T111300Z-00ca7096/trainer/checkpoint-2548/vocab.txt +0 -0
- ood-vifactcheck-phobert-topic-domain-03-s42-20260814T112759Z-00ca7096/trainer/checkpoint-1560/added_tokens.json +3 -0
- ood-vifactcheck-phobert-topic-domain-03-s42-20260814T112759Z-00ca7096/trainer/checkpoint-1560/bpe.codes +0 -0
- ood-vifactcheck-phobert-topic-domain-03-s42-20260814T112759Z-00ca7096/trainer/checkpoint-1560/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-topic-domain-03-s42-20260814T112759Z-00ca7096/trainer/checkpoint-1560/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-topic-domain-03-s42-20260814T112759Z-00ca7096/trainer/checkpoint-1560/trainer_state.json +565 -0
- ood-vifactcheck-phobert-topic-domain-03-s42-20260814T112759Z-00ca7096/trainer/checkpoint-1560/vocab.txt +0 -0
- ood-vifactcheck-phobert-topic-domain-03-s42-20260814T112759Z-00ca7096/trainer/checkpoint-2145/added_tokens.json +3 -0
- ood-vifactcheck-phobert-topic-domain-03-s42-20260814T112759Z-00ca7096/trainer/checkpoint-2145/bpe.codes +0 -0
- ood-vifactcheck-phobert-topic-domain-03-s42-20260814T112759Z-00ca7096/trainer/checkpoint-2145/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-topic-domain-03-s42-20260814T112759Z-00ca7096/trainer/checkpoint-2145/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-topic-domain-03-s42-20260814T112759Z-00ca7096/trainer/checkpoint-2145/trainer_state.json +759 -0
- ood-vifactcheck-phobert-topic-domain-03-s42-20260814T112759Z-00ca7096/trainer/checkpoint-2145/vocab.txt +0 -0
- ood-vifactcheck-phobert-topic-domain-04-s42-20260814T114226Z-00ca7096/trainer/checkpoint-752/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-topic-domain-04-s42-20260814T114226Z-00ca7096/trainer/checkpoint-752/tokenizer_config.json +55 -0
ood-vifactcheck-phobert-topic-domain-00-s42-20260814T103527Z-00ca7096/trainer/checkpoint-2805/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-topic-domain-00-s42-20260814T103527Z-00ca7096/trainer/checkpoint-2805/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-00-s42-20260814T103527Z-00ca7096/trainer/checkpoint-2805/special_tokens_map.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": "<mask>",
|
| 6 |
+
"pad_token": "<pad>",
|
| 7 |
+
"sep_token": "</s>",
|
| 8 |
+
"unk_token": "<unk>"
|
| 9 |
+
}
|
ood-vifactcheck-phobert-topic-domain-00-s42-20260814T103527Z-00ca7096/trainer/checkpoint-2805/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 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": "</s>",
|
| 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 |
+
"64000": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 256,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "PhobertTokenizer",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-phobert-topic-domain-00-s42-20260814T103527Z-00ca7096/trainer/checkpoint-2805/trainer_state.json
ADDED
|
@@ -0,0 +1,992 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 2805,
|
| 3 |
+
"best_metric": 0.8658226718172223,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-topic-domain-00-s42-20260814T103527Z-00ca7096/trainer/checkpoint-2805",
|
| 5 |
+
"epoch": 15.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 2805,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.13368983957219252,
|
| 14 |
+
"grad_norm": 7.314456939697266,
|
| 15 |
+
"learning_rate": 1.2834224598930483e-06,
|
| 16 |
+
"loss": 4.426,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.26737967914438504,
|
| 21 |
+
"grad_norm": 7.493458271026611,
|
| 22 |
+
"learning_rate": 2.6203208556149735e-06,
|
| 23 |
+
"loss": 4.3921,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.40106951871657753,
|
| 28 |
+
"grad_norm": 8.1897611618042,
|
| 29 |
+
"learning_rate": 3.957219251336899e-06,
|
| 30 |
+
"loss": 4.3821,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5347593582887701,
|
| 35 |
+
"grad_norm": 8.699569702148438,
|
| 36 |
+
"learning_rate": 5.294117647058824e-06,
|
| 37 |
+
"loss": 4.3957,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6684491978609626,
|
| 42 |
+
"grad_norm": 7.601034164428711,
|
| 43 |
+
"learning_rate": 6.631016042780749e-06,
|
| 44 |
+
"loss": 4.3447,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8021390374331551,
|
| 49 |
+
"grad_norm": 10.49255084991455,
|
| 50 |
+
"learning_rate": 7.967914438502674e-06,
|
| 51 |
+
"loss": 4.258,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9358288770053476,
|
| 56 |
+
"grad_norm": 9.23024845123291,
|
| 57 |
+
"learning_rate": 9.3048128342246e-06,
|
| 58 |
+
"loss": 4.1945,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.513986013986014,
|
| 64 |
+
"eval_loss": 0.9784097671508789,
|
| 65 |
+
"eval_macro_f1": 0.4891629757898957,
|
| 66 |
+
"eval_runtime": 1.1751,
|
| 67 |
+
"eval_samples_per_second": 486.777,
|
| 68 |
+
"eval_steps_per_second": 30.636,
|
| 69 |
+
"eval_weighted_f1": 0.4915975808541056,
|
| 70 |
+
"step": 187
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0695187165775402,
|
| 74 |
+
"grad_norm": 19.759023666381836,
|
| 75 |
+
"learning_rate": 1.0641711229946525e-05,
|
| 76 |
+
"loss": 3.8526,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.2032085561497325,
|
| 81 |
+
"grad_norm": 48.95859909057617,
|
| 82 |
+
"learning_rate": 1.197860962566845e-05,
|
| 83 |
+
"loss": 3.4387,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.3368983957219251,
|
| 88 |
+
"grad_norm": 22.45811653137207,
|
| 89 |
+
"learning_rate": 1.3315508021390375e-05,
|
| 90 |
+
"loss": 3.0088,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.4705882352941178,
|
| 95 |
+
"grad_norm": 34.411354064941406,
|
| 96 |
+
"learning_rate": 1.4652406417112302e-05,
|
| 97 |
+
"loss": 2.7125,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.6042780748663101,
|
| 102 |
+
"grad_norm": 26.857059478759766,
|
| 103 |
+
"learning_rate": 1.5989304812834226e-05,
|
| 104 |
+
"loss": 2.5995,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7379679144385025,
|
| 109 |
+
"grad_norm": 21.47177505493164,
|
| 110 |
+
"learning_rate": 1.7326203208556153e-05,
|
| 111 |
+
"loss": 2.2224,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.8716577540106951,
|
| 116 |
+
"grad_norm": 44.43115234375,
|
| 117 |
+
"learning_rate": 1.8663101604278075e-05,
|
| 118 |
+
"loss": 2.3895,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.8006993006993007,
|
| 124 |
+
"eval_loss": 0.5312798619270325,
|
| 125 |
+
"eval_macro_f1": 0.7993024246641176,
|
| 126 |
+
"eval_runtime": 1.1982,
|
| 127 |
+
"eval_samples_per_second": 477.377,
|
| 128 |
+
"eval_steps_per_second": 30.045,
|
| 129 |
+
"eval_weighted_f1": 0.7990095936053023,
|
| 130 |
+
"step": 374
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.0053475935828877,
|
| 134 |
+
"grad_norm": 38.94443893432617,
|
| 135 |
+
"learning_rate": 2e-05,
|
| 136 |
+
"loss": 2.1817,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.1390374331550803,
|
| 141 |
+
"grad_norm": 33.59406661987305,
|
| 142 |
+
"learning_rate": 1.9851455733808676e-05,
|
| 143 |
+
"loss": 1.6277,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.2727272727272725,
|
| 148 |
+
"grad_norm": 43.400596618652344,
|
| 149 |
+
"learning_rate": 1.9702911467617353e-05,
|
| 150 |
+
"loss": 1.3372,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.406417112299465,
|
| 155 |
+
"grad_norm": 49.194908142089844,
|
| 156 |
+
"learning_rate": 1.9554367201426024e-05,
|
| 157 |
+
"loss": 1.4452,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.5401069518716577,
|
| 162 |
+
"grad_norm": 23.138385772705078,
|
| 163 |
+
"learning_rate": 1.94058229352347e-05,
|
| 164 |
+
"loss": 1.6532,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.6737967914438503,
|
| 169 |
+
"grad_norm": 21.255672454833984,
|
| 170 |
+
"learning_rate": 1.9257278669043376e-05,
|
| 171 |
+
"loss": 1.8128,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.807486631016043,
|
| 176 |
+
"grad_norm": 68.28426361083984,
|
| 177 |
+
"learning_rate": 1.9108734402852053e-05,
|
| 178 |
+
"loss": 1.4082,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 2.9411764705882355,
|
| 183 |
+
"grad_norm": 37.87765884399414,
|
| 184 |
+
"learning_rate": 1.8960190136660727e-05,
|
| 185 |
+
"loss": 1.5379,
|
| 186 |
+
"step": 550
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 3.0,
|
| 190 |
+
"eval_accuracy": 0.8111888111888111,
|
| 191 |
+
"eval_loss": 0.5604568123817444,
|
| 192 |
+
"eval_macro_f1": 0.8103632478632479,
|
| 193 |
+
"eval_runtime": 1.26,
|
| 194 |
+
"eval_samples_per_second": 453.956,
|
| 195 |
+
"eval_steps_per_second": 28.571,
|
| 196 |
+
"eval_weighted_f1": 0.8106733010579164,
|
| 197 |
+
"step": 561
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.0748663101604277,
|
| 201 |
+
"grad_norm": 60.086856842041016,
|
| 202 |
+
"learning_rate": 1.88116458704694e-05,
|
| 203 |
+
"loss": 1.4486,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.2085561497326203,
|
| 208 |
+
"grad_norm": 42.22616195678711,
|
| 209 |
+
"learning_rate": 1.8663101604278075e-05,
|
| 210 |
+
"loss": 1.023,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.342245989304813,
|
| 215 |
+
"grad_norm": 46.136993408203125,
|
| 216 |
+
"learning_rate": 1.8514557338086753e-05,
|
| 217 |
+
"loss": 0.9288,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.4759358288770055,
|
| 222 |
+
"grad_norm": 30.66832733154297,
|
| 223 |
+
"learning_rate": 1.8366013071895427e-05,
|
| 224 |
+
"loss": 0.9943,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.6096256684491976,
|
| 229 |
+
"grad_norm": 34.54519271850586,
|
| 230 |
+
"learning_rate": 1.82174688057041e-05,
|
| 231 |
+
"loss": 0.9114,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.7433155080213902,
|
| 236 |
+
"grad_norm": 34.5118293762207,
|
| 237 |
+
"learning_rate": 1.8068924539512775e-05,
|
| 238 |
+
"loss": 1.0367,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.877005347593583,
|
| 243 |
+
"grad_norm": 36.123531341552734,
|
| 244 |
+
"learning_rate": 1.7920380273321453e-05,
|
| 245 |
+
"loss": 1.0647,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 4.0,
|
| 250 |
+
"eval_accuracy": 0.8356643356643356,
|
| 251 |
+
"eval_loss": 0.49698901176452637,
|
| 252 |
+
"eval_macro_f1": 0.8364840664131773,
|
| 253 |
+
"eval_runtime": 1.1995,
|
| 254 |
+
"eval_samples_per_second": 476.884,
|
| 255 |
+
"eval_steps_per_second": 30.014,
|
| 256 |
+
"eval_weighted_f1": 0.8362474770490761,
|
| 257 |
+
"step": 748
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.010695187165775,
|
| 261 |
+
"grad_norm": 39.13850402832031,
|
| 262 |
+
"learning_rate": 1.7771836007130127e-05,
|
| 263 |
+
"loss": 0.9819,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.144385026737968,
|
| 268 |
+
"grad_norm": 41.74605178833008,
|
| 269 |
+
"learning_rate": 1.76232917409388e-05,
|
| 270 |
+
"loss": 0.6016,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.278074866310161,
|
| 275 |
+
"grad_norm": 43.969852447509766,
|
| 276 |
+
"learning_rate": 1.7474747474747475e-05,
|
| 277 |
+
"loss": 0.6304,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.411764705882353,
|
| 282 |
+
"grad_norm": 157.08010864257812,
|
| 283 |
+
"learning_rate": 1.7326203208556153e-05,
|
| 284 |
+
"loss": 0.8199,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.545454545454545,
|
| 289 |
+
"grad_norm": 3.2522642612457275,
|
| 290 |
+
"learning_rate": 1.7177658942364827e-05,
|
| 291 |
+
"loss": 0.7537,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.6791443850267385,
|
| 296 |
+
"grad_norm": 14.299264907836914,
|
| 297 |
+
"learning_rate": 1.70291146761735e-05,
|
| 298 |
+
"loss": 0.6774,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.81283422459893,
|
| 303 |
+
"grad_norm": 63.34947967529297,
|
| 304 |
+
"learning_rate": 1.6880570409982175e-05,
|
| 305 |
+
"loss": 0.6664,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 4.946524064171123,
|
| 310 |
+
"grad_norm": 7.942429542541504,
|
| 311 |
+
"learning_rate": 1.6732026143790852e-05,
|
| 312 |
+
"loss": 0.5354,
|
| 313 |
+
"step": 925
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 5.0,
|
| 317 |
+
"eval_accuracy": 0.8251748251748252,
|
| 318 |
+
"eval_loss": 0.5805322527885437,
|
| 319 |
+
"eval_macro_f1": 0.8267569416332092,
|
| 320 |
+
"eval_runtime": 1.1759,
|
| 321 |
+
"eval_samples_per_second": 486.438,
|
| 322 |
+
"eval_steps_per_second": 30.615,
|
| 323 |
+
"eval_weighted_f1": 0.8262792721651135,
|
| 324 |
+
"step": 935
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.080213903743315,
|
| 328 |
+
"grad_norm": 71.38935089111328,
|
| 329 |
+
"learning_rate": 1.6583481877599526e-05,
|
| 330 |
+
"loss": 0.804,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.213903743315508,
|
| 335 |
+
"grad_norm": 60.39769744873047,
|
| 336 |
+
"learning_rate": 1.64349376114082e-05,
|
| 337 |
+
"loss": 0.7228,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.347593582887701,
|
| 342 |
+
"grad_norm": 8.537591934204102,
|
| 343 |
+
"learning_rate": 1.6286393345216875e-05,
|
| 344 |
+
"loss": 0.4411,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.481283422459893,
|
| 349 |
+
"grad_norm": 34.465782165527344,
|
| 350 |
+
"learning_rate": 1.6137849079025552e-05,
|
| 351 |
+
"loss": 0.4283,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.614973262032086,
|
| 356 |
+
"grad_norm": 37.07090759277344,
|
| 357 |
+
"learning_rate": 1.5989304812834226e-05,
|
| 358 |
+
"loss": 0.7149,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.748663101604278,
|
| 363 |
+
"grad_norm": 20.84925651550293,
|
| 364 |
+
"learning_rate": 1.58407605466429e-05,
|
| 365 |
+
"loss": 0.4058,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 5.882352941176471,
|
| 370 |
+
"grad_norm": 62.239959716796875,
|
| 371 |
+
"learning_rate": 1.5692216280451574e-05,
|
| 372 |
+
"loss": 0.5234,
|
| 373 |
+
"step": 1100
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 6.0,
|
| 377 |
+
"eval_accuracy": 0.8513986013986014,
|
| 378 |
+
"eval_loss": 0.5905094742774963,
|
| 379 |
+
"eval_macro_f1": 0.8518230504677397,
|
| 380 |
+
"eval_runtime": 1.1783,
|
| 381 |
+
"eval_samples_per_second": 485.44,
|
| 382 |
+
"eval_steps_per_second": 30.552,
|
| 383 |
+
"eval_weighted_f1": 0.8513321555480611,
|
| 384 |
+
"step": 1122
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 6.016042780748663,
|
| 388 |
+
"grad_norm": 5.293822288513184,
|
| 389 |
+
"learning_rate": 1.5543672014260252e-05,
|
| 390 |
+
"loss": 0.4593,
|
| 391 |
+
"step": 1125
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 6.149732620320855,
|
| 395 |
+
"grad_norm": 67.31571960449219,
|
| 396 |
+
"learning_rate": 1.5395127748068926e-05,
|
| 397 |
+
"loss": 0.4125,
|
| 398 |
+
"step": 1150
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.283422459893048,
|
| 402 |
+
"grad_norm": 0.43323785066604614,
|
| 403 |
+
"learning_rate": 1.5246583481877602e-05,
|
| 404 |
+
"loss": 0.4011,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.4171122994652405,
|
| 409 |
+
"grad_norm": 1.9905824661254883,
|
| 410 |
+
"learning_rate": 1.5098039215686276e-05,
|
| 411 |
+
"loss": 0.4706,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.550802139037433,
|
| 416 |
+
"grad_norm": 2.5944104194641113,
|
| 417 |
+
"learning_rate": 1.4949494949494952e-05,
|
| 418 |
+
"loss": 0.2873,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.684491978609626,
|
| 423 |
+
"grad_norm": 78.01982116699219,
|
| 424 |
+
"learning_rate": 1.4800950683303626e-05,
|
| 425 |
+
"loss": 0.4533,
|
| 426 |
+
"step": 1250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 6.818181818181818,
|
| 430 |
+
"grad_norm": 11.068941116333008,
|
| 431 |
+
"learning_rate": 1.4652406417112302e-05,
|
| 432 |
+
"loss": 0.4868,
|
| 433 |
+
"step": 1275
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 6.951871657754011,
|
| 437 |
+
"grad_norm": 56.35197448730469,
|
| 438 |
+
"learning_rate": 1.4503862150920976e-05,
|
| 439 |
+
"loss": 0.4285,
|
| 440 |
+
"step": 1300
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 7.0,
|
| 444 |
+
"eval_accuracy": 0.8479020979020979,
|
| 445 |
+
"eval_loss": 0.719183087348938,
|
| 446 |
+
"eval_macro_f1": 0.8482810941137346,
|
| 447 |
+
"eval_runtime": 1.2092,
|
| 448 |
+
"eval_samples_per_second": 473.054,
|
| 449 |
+
"eval_steps_per_second": 29.773,
|
| 450 |
+
"eval_weighted_f1": 0.8478778489274831,
|
| 451 |
+
"step": 1309
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 7.0855614973262036,
|
| 455 |
+
"grad_norm": 85.26020812988281,
|
| 456 |
+
"learning_rate": 1.4355317884729652e-05,
|
| 457 |
+
"loss": 0.315,
|
| 458 |
+
"step": 1325
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 7.219251336898395,
|
| 462 |
+
"grad_norm": 0.48883774876594543,
|
| 463 |
+
"learning_rate": 1.4206773618538326e-05,
|
| 464 |
+
"loss": 0.2551,
|
| 465 |
+
"step": 1350
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 7.352941176470588,
|
| 469 |
+
"grad_norm": 166.37911987304688,
|
| 470 |
+
"learning_rate": 1.4058229352347002e-05,
|
| 471 |
+
"loss": 0.2635,
|
| 472 |
+
"step": 1375
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 7.4866310160427805,
|
| 476 |
+
"grad_norm": 8.829872131347656,
|
| 477 |
+
"learning_rate": 1.3909685086155676e-05,
|
| 478 |
+
"loss": 0.3023,
|
| 479 |
+
"step": 1400
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 7.620320855614973,
|
| 483 |
+
"grad_norm": 0.3614761531352997,
|
| 484 |
+
"learning_rate": 1.3761140819964351e-05,
|
| 485 |
+
"loss": 0.2544,
|
| 486 |
+
"step": 1425
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 7.754010695187166,
|
| 490 |
+
"grad_norm": 16.760665893554688,
|
| 491 |
+
"learning_rate": 1.3612596553773026e-05,
|
| 492 |
+
"loss": 0.3025,
|
| 493 |
+
"step": 1450
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 7.887700534759358,
|
| 497 |
+
"grad_norm": 55.553958892822266,
|
| 498 |
+
"learning_rate": 1.3464052287581701e-05,
|
| 499 |
+
"loss": 0.23,
|
| 500 |
+
"step": 1475
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 8.0,
|
| 504 |
+
"eval_accuracy": 0.8479020979020979,
|
| 505 |
+
"eval_loss": 0.7581298351287842,
|
| 506 |
+
"eval_macro_f1": 0.8476354181681142,
|
| 507 |
+
"eval_runtime": 1.2452,
|
| 508 |
+
"eval_samples_per_second": 459.346,
|
| 509 |
+
"eval_steps_per_second": 28.91,
|
| 510 |
+
"eval_weighted_f1": 0.8469018968969574,
|
| 511 |
+
"step": 1496
|
| 512 |
+
},
|
| 513 |
+
{
|
| 514 |
+
"epoch": 8.02139037433155,
|
| 515 |
+
"grad_norm": 0.3753194510936737,
|
| 516 |
+
"learning_rate": 1.3315508021390375e-05,
|
| 517 |
+
"loss": 0.2856,
|
| 518 |
+
"step": 1500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 8.155080213903743,
|
| 522 |
+
"grad_norm": 143.26071166992188,
|
| 523 |
+
"learning_rate": 1.3166963755199051e-05,
|
| 524 |
+
"loss": 0.245,
|
| 525 |
+
"step": 1525
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 8.288770053475936,
|
| 529 |
+
"grad_norm": 0.9587932825088501,
|
| 530 |
+
"learning_rate": 1.3018419489007725e-05,
|
| 531 |
+
"loss": 0.1613,
|
| 532 |
+
"step": 1550
|
| 533 |
+
},
|
| 534 |
+
{
|
| 535 |
+
"epoch": 8.422459893048128,
|
| 536 |
+
"grad_norm": 0.20261462032794952,
|
| 537 |
+
"learning_rate": 1.2869875222816401e-05,
|
| 538 |
+
"loss": 0.1345,
|
| 539 |
+
"step": 1575
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 8.556149732620321,
|
| 543 |
+
"grad_norm": 0.36046233773231506,
|
| 544 |
+
"learning_rate": 1.2721330956625075e-05,
|
| 545 |
+
"loss": 0.2666,
|
| 546 |
+
"step": 1600
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 8.689839572192513,
|
| 550 |
+
"grad_norm": 0.18702402710914612,
|
| 551 |
+
"learning_rate": 1.257278669043375e-05,
|
| 552 |
+
"loss": 0.1555,
|
| 553 |
+
"step": 1625
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 8.823529411764707,
|
| 557 |
+
"grad_norm": 117.78833770751953,
|
| 558 |
+
"learning_rate": 1.2424242424242425e-05,
|
| 559 |
+
"loss": 0.3232,
|
| 560 |
+
"step": 1650
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 8.957219251336898,
|
| 564 |
+
"grad_norm": 2.0969324111938477,
|
| 565 |
+
"learning_rate": 1.22756981580511e-05,
|
| 566 |
+
"loss": 0.281,
|
| 567 |
+
"step": 1675
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 9.0,
|
| 571 |
+
"eval_accuracy": 0.8548951048951049,
|
| 572 |
+
"eval_loss": 0.7213868498802185,
|
| 573 |
+
"eval_macro_f1": 0.8560161429709648,
|
| 574 |
+
"eval_runtime": 1.2716,
|
| 575 |
+
"eval_samples_per_second": 449.813,
|
| 576 |
+
"eval_steps_per_second": 28.31,
|
| 577 |
+
"eval_weighted_f1": 0.8551216076125248,
|
| 578 |
+
"step": 1683
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 9.090909090909092,
|
| 582 |
+
"grad_norm": 0.20885683596134186,
|
| 583 |
+
"learning_rate": 1.2127153891859775e-05,
|
| 584 |
+
"loss": 0.277,
|
| 585 |
+
"step": 1700
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 9.224598930481283,
|
| 589 |
+
"grad_norm": 138.6304473876953,
|
| 590 |
+
"learning_rate": 1.197860962566845e-05,
|
| 591 |
+
"loss": 0.2499,
|
| 592 |
+
"step": 1725
|
| 593 |
+
},
|
| 594 |
+
{
|
| 595 |
+
"epoch": 9.358288770053475,
|
| 596 |
+
"grad_norm": 0.8483990430831909,
|
| 597 |
+
"learning_rate": 1.1830065359477125e-05,
|
| 598 |
+
"loss": 0.2044,
|
| 599 |
+
"step": 1750
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 9.491978609625669,
|
| 603 |
+
"grad_norm": 2.4829342365264893,
|
| 604 |
+
"learning_rate": 1.1681521093285799e-05,
|
| 605 |
+
"loss": 0.207,
|
| 606 |
+
"step": 1775
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"epoch": 9.62566844919786,
|
| 610 |
+
"grad_norm": 0.23277647793293,
|
| 611 |
+
"learning_rate": 1.1532976827094475e-05,
|
| 612 |
+
"loss": 0.1672,
|
| 613 |
+
"step": 1800
|
| 614 |
+
},
|
| 615 |
+
{
|
| 616 |
+
"epoch": 9.759358288770054,
|
| 617 |
+
"grad_norm": 0.19339442253112793,
|
| 618 |
+
"learning_rate": 1.1384432560903149e-05,
|
| 619 |
+
"loss": 0.2213,
|
| 620 |
+
"step": 1825
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 9.893048128342246,
|
| 624 |
+
"grad_norm": 0.14063772559165955,
|
| 625 |
+
"learning_rate": 1.1235888294711825e-05,
|
| 626 |
+
"loss": 0.188,
|
| 627 |
+
"step": 1850
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 10.0,
|
| 631 |
+
"eval_accuracy": 0.8583916083916084,
|
| 632 |
+
"eval_loss": 0.7773720622062683,
|
| 633 |
+
"eval_macro_f1": 0.8591427241630178,
|
| 634 |
+
"eval_runtime": 1.1881,
|
| 635 |
+
"eval_samples_per_second": 481.453,
|
| 636 |
+
"eval_steps_per_second": 30.301,
|
| 637 |
+
"eval_weighted_f1": 0.8584397203783107,
|
| 638 |
+
"step": 1870
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 10.026737967914439,
|
| 642 |
+
"grad_norm": 0.1475857049226761,
|
| 643 |
+
"learning_rate": 1.1087344028520499e-05,
|
| 644 |
+
"loss": 0.2442,
|
| 645 |
+
"step": 1875
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 10.16042780748663,
|
| 649 |
+
"grad_norm": 0.19178685545921326,
|
| 650 |
+
"learning_rate": 1.0938799762329175e-05,
|
| 651 |
+
"loss": 0.0769,
|
| 652 |
+
"step": 1900
|
| 653 |
+
},
|
| 654 |
+
{
|
| 655 |
+
"epoch": 10.294117647058824,
|
| 656 |
+
"grad_norm": 125.27078247070312,
|
| 657 |
+
"learning_rate": 1.0790255496137849e-05,
|
| 658 |
+
"loss": 0.0851,
|
| 659 |
+
"step": 1925
|
| 660 |
+
},
|
| 661 |
+
{
|
| 662 |
+
"epoch": 10.427807486631016,
|
| 663 |
+
"grad_norm": 9.923179626464844,
|
| 664 |
+
"learning_rate": 1.0641711229946525e-05,
|
| 665 |
+
"loss": 0.0508,
|
| 666 |
+
"step": 1950
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"epoch": 10.56149732620321,
|
| 670 |
+
"grad_norm": 0.1864607185125351,
|
| 671 |
+
"learning_rate": 1.0493166963755199e-05,
|
| 672 |
+
"loss": 0.1654,
|
| 673 |
+
"step": 1975
|
| 674 |
+
},
|
| 675 |
+
{
|
| 676 |
+
"epoch": 10.695187165775401,
|
| 677 |
+
"grad_norm": 0.13694506883621216,
|
| 678 |
+
"learning_rate": 1.0344622697563875e-05,
|
| 679 |
+
"loss": 0.188,
|
| 680 |
+
"step": 2000
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 10.828877005347593,
|
| 684 |
+
"grad_norm": 0.12389884889125824,
|
| 685 |
+
"learning_rate": 1.0196078431372549e-05,
|
| 686 |
+
"loss": 0.201,
|
| 687 |
+
"step": 2025
|
| 688 |
+
},
|
| 689 |
+
{
|
| 690 |
+
"epoch": 10.962566844919786,
|
| 691 |
+
"grad_norm": 1.6487425565719604,
|
| 692 |
+
"learning_rate": 1.0047534165181224e-05,
|
| 693 |
+
"loss": 0.4032,
|
| 694 |
+
"step": 2050
|
| 695 |
+
},
|
| 696 |
+
{
|
| 697 |
+
"epoch": 11.0,
|
| 698 |
+
"eval_accuracy": 0.8636363636363636,
|
| 699 |
+
"eval_loss": 0.7823806405067444,
|
| 700 |
+
"eval_macro_f1": 0.8625501336898395,
|
| 701 |
+
"eval_runtime": 1.1589,
|
| 702 |
+
"eval_samples_per_second": 493.59,
|
| 703 |
+
"eval_steps_per_second": 31.065,
|
| 704 |
+
"eval_weighted_f1": 0.8622850522917368,
|
| 705 |
+
"step": 2057
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"epoch": 11.096256684491978,
|
| 709 |
+
"grad_norm": 0.11938251554965973,
|
| 710 |
+
"learning_rate": 9.8989898989899e-06,
|
| 711 |
+
"loss": 0.1116,
|
| 712 |
+
"step": 2075
|
| 713 |
+
},
|
| 714 |
+
{
|
| 715 |
+
"epoch": 11.229946524064172,
|
| 716 |
+
"grad_norm": 0.11674272269010544,
|
| 717 |
+
"learning_rate": 9.750445632798576e-06,
|
| 718 |
+
"loss": 0.0692,
|
| 719 |
+
"step": 2100
|
| 720 |
+
},
|
| 721 |
+
{
|
| 722 |
+
"epoch": 11.363636363636363,
|
| 723 |
+
"grad_norm": 0.1071222648024559,
|
| 724 |
+
"learning_rate": 9.60190136660725e-06,
|
| 725 |
+
"loss": 0.0477,
|
| 726 |
+
"step": 2125
|
| 727 |
+
},
|
| 728 |
+
{
|
| 729 |
+
"epoch": 11.497326203208557,
|
| 730 |
+
"grad_norm": 13.00842571258545,
|
| 731 |
+
"learning_rate": 9.453357100415926e-06,
|
| 732 |
+
"loss": 0.157,
|
| 733 |
+
"step": 2150
|
| 734 |
+
},
|
| 735 |
+
{
|
| 736 |
+
"epoch": 11.631016042780749,
|
| 737 |
+
"grad_norm": 0.11801385134458542,
|
| 738 |
+
"learning_rate": 9.3048128342246e-06,
|
| 739 |
+
"loss": 0.1807,
|
| 740 |
+
"step": 2175
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 11.764705882352942,
|
| 744 |
+
"grad_norm": 0.11856313794851303,
|
| 745 |
+
"learning_rate": 9.156268568033274e-06,
|
| 746 |
+
"loss": 0.1546,
|
| 747 |
+
"step": 2200
|
| 748 |
+
},
|
| 749 |
+
{
|
| 750 |
+
"epoch": 11.898395721925134,
|
| 751 |
+
"grad_norm": 0.23650076985359192,
|
| 752 |
+
"learning_rate": 9.007724301841948e-06,
|
| 753 |
+
"loss": 0.121,
|
| 754 |
+
"step": 2225
|
| 755 |
+
},
|
| 756 |
+
{
|
| 757 |
+
"epoch": 12.0,
|
| 758 |
+
"eval_accuracy": 0.8513986013986014,
|
| 759 |
+
"eval_loss": 0.8625929951667786,
|
| 760 |
+
"eval_macro_f1": 0.8520924276325363,
|
| 761 |
+
"eval_runtime": 1.2366,
|
| 762 |
+
"eval_samples_per_second": 462.549,
|
| 763 |
+
"eval_steps_per_second": 29.111,
|
| 764 |
+
"eval_weighted_f1": 0.8513394678349749,
|
| 765 |
+
"step": 2244
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 12.032085561497325,
|
| 769 |
+
"grad_norm": 0.11647474020719528,
|
| 770 |
+
"learning_rate": 8.859180035650624e-06,
|
| 771 |
+
"loss": 0.1459,
|
| 772 |
+
"step": 2250
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 12.165775401069519,
|
| 776 |
+
"grad_norm": 0.09729824960231781,
|
| 777 |
+
"learning_rate": 8.7106357694593e-06,
|
| 778 |
+
"loss": 0.0841,
|
| 779 |
+
"step": 2275
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 12.29946524064171,
|
| 783 |
+
"grad_norm": 89.26582336425781,
|
| 784 |
+
"learning_rate": 8.562091503267974e-06,
|
| 785 |
+
"loss": 0.1326,
|
| 786 |
+
"step": 2300
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 12.433155080213904,
|
| 790 |
+
"grad_norm": 0.34565767645835876,
|
| 791 |
+
"learning_rate": 8.41354723707665e-06,
|
| 792 |
+
"loss": 0.0892,
|
| 793 |
+
"step": 2325
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 12.566844919786096,
|
| 797 |
+
"grad_norm": 0.10946846753358841,
|
| 798 |
+
"learning_rate": 8.265002970885324e-06,
|
| 799 |
+
"loss": 0.079,
|
| 800 |
+
"step": 2350
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 12.70053475935829,
|
| 804 |
+
"grad_norm": 0.08815819025039673,
|
| 805 |
+
"learning_rate": 8.116458704694e-06,
|
| 806 |
+
"loss": 0.0555,
|
| 807 |
+
"step": 2375
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 12.834224598930481,
|
| 811 |
+
"grad_norm": 0.08501690626144409,
|
| 812 |
+
"learning_rate": 7.967914438502674e-06,
|
| 813 |
+
"loss": 0.0625,
|
| 814 |
+
"step": 2400
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 12.967914438502675,
|
| 818 |
+
"grad_norm": 0.0843612551689148,
|
| 819 |
+
"learning_rate": 7.81937017231135e-06,
|
| 820 |
+
"loss": 0.1327,
|
| 821 |
+
"step": 2425
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 13.0,
|
| 825 |
+
"eval_accuracy": 0.8636363636363636,
|
| 826 |
+
"eval_loss": 0.8263055682182312,
|
| 827 |
+
"eval_macro_f1": 0.86499887362837,
|
| 828 |
+
"eval_runtime": 1.2319,
|
| 829 |
+
"eval_samples_per_second": 464.32,
|
| 830 |
+
"eval_steps_per_second": 29.223,
|
| 831 |
+
"eval_weighted_f1": 0.8643929697945394,
|
| 832 |
+
"step": 2431
|
| 833 |
+
},
|
| 834 |
+
{
|
| 835 |
+
"epoch": 13.101604278074866,
|
| 836 |
+
"grad_norm": 0.0848432332277298,
|
| 837 |
+
"learning_rate": 7.670825906120024e-06,
|
| 838 |
+
"loss": 0.0882,
|
| 839 |
+
"step": 2450
|
| 840 |
+
},
|
| 841 |
+
{
|
| 842 |
+
"epoch": 13.235294117647058,
|
| 843 |
+
"grad_norm": 0.9270504117012024,
|
| 844 |
+
"learning_rate": 7.522281639928699e-06,
|
| 845 |
+
"loss": 0.0534,
|
| 846 |
+
"step": 2475
|
| 847 |
+
},
|
| 848 |
+
{
|
| 849 |
+
"epoch": 13.368983957219251,
|
| 850 |
+
"grad_norm": 0.1587558537721634,
|
| 851 |
+
"learning_rate": 7.373737373737374e-06,
|
| 852 |
+
"loss": 0.1242,
|
| 853 |
+
"step": 2500
|
| 854 |
+
},
|
| 855 |
+
{
|
| 856 |
+
"epoch": 13.502673796791443,
|
| 857 |
+
"grad_norm": 0.4047185480594635,
|
| 858 |
+
"learning_rate": 7.2251931075460495e-06,
|
| 859 |
+
"loss": 0.0957,
|
| 860 |
+
"step": 2525
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 13.636363636363637,
|
| 864 |
+
"grad_norm": 0.08490663021802902,
|
| 865 |
+
"learning_rate": 7.0766488413547244e-06,
|
| 866 |
+
"loss": 0.0247,
|
| 867 |
+
"step": 2550
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"epoch": 13.770053475935828,
|
| 871 |
+
"grad_norm": 0.43225857615470886,
|
| 872 |
+
"learning_rate": 6.928104575163399e-06,
|
| 873 |
+
"loss": 0.0287,
|
| 874 |
+
"step": 2575
|
| 875 |
+
},
|
| 876 |
+
{
|
| 877 |
+
"epoch": 13.903743315508022,
|
| 878 |
+
"grad_norm": 0.08818015456199646,
|
| 879 |
+
"learning_rate": 6.779560308972074e-06,
|
| 880 |
+
"loss": 0.1099,
|
| 881 |
+
"step": 2600
|
| 882 |
+
},
|
| 883 |
+
{
|
| 884 |
+
"epoch": 14.0,
|
| 885 |
+
"eval_accuracy": 0.8513986013986014,
|
| 886 |
+
"eval_loss": 0.9054542183876038,
|
| 887 |
+
"eval_macro_f1": 0.8515364038625416,
|
| 888 |
+
"eval_runtime": 1.1902,
|
| 889 |
+
"eval_samples_per_second": 480.604,
|
| 890 |
+
"eval_steps_per_second": 30.248,
|
| 891 |
+
"eval_weighted_f1": 0.8507822307431314,
|
| 892 |
+
"step": 2618
|
| 893 |
+
},
|
| 894 |
+
{
|
| 895 |
+
"epoch": 14.037433155080214,
|
| 896 |
+
"grad_norm": 0.07318533211946487,
|
| 897 |
+
"learning_rate": 6.631016042780749e-06,
|
| 898 |
+
"loss": 0.0676,
|
| 899 |
+
"step": 2625
|
| 900 |
+
},
|
| 901 |
+
{
|
| 902 |
+
"epoch": 14.171122994652407,
|
| 903 |
+
"grad_norm": 0.07690595835447311,
|
| 904 |
+
"learning_rate": 6.482471776589424e-06,
|
| 905 |
+
"loss": 0.0183,
|
| 906 |
+
"step": 2650
|
| 907 |
+
},
|
| 908 |
+
{
|
| 909 |
+
"epoch": 14.304812834224599,
|
| 910 |
+
"grad_norm": 22.4063663482666,
|
| 911 |
+
"learning_rate": 6.333927510398099e-06,
|
| 912 |
+
"loss": 0.0437,
|
| 913 |
+
"step": 2675
|
| 914 |
+
},
|
| 915 |
+
{
|
| 916 |
+
"epoch": 14.43850267379679,
|
| 917 |
+
"grad_norm": 0.07405930757522583,
|
| 918 |
+
"learning_rate": 6.185383244206774e-06,
|
| 919 |
+
"loss": 0.0428,
|
| 920 |
+
"step": 2700
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 14.572192513368984,
|
| 924 |
+
"grad_norm": 0.07327538728713989,
|
| 925 |
+
"learning_rate": 6.036838978015449e-06,
|
| 926 |
+
"loss": 0.0111,
|
| 927 |
+
"step": 2725
|
| 928 |
+
},
|
| 929 |
+
{
|
| 930 |
+
"epoch": 14.705882352941176,
|
| 931 |
+
"grad_norm": 0.07578103244304657,
|
| 932 |
+
"learning_rate": 5.888294711824124e-06,
|
| 933 |
+
"loss": 0.076,
|
| 934 |
+
"step": 2750
|
| 935 |
+
},
|
| 936 |
+
{
|
| 937 |
+
"epoch": 14.83957219251337,
|
| 938 |
+
"grad_norm": 0.08358561992645264,
|
| 939 |
+
"learning_rate": 5.739750445632799e-06,
|
| 940 |
+
"loss": 0.0187,
|
| 941 |
+
"step": 2775
|
| 942 |
+
},
|
| 943 |
+
{
|
| 944 |
+
"epoch": 14.973262032085561,
|
| 945 |
+
"grad_norm": 0.07746005803346634,
|
| 946 |
+
"learning_rate": 5.591206179441474e-06,
|
| 947 |
+
"loss": 0.0135,
|
| 948 |
+
"step": 2800
|
| 949 |
+
},
|
| 950 |
+
{
|
| 951 |
+
"epoch": 15.0,
|
| 952 |
+
"eval_accuracy": 0.8653846153846154,
|
| 953 |
+
"eval_loss": 0.874975860118866,
|
| 954 |
+
"eval_macro_f1": 0.8658226718172223,
|
| 955 |
+
"eval_runtime": 1.2051,
|
| 956 |
+
"eval_samples_per_second": 474.636,
|
| 957 |
+
"eval_steps_per_second": 29.872,
|
| 958 |
+
"eval_weighted_f1": 0.8651976025768445,
|
| 959 |
+
"step": 2805
|
| 960 |
+
}
|
| 961 |
+
],
|
| 962 |
+
"logging_steps": 25,
|
| 963 |
+
"max_steps": 3740,
|
| 964 |
+
"num_input_tokens_seen": 0,
|
| 965 |
+
"num_train_epochs": 20,
|
| 966 |
+
"save_steps": 500,
|
| 967 |
+
"stateful_callbacks": {
|
| 968 |
+
"EarlyStoppingCallback": {
|
| 969 |
+
"args": {
|
| 970 |
+
"early_stopping_patience": 3,
|
| 971 |
+
"early_stopping_threshold": 0.0
|
| 972 |
+
},
|
| 973 |
+
"attributes": {
|
| 974 |
+
"early_stopping_patience_counter": 0
|
| 975 |
+
}
|
| 976 |
+
},
|
| 977 |
+
"TrainerControl": {
|
| 978 |
+
"args": {
|
| 979 |
+
"should_epoch_stop": false,
|
| 980 |
+
"should_evaluate": false,
|
| 981 |
+
"should_log": false,
|
| 982 |
+
"should_save": true,
|
| 983 |
+
"should_training_stop": false
|
| 984 |
+
},
|
| 985 |
+
"attributes": {}
|
| 986 |
+
}
|
| 987 |
+
},
|
| 988 |
+
"total_flos": 0.0,
|
| 989 |
+
"train_batch_size": 8,
|
| 990 |
+
"trial_name": null,
|
| 991 |
+
"trial_params": null
|
| 992 |
+
}
|
ood-vifactcheck-phobert-topic-domain-00-s42-20260814T103527Z-00ca7096/trainer/checkpoint-2805/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-00-s42-20260814T103527Z-00ca7096/trainer/checkpoint-3366/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-topic-domain-00-s42-20260814T103527Z-00ca7096/trainer/checkpoint-3366/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-00-s42-20260814T103527Z-00ca7096/trainer/checkpoint-3366/special_tokens_map.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": "<mask>",
|
| 6 |
+
"pad_token": "<pad>",
|
| 7 |
+
"sep_token": "</s>",
|
| 8 |
+
"unk_token": "<unk>"
|
| 9 |
+
}
|
ood-vifactcheck-phobert-topic-domain-00-s42-20260814T103527Z-00ca7096/trainer/checkpoint-3366/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 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": "</s>",
|
| 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 |
+
"64000": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 256,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "PhobertTokenizer",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-phobert-topic-domain-00-s42-20260814T103527Z-00ca7096/trainer/checkpoint-3366/trainer_state.json
ADDED
|
@@ -0,0 +1,1179 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 2805,
|
| 3 |
+
"best_metric": 0.8658226718172223,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-topic-domain-00-s42-20260814T103527Z-00ca7096/trainer/checkpoint-2805",
|
| 5 |
+
"epoch": 18.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 3366,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.13368983957219252,
|
| 14 |
+
"grad_norm": 7.314456939697266,
|
| 15 |
+
"learning_rate": 1.2834224598930483e-06,
|
| 16 |
+
"loss": 4.426,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.26737967914438504,
|
| 21 |
+
"grad_norm": 7.493458271026611,
|
| 22 |
+
"learning_rate": 2.6203208556149735e-06,
|
| 23 |
+
"loss": 4.3921,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.40106951871657753,
|
| 28 |
+
"grad_norm": 8.1897611618042,
|
| 29 |
+
"learning_rate": 3.957219251336899e-06,
|
| 30 |
+
"loss": 4.3821,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5347593582887701,
|
| 35 |
+
"grad_norm": 8.699569702148438,
|
| 36 |
+
"learning_rate": 5.294117647058824e-06,
|
| 37 |
+
"loss": 4.3957,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6684491978609626,
|
| 42 |
+
"grad_norm": 7.601034164428711,
|
| 43 |
+
"learning_rate": 6.631016042780749e-06,
|
| 44 |
+
"loss": 4.3447,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8021390374331551,
|
| 49 |
+
"grad_norm": 10.49255084991455,
|
| 50 |
+
"learning_rate": 7.967914438502674e-06,
|
| 51 |
+
"loss": 4.258,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9358288770053476,
|
| 56 |
+
"grad_norm": 9.23024845123291,
|
| 57 |
+
"learning_rate": 9.3048128342246e-06,
|
| 58 |
+
"loss": 4.1945,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.513986013986014,
|
| 64 |
+
"eval_loss": 0.9784097671508789,
|
| 65 |
+
"eval_macro_f1": 0.4891629757898957,
|
| 66 |
+
"eval_runtime": 1.1751,
|
| 67 |
+
"eval_samples_per_second": 486.777,
|
| 68 |
+
"eval_steps_per_second": 30.636,
|
| 69 |
+
"eval_weighted_f1": 0.4915975808541056,
|
| 70 |
+
"step": 187
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0695187165775402,
|
| 74 |
+
"grad_norm": 19.759023666381836,
|
| 75 |
+
"learning_rate": 1.0641711229946525e-05,
|
| 76 |
+
"loss": 3.8526,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.2032085561497325,
|
| 81 |
+
"grad_norm": 48.95859909057617,
|
| 82 |
+
"learning_rate": 1.197860962566845e-05,
|
| 83 |
+
"loss": 3.4387,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.3368983957219251,
|
| 88 |
+
"grad_norm": 22.45811653137207,
|
| 89 |
+
"learning_rate": 1.3315508021390375e-05,
|
| 90 |
+
"loss": 3.0088,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.4705882352941178,
|
| 95 |
+
"grad_norm": 34.411354064941406,
|
| 96 |
+
"learning_rate": 1.4652406417112302e-05,
|
| 97 |
+
"loss": 2.7125,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.6042780748663101,
|
| 102 |
+
"grad_norm": 26.857059478759766,
|
| 103 |
+
"learning_rate": 1.5989304812834226e-05,
|
| 104 |
+
"loss": 2.5995,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7379679144385025,
|
| 109 |
+
"grad_norm": 21.47177505493164,
|
| 110 |
+
"learning_rate": 1.7326203208556153e-05,
|
| 111 |
+
"loss": 2.2224,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.8716577540106951,
|
| 116 |
+
"grad_norm": 44.43115234375,
|
| 117 |
+
"learning_rate": 1.8663101604278075e-05,
|
| 118 |
+
"loss": 2.3895,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.8006993006993007,
|
| 124 |
+
"eval_loss": 0.5312798619270325,
|
| 125 |
+
"eval_macro_f1": 0.7993024246641176,
|
| 126 |
+
"eval_runtime": 1.1982,
|
| 127 |
+
"eval_samples_per_second": 477.377,
|
| 128 |
+
"eval_steps_per_second": 30.045,
|
| 129 |
+
"eval_weighted_f1": 0.7990095936053023,
|
| 130 |
+
"step": 374
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.0053475935828877,
|
| 134 |
+
"grad_norm": 38.94443893432617,
|
| 135 |
+
"learning_rate": 2e-05,
|
| 136 |
+
"loss": 2.1817,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.1390374331550803,
|
| 141 |
+
"grad_norm": 33.59406661987305,
|
| 142 |
+
"learning_rate": 1.9851455733808676e-05,
|
| 143 |
+
"loss": 1.6277,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.2727272727272725,
|
| 148 |
+
"grad_norm": 43.400596618652344,
|
| 149 |
+
"learning_rate": 1.9702911467617353e-05,
|
| 150 |
+
"loss": 1.3372,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.406417112299465,
|
| 155 |
+
"grad_norm": 49.194908142089844,
|
| 156 |
+
"learning_rate": 1.9554367201426024e-05,
|
| 157 |
+
"loss": 1.4452,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.5401069518716577,
|
| 162 |
+
"grad_norm": 23.138385772705078,
|
| 163 |
+
"learning_rate": 1.94058229352347e-05,
|
| 164 |
+
"loss": 1.6532,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.6737967914438503,
|
| 169 |
+
"grad_norm": 21.255672454833984,
|
| 170 |
+
"learning_rate": 1.9257278669043376e-05,
|
| 171 |
+
"loss": 1.8128,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.807486631016043,
|
| 176 |
+
"grad_norm": 68.28426361083984,
|
| 177 |
+
"learning_rate": 1.9108734402852053e-05,
|
| 178 |
+
"loss": 1.4082,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 2.9411764705882355,
|
| 183 |
+
"grad_norm": 37.87765884399414,
|
| 184 |
+
"learning_rate": 1.8960190136660727e-05,
|
| 185 |
+
"loss": 1.5379,
|
| 186 |
+
"step": 550
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 3.0,
|
| 190 |
+
"eval_accuracy": 0.8111888111888111,
|
| 191 |
+
"eval_loss": 0.5604568123817444,
|
| 192 |
+
"eval_macro_f1": 0.8103632478632479,
|
| 193 |
+
"eval_runtime": 1.26,
|
| 194 |
+
"eval_samples_per_second": 453.956,
|
| 195 |
+
"eval_steps_per_second": 28.571,
|
| 196 |
+
"eval_weighted_f1": 0.8106733010579164,
|
| 197 |
+
"step": 561
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.0748663101604277,
|
| 201 |
+
"grad_norm": 60.086856842041016,
|
| 202 |
+
"learning_rate": 1.88116458704694e-05,
|
| 203 |
+
"loss": 1.4486,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.2085561497326203,
|
| 208 |
+
"grad_norm": 42.22616195678711,
|
| 209 |
+
"learning_rate": 1.8663101604278075e-05,
|
| 210 |
+
"loss": 1.023,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.342245989304813,
|
| 215 |
+
"grad_norm": 46.136993408203125,
|
| 216 |
+
"learning_rate": 1.8514557338086753e-05,
|
| 217 |
+
"loss": 0.9288,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.4759358288770055,
|
| 222 |
+
"grad_norm": 30.66832733154297,
|
| 223 |
+
"learning_rate": 1.8366013071895427e-05,
|
| 224 |
+
"loss": 0.9943,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.6096256684491976,
|
| 229 |
+
"grad_norm": 34.54519271850586,
|
| 230 |
+
"learning_rate": 1.82174688057041e-05,
|
| 231 |
+
"loss": 0.9114,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.7433155080213902,
|
| 236 |
+
"grad_norm": 34.5118293762207,
|
| 237 |
+
"learning_rate": 1.8068924539512775e-05,
|
| 238 |
+
"loss": 1.0367,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.877005347593583,
|
| 243 |
+
"grad_norm": 36.123531341552734,
|
| 244 |
+
"learning_rate": 1.7920380273321453e-05,
|
| 245 |
+
"loss": 1.0647,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 4.0,
|
| 250 |
+
"eval_accuracy": 0.8356643356643356,
|
| 251 |
+
"eval_loss": 0.49698901176452637,
|
| 252 |
+
"eval_macro_f1": 0.8364840664131773,
|
| 253 |
+
"eval_runtime": 1.1995,
|
| 254 |
+
"eval_samples_per_second": 476.884,
|
| 255 |
+
"eval_steps_per_second": 30.014,
|
| 256 |
+
"eval_weighted_f1": 0.8362474770490761,
|
| 257 |
+
"step": 748
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.010695187165775,
|
| 261 |
+
"grad_norm": 39.13850402832031,
|
| 262 |
+
"learning_rate": 1.7771836007130127e-05,
|
| 263 |
+
"loss": 0.9819,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.144385026737968,
|
| 268 |
+
"grad_norm": 41.74605178833008,
|
| 269 |
+
"learning_rate": 1.76232917409388e-05,
|
| 270 |
+
"loss": 0.6016,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.278074866310161,
|
| 275 |
+
"grad_norm": 43.969852447509766,
|
| 276 |
+
"learning_rate": 1.7474747474747475e-05,
|
| 277 |
+
"loss": 0.6304,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.411764705882353,
|
| 282 |
+
"grad_norm": 157.08010864257812,
|
| 283 |
+
"learning_rate": 1.7326203208556153e-05,
|
| 284 |
+
"loss": 0.8199,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.545454545454545,
|
| 289 |
+
"grad_norm": 3.2522642612457275,
|
| 290 |
+
"learning_rate": 1.7177658942364827e-05,
|
| 291 |
+
"loss": 0.7537,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.6791443850267385,
|
| 296 |
+
"grad_norm": 14.299264907836914,
|
| 297 |
+
"learning_rate": 1.70291146761735e-05,
|
| 298 |
+
"loss": 0.6774,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.81283422459893,
|
| 303 |
+
"grad_norm": 63.34947967529297,
|
| 304 |
+
"learning_rate": 1.6880570409982175e-05,
|
| 305 |
+
"loss": 0.6664,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 4.946524064171123,
|
| 310 |
+
"grad_norm": 7.942429542541504,
|
| 311 |
+
"learning_rate": 1.6732026143790852e-05,
|
| 312 |
+
"loss": 0.5354,
|
| 313 |
+
"step": 925
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 5.0,
|
| 317 |
+
"eval_accuracy": 0.8251748251748252,
|
| 318 |
+
"eval_loss": 0.5805322527885437,
|
| 319 |
+
"eval_macro_f1": 0.8267569416332092,
|
| 320 |
+
"eval_runtime": 1.1759,
|
| 321 |
+
"eval_samples_per_second": 486.438,
|
| 322 |
+
"eval_steps_per_second": 30.615,
|
| 323 |
+
"eval_weighted_f1": 0.8262792721651135,
|
| 324 |
+
"step": 935
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.080213903743315,
|
| 328 |
+
"grad_norm": 71.38935089111328,
|
| 329 |
+
"learning_rate": 1.6583481877599526e-05,
|
| 330 |
+
"loss": 0.804,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.213903743315508,
|
| 335 |
+
"grad_norm": 60.39769744873047,
|
| 336 |
+
"learning_rate": 1.64349376114082e-05,
|
| 337 |
+
"loss": 0.7228,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.347593582887701,
|
| 342 |
+
"grad_norm": 8.537591934204102,
|
| 343 |
+
"learning_rate": 1.6286393345216875e-05,
|
| 344 |
+
"loss": 0.4411,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.481283422459893,
|
| 349 |
+
"grad_norm": 34.465782165527344,
|
| 350 |
+
"learning_rate": 1.6137849079025552e-05,
|
| 351 |
+
"loss": 0.4283,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.614973262032086,
|
| 356 |
+
"grad_norm": 37.07090759277344,
|
| 357 |
+
"learning_rate": 1.5989304812834226e-05,
|
| 358 |
+
"loss": 0.7149,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.748663101604278,
|
| 363 |
+
"grad_norm": 20.84925651550293,
|
| 364 |
+
"learning_rate": 1.58407605466429e-05,
|
| 365 |
+
"loss": 0.4058,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 5.882352941176471,
|
| 370 |
+
"grad_norm": 62.239959716796875,
|
| 371 |
+
"learning_rate": 1.5692216280451574e-05,
|
| 372 |
+
"loss": 0.5234,
|
| 373 |
+
"step": 1100
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 6.0,
|
| 377 |
+
"eval_accuracy": 0.8513986013986014,
|
| 378 |
+
"eval_loss": 0.5905094742774963,
|
| 379 |
+
"eval_macro_f1": 0.8518230504677397,
|
| 380 |
+
"eval_runtime": 1.1783,
|
| 381 |
+
"eval_samples_per_second": 485.44,
|
| 382 |
+
"eval_steps_per_second": 30.552,
|
| 383 |
+
"eval_weighted_f1": 0.8513321555480611,
|
| 384 |
+
"step": 1122
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 6.016042780748663,
|
| 388 |
+
"grad_norm": 5.293822288513184,
|
| 389 |
+
"learning_rate": 1.5543672014260252e-05,
|
| 390 |
+
"loss": 0.4593,
|
| 391 |
+
"step": 1125
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 6.149732620320855,
|
| 395 |
+
"grad_norm": 67.31571960449219,
|
| 396 |
+
"learning_rate": 1.5395127748068926e-05,
|
| 397 |
+
"loss": 0.4125,
|
| 398 |
+
"step": 1150
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.283422459893048,
|
| 402 |
+
"grad_norm": 0.43323785066604614,
|
| 403 |
+
"learning_rate": 1.5246583481877602e-05,
|
| 404 |
+
"loss": 0.4011,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.4171122994652405,
|
| 409 |
+
"grad_norm": 1.9905824661254883,
|
| 410 |
+
"learning_rate": 1.5098039215686276e-05,
|
| 411 |
+
"loss": 0.4706,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.550802139037433,
|
| 416 |
+
"grad_norm": 2.5944104194641113,
|
| 417 |
+
"learning_rate": 1.4949494949494952e-05,
|
| 418 |
+
"loss": 0.2873,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.684491978609626,
|
| 423 |
+
"grad_norm": 78.01982116699219,
|
| 424 |
+
"learning_rate": 1.4800950683303626e-05,
|
| 425 |
+
"loss": 0.4533,
|
| 426 |
+
"step": 1250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 6.818181818181818,
|
| 430 |
+
"grad_norm": 11.068941116333008,
|
| 431 |
+
"learning_rate": 1.4652406417112302e-05,
|
| 432 |
+
"loss": 0.4868,
|
| 433 |
+
"step": 1275
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 6.951871657754011,
|
| 437 |
+
"grad_norm": 56.35197448730469,
|
| 438 |
+
"learning_rate": 1.4503862150920976e-05,
|
| 439 |
+
"loss": 0.4285,
|
| 440 |
+
"step": 1300
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 7.0,
|
| 444 |
+
"eval_accuracy": 0.8479020979020979,
|
| 445 |
+
"eval_loss": 0.719183087348938,
|
| 446 |
+
"eval_macro_f1": 0.8482810941137346,
|
| 447 |
+
"eval_runtime": 1.2092,
|
| 448 |
+
"eval_samples_per_second": 473.054,
|
| 449 |
+
"eval_steps_per_second": 29.773,
|
| 450 |
+
"eval_weighted_f1": 0.8478778489274831,
|
| 451 |
+
"step": 1309
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 7.0855614973262036,
|
| 455 |
+
"grad_norm": 85.26020812988281,
|
| 456 |
+
"learning_rate": 1.4355317884729652e-05,
|
| 457 |
+
"loss": 0.315,
|
| 458 |
+
"step": 1325
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 7.219251336898395,
|
| 462 |
+
"grad_norm": 0.48883774876594543,
|
| 463 |
+
"learning_rate": 1.4206773618538326e-05,
|
| 464 |
+
"loss": 0.2551,
|
| 465 |
+
"step": 1350
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 7.352941176470588,
|
| 469 |
+
"grad_norm": 166.37911987304688,
|
| 470 |
+
"learning_rate": 1.4058229352347002e-05,
|
| 471 |
+
"loss": 0.2635,
|
| 472 |
+
"step": 1375
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 7.4866310160427805,
|
| 476 |
+
"grad_norm": 8.829872131347656,
|
| 477 |
+
"learning_rate": 1.3909685086155676e-05,
|
| 478 |
+
"loss": 0.3023,
|
| 479 |
+
"step": 1400
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 7.620320855614973,
|
| 483 |
+
"grad_norm": 0.3614761531352997,
|
| 484 |
+
"learning_rate": 1.3761140819964351e-05,
|
| 485 |
+
"loss": 0.2544,
|
| 486 |
+
"step": 1425
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 7.754010695187166,
|
| 490 |
+
"grad_norm": 16.760665893554688,
|
| 491 |
+
"learning_rate": 1.3612596553773026e-05,
|
| 492 |
+
"loss": 0.3025,
|
| 493 |
+
"step": 1450
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 7.887700534759358,
|
| 497 |
+
"grad_norm": 55.553958892822266,
|
| 498 |
+
"learning_rate": 1.3464052287581701e-05,
|
| 499 |
+
"loss": 0.23,
|
| 500 |
+
"step": 1475
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 8.0,
|
| 504 |
+
"eval_accuracy": 0.8479020979020979,
|
| 505 |
+
"eval_loss": 0.7581298351287842,
|
| 506 |
+
"eval_macro_f1": 0.8476354181681142,
|
| 507 |
+
"eval_runtime": 1.2452,
|
| 508 |
+
"eval_samples_per_second": 459.346,
|
| 509 |
+
"eval_steps_per_second": 28.91,
|
| 510 |
+
"eval_weighted_f1": 0.8469018968969574,
|
| 511 |
+
"step": 1496
|
| 512 |
+
},
|
| 513 |
+
{
|
| 514 |
+
"epoch": 8.02139037433155,
|
| 515 |
+
"grad_norm": 0.3753194510936737,
|
| 516 |
+
"learning_rate": 1.3315508021390375e-05,
|
| 517 |
+
"loss": 0.2856,
|
| 518 |
+
"step": 1500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 8.155080213903743,
|
| 522 |
+
"grad_norm": 143.26071166992188,
|
| 523 |
+
"learning_rate": 1.3166963755199051e-05,
|
| 524 |
+
"loss": 0.245,
|
| 525 |
+
"step": 1525
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 8.288770053475936,
|
| 529 |
+
"grad_norm": 0.9587932825088501,
|
| 530 |
+
"learning_rate": 1.3018419489007725e-05,
|
| 531 |
+
"loss": 0.1613,
|
| 532 |
+
"step": 1550
|
| 533 |
+
},
|
| 534 |
+
{
|
| 535 |
+
"epoch": 8.422459893048128,
|
| 536 |
+
"grad_norm": 0.20261462032794952,
|
| 537 |
+
"learning_rate": 1.2869875222816401e-05,
|
| 538 |
+
"loss": 0.1345,
|
| 539 |
+
"step": 1575
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 8.556149732620321,
|
| 543 |
+
"grad_norm": 0.36046233773231506,
|
| 544 |
+
"learning_rate": 1.2721330956625075e-05,
|
| 545 |
+
"loss": 0.2666,
|
| 546 |
+
"step": 1600
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 8.689839572192513,
|
| 550 |
+
"grad_norm": 0.18702402710914612,
|
| 551 |
+
"learning_rate": 1.257278669043375e-05,
|
| 552 |
+
"loss": 0.1555,
|
| 553 |
+
"step": 1625
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 8.823529411764707,
|
| 557 |
+
"grad_norm": 117.78833770751953,
|
| 558 |
+
"learning_rate": 1.2424242424242425e-05,
|
| 559 |
+
"loss": 0.3232,
|
| 560 |
+
"step": 1650
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 8.957219251336898,
|
| 564 |
+
"grad_norm": 2.0969324111938477,
|
| 565 |
+
"learning_rate": 1.22756981580511e-05,
|
| 566 |
+
"loss": 0.281,
|
| 567 |
+
"step": 1675
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 9.0,
|
| 571 |
+
"eval_accuracy": 0.8548951048951049,
|
| 572 |
+
"eval_loss": 0.7213868498802185,
|
| 573 |
+
"eval_macro_f1": 0.8560161429709648,
|
| 574 |
+
"eval_runtime": 1.2716,
|
| 575 |
+
"eval_samples_per_second": 449.813,
|
| 576 |
+
"eval_steps_per_second": 28.31,
|
| 577 |
+
"eval_weighted_f1": 0.8551216076125248,
|
| 578 |
+
"step": 1683
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 9.090909090909092,
|
| 582 |
+
"grad_norm": 0.20885683596134186,
|
| 583 |
+
"learning_rate": 1.2127153891859775e-05,
|
| 584 |
+
"loss": 0.277,
|
| 585 |
+
"step": 1700
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 9.224598930481283,
|
| 589 |
+
"grad_norm": 138.6304473876953,
|
| 590 |
+
"learning_rate": 1.197860962566845e-05,
|
| 591 |
+
"loss": 0.2499,
|
| 592 |
+
"step": 1725
|
| 593 |
+
},
|
| 594 |
+
{
|
| 595 |
+
"epoch": 9.358288770053475,
|
| 596 |
+
"grad_norm": 0.8483990430831909,
|
| 597 |
+
"learning_rate": 1.1830065359477125e-05,
|
| 598 |
+
"loss": 0.2044,
|
| 599 |
+
"step": 1750
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 9.491978609625669,
|
| 603 |
+
"grad_norm": 2.4829342365264893,
|
| 604 |
+
"learning_rate": 1.1681521093285799e-05,
|
| 605 |
+
"loss": 0.207,
|
| 606 |
+
"step": 1775
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"epoch": 9.62566844919786,
|
| 610 |
+
"grad_norm": 0.23277647793293,
|
| 611 |
+
"learning_rate": 1.1532976827094475e-05,
|
| 612 |
+
"loss": 0.1672,
|
| 613 |
+
"step": 1800
|
| 614 |
+
},
|
| 615 |
+
{
|
| 616 |
+
"epoch": 9.759358288770054,
|
| 617 |
+
"grad_norm": 0.19339442253112793,
|
| 618 |
+
"learning_rate": 1.1384432560903149e-05,
|
| 619 |
+
"loss": 0.2213,
|
| 620 |
+
"step": 1825
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 9.893048128342246,
|
| 624 |
+
"grad_norm": 0.14063772559165955,
|
| 625 |
+
"learning_rate": 1.1235888294711825e-05,
|
| 626 |
+
"loss": 0.188,
|
| 627 |
+
"step": 1850
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 10.0,
|
| 631 |
+
"eval_accuracy": 0.8583916083916084,
|
| 632 |
+
"eval_loss": 0.7773720622062683,
|
| 633 |
+
"eval_macro_f1": 0.8591427241630178,
|
| 634 |
+
"eval_runtime": 1.1881,
|
| 635 |
+
"eval_samples_per_second": 481.453,
|
| 636 |
+
"eval_steps_per_second": 30.301,
|
| 637 |
+
"eval_weighted_f1": 0.8584397203783107,
|
| 638 |
+
"step": 1870
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 10.026737967914439,
|
| 642 |
+
"grad_norm": 0.1475857049226761,
|
| 643 |
+
"learning_rate": 1.1087344028520499e-05,
|
| 644 |
+
"loss": 0.2442,
|
| 645 |
+
"step": 1875
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 10.16042780748663,
|
| 649 |
+
"grad_norm": 0.19178685545921326,
|
| 650 |
+
"learning_rate": 1.0938799762329175e-05,
|
| 651 |
+
"loss": 0.0769,
|
| 652 |
+
"step": 1900
|
| 653 |
+
},
|
| 654 |
+
{
|
| 655 |
+
"epoch": 10.294117647058824,
|
| 656 |
+
"grad_norm": 125.27078247070312,
|
| 657 |
+
"learning_rate": 1.0790255496137849e-05,
|
| 658 |
+
"loss": 0.0851,
|
| 659 |
+
"step": 1925
|
| 660 |
+
},
|
| 661 |
+
{
|
| 662 |
+
"epoch": 10.427807486631016,
|
| 663 |
+
"grad_norm": 9.923179626464844,
|
| 664 |
+
"learning_rate": 1.0641711229946525e-05,
|
| 665 |
+
"loss": 0.0508,
|
| 666 |
+
"step": 1950
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"epoch": 10.56149732620321,
|
| 670 |
+
"grad_norm": 0.1864607185125351,
|
| 671 |
+
"learning_rate": 1.0493166963755199e-05,
|
| 672 |
+
"loss": 0.1654,
|
| 673 |
+
"step": 1975
|
| 674 |
+
},
|
| 675 |
+
{
|
| 676 |
+
"epoch": 10.695187165775401,
|
| 677 |
+
"grad_norm": 0.13694506883621216,
|
| 678 |
+
"learning_rate": 1.0344622697563875e-05,
|
| 679 |
+
"loss": 0.188,
|
| 680 |
+
"step": 2000
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 10.828877005347593,
|
| 684 |
+
"grad_norm": 0.12389884889125824,
|
| 685 |
+
"learning_rate": 1.0196078431372549e-05,
|
| 686 |
+
"loss": 0.201,
|
| 687 |
+
"step": 2025
|
| 688 |
+
},
|
| 689 |
+
{
|
| 690 |
+
"epoch": 10.962566844919786,
|
| 691 |
+
"grad_norm": 1.6487425565719604,
|
| 692 |
+
"learning_rate": 1.0047534165181224e-05,
|
| 693 |
+
"loss": 0.4032,
|
| 694 |
+
"step": 2050
|
| 695 |
+
},
|
| 696 |
+
{
|
| 697 |
+
"epoch": 11.0,
|
| 698 |
+
"eval_accuracy": 0.8636363636363636,
|
| 699 |
+
"eval_loss": 0.7823806405067444,
|
| 700 |
+
"eval_macro_f1": 0.8625501336898395,
|
| 701 |
+
"eval_runtime": 1.1589,
|
| 702 |
+
"eval_samples_per_second": 493.59,
|
| 703 |
+
"eval_steps_per_second": 31.065,
|
| 704 |
+
"eval_weighted_f1": 0.8622850522917368,
|
| 705 |
+
"step": 2057
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"epoch": 11.096256684491978,
|
| 709 |
+
"grad_norm": 0.11938251554965973,
|
| 710 |
+
"learning_rate": 9.8989898989899e-06,
|
| 711 |
+
"loss": 0.1116,
|
| 712 |
+
"step": 2075
|
| 713 |
+
},
|
| 714 |
+
{
|
| 715 |
+
"epoch": 11.229946524064172,
|
| 716 |
+
"grad_norm": 0.11674272269010544,
|
| 717 |
+
"learning_rate": 9.750445632798576e-06,
|
| 718 |
+
"loss": 0.0692,
|
| 719 |
+
"step": 2100
|
| 720 |
+
},
|
| 721 |
+
{
|
| 722 |
+
"epoch": 11.363636363636363,
|
| 723 |
+
"grad_norm": 0.1071222648024559,
|
| 724 |
+
"learning_rate": 9.60190136660725e-06,
|
| 725 |
+
"loss": 0.0477,
|
| 726 |
+
"step": 2125
|
| 727 |
+
},
|
| 728 |
+
{
|
| 729 |
+
"epoch": 11.497326203208557,
|
| 730 |
+
"grad_norm": 13.00842571258545,
|
| 731 |
+
"learning_rate": 9.453357100415926e-06,
|
| 732 |
+
"loss": 0.157,
|
| 733 |
+
"step": 2150
|
| 734 |
+
},
|
| 735 |
+
{
|
| 736 |
+
"epoch": 11.631016042780749,
|
| 737 |
+
"grad_norm": 0.11801385134458542,
|
| 738 |
+
"learning_rate": 9.3048128342246e-06,
|
| 739 |
+
"loss": 0.1807,
|
| 740 |
+
"step": 2175
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 11.764705882352942,
|
| 744 |
+
"grad_norm": 0.11856313794851303,
|
| 745 |
+
"learning_rate": 9.156268568033274e-06,
|
| 746 |
+
"loss": 0.1546,
|
| 747 |
+
"step": 2200
|
| 748 |
+
},
|
| 749 |
+
{
|
| 750 |
+
"epoch": 11.898395721925134,
|
| 751 |
+
"grad_norm": 0.23650076985359192,
|
| 752 |
+
"learning_rate": 9.007724301841948e-06,
|
| 753 |
+
"loss": 0.121,
|
| 754 |
+
"step": 2225
|
| 755 |
+
},
|
| 756 |
+
{
|
| 757 |
+
"epoch": 12.0,
|
| 758 |
+
"eval_accuracy": 0.8513986013986014,
|
| 759 |
+
"eval_loss": 0.8625929951667786,
|
| 760 |
+
"eval_macro_f1": 0.8520924276325363,
|
| 761 |
+
"eval_runtime": 1.2366,
|
| 762 |
+
"eval_samples_per_second": 462.549,
|
| 763 |
+
"eval_steps_per_second": 29.111,
|
| 764 |
+
"eval_weighted_f1": 0.8513394678349749,
|
| 765 |
+
"step": 2244
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 12.032085561497325,
|
| 769 |
+
"grad_norm": 0.11647474020719528,
|
| 770 |
+
"learning_rate": 8.859180035650624e-06,
|
| 771 |
+
"loss": 0.1459,
|
| 772 |
+
"step": 2250
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 12.165775401069519,
|
| 776 |
+
"grad_norm": 0.09729824960231781,
|
| 777 |
+
"learning_rate": 8.7106357694593e-06,
|
| 778 |
+
"loss": 0.0841,
|
| 779 |
+
"step": 2275
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 12.29946524064171,
|
| 783 |
+
"grad_norm": 89.26582336425781,
|
| 784 |
+
"learning_rate": 8.562091503267974e-06,
|
| 785 |
+
"loss": 0.1326,
|
| 786 |
+
"step": 2300
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 12.433155080213904,
|
| 790 |
+
"grad_norm": 0.34565767645835876,
|
| 791 |
+
"learning_rate": 8.41354723707665e-06,
|
| 792 |
+
"loss": 0.0892,
|
| 793 |
+
"step": 2325
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 12.566844919786096,
|
| 797 |
+
"grad_norm": 0.10946846753358841,
|
| 798 |
+
"learning_rate": 8.265002970885324e-06,
|
| 799 |
+
"loss": 0.079,
|
| 800 |
+
"step": 2350
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 12.70053475935829,
|
| 804 |
+
"grad_norm": 0.08815819025039673,
|
| 805 |
+
"learning_rate": 8.116458704694e-06,
|
| 806 |
+
"loss": 0.0555,
|
| 807 |
+
"step": 2375
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 12.834224598930481,
|
| 811 |
+
"grad_norm": 0.08501690626144409,
|
| 812 |
+
"learning_rate": 7.967914438502674e-06,
|
| 813 |
+
"loss": 0.0625,
|
| 814 |
+
"step": 2400
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 12.967914438502675,
|
| 818 |
+
"grad_norm": 0.0843612551689148,
|
| 819 |
+
"learning_rate": 7.81937017231135e-06,
|
| 820 |
+
"loss": 0.1327,
|
| 821 |
+
"step": 2425
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 13.0,
|
| 825 |
+
"eval_accuracy": 0.8636363636363636,
|
| 826 |
+
"eval_loss": 0.8263055682182312,
|
| 827 |
+
"eval_macro_f1": 0.86499887362837,
|
| 828 |
+
"eval_runtime": 1.2319,
|
| 829 |
+
"eval_samples_per_second": 464.32,
|
| 830 |
+
"eval_steps_per_second": 29.223,
|
| 831 |
+
"eval_weighted_f1": 0.8643929697945394,
|
| 832 |
+
"step": 2431
|
| 833 |
+
},
|
| 834 |
+
{
|
| 835 |
+
"epoch": 13.101604278074866,
|
| 836 |
+
"grad_norm": 0.0848432332277298,
|
| 837 |
+
"learning_rate": 7.670825906120024e-06,
|
| 838 |
+
"loss": 0.0882,
|
| 839 |
+
"step": 2450
|
| 840 |
+
},
|
| 841 |
+
{
|
| 842 |
+
"epoch": 13.235294117647058,
|
| 843 |
+
"grad_norm": 0.9270504117012024,
|
| 844 |
+
"learning_rate": 7.522281639928699e-06,
|
| 845 |
+
"loss": 0.0534,
|
| 846 |
+
"step": 2475
|
| 847 |
+
},
|
| 848 |
+
{
|
| 849 |
+
"epoch": 13.368983957219251,
|
| 850 |
+
"grad_norm": 0.1587558537721634,
|
| 851 |
+
"learning_rate": 7.373737373737374e-06,
|
| 852 |
+
"loss": 0.1242,
|
| 853 |
+
"step": 2500
|
| 854 |
+
},
|
| 855 |
+
{
|
| 856 |
+
"epoch": 13.502673796791443,
|
| 857 |
+
"grad_norm": 0.4047185480594635,
|
| 858 |
+
"learning_rate": 7.2251931075460495e-06,
|
| 859 |
+
"loss": 0.0957,
|
| 860 |
+
"step": 2525
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 13.636363636363637,
|
| 864 |
+
"grad_norm": 0.08490663021802902,
|
| 865 |
+
"learning_rate": 7.0766488413547244e-06,
|
| 866 |
+
"loss": 0.0247,
|
| 867 |
+
"step": 2550
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"epoch": 13.770053475935828,
|
| 871 |
+
"grad_norm": 0.43225857615470886,
|
| 872 |
+
"learning_rate": 6.928104575163399e-06,
|
| 873 |
+
"loss": 0.0287,
|
| 874 |
+
"step": 2575
|
| 875 |
+
},
|
| 876 |
+
{
|
| 877 |
+
"epoch": 13.903743315508022,
|
| 878 |
+
"grad_norm": 0.08818015456199646,
|
| 879 |
+
"learning_rate": 6.779560308972074e-06,
|
| 880 |
+
"loss": 0.1099,
|
| 881 |
+
"step": 2600
|
| 882 |
+
},
|
| 883 |
+
{
|
| 884 |
+
"epoch": 14.0,
|
| 885 |
+
"eval_accuracy": 0.8513986013986014,
|
| 886 |
+
"eval_loss": 0.9054542183876038,
|
| 887 |
+
"eval_macro_f1": 0.8515364038625416,
|
| 888 |
+
"eval_runtime": 1.1902,
|
| 889 |
+
"eval_samples_per_second": 480.604,
|
| 890 |
+
"eval_steps_per_second": 30.248,
|
| 891 |
+
"eval_weighted_f1": 0.8507822307431314,
|
| 892 |
+
"step": 2618
|
| 893 |
+
},
|
| 894 |
+
{
|
| 895 |
+
"epoch": 14.037433155080214,
|
| 896 |
+
"grad_norm": 0.07318533211946487,
|
| 897 |
+
"learning_rate": 6.631016042780749e-06,
|
| 898 |
+
"loss": 0.0676,
|
| 899 |
+
"step": 2625
|
| 900 |
+
},
|
| 901 |
+
{
|
| 902 |
+
"epoch": 14.171122994652407,
|
| 903 |
+
"grad_norm": 0.07690595835447311,
|
| 904 |
+
"learning_rate": 6.482471776589424e-06,
|
| 905 |
+
"loss": 0.0183,
|
| 906 |
+
"step": 2650
|
| 907 |
+
},
|
| 908 |
+
{
|
| 909 |
+
"epoch": 14.304812834224599,
|
| 910 |
+
"grad_norm": 22.4063663482666,
|
| 911 |
+
"learning_rate": 6.333927510398099e-06,
|
| 912 |
+
"loss": 0.0437,
|
| 913 |
+
"step": 2675
|
| 914 |
+
},
|
| 915 |
+
{
|
| 916 |
+
"epoch": 14.43850267379679,
|
| 917 |
+
"grad_norm": 0.07405930757522583,
|
| 918 |
+
"learning_rate": 6.185383244206774e-06,
|
| 919 |
+
"loss": 0.0428,
|
| 920 |
+
"step": 2700
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 14.572192513368984,
|
| 924 |
+
"grad_norm": 0.07327538728713989,
|
| 925 |
+
"learning_rate": 6.036838978015449e-06,
|
| 926 |
+
"loss": 0.0111,
|
| 927 |
+
"step": 2725
|
| 928 |
+
},
|
| 929 |
+
{
|
| 930 |
+
"epoch": 14.705882352941176,
|
| 931 |
+
"grad_norm": 0.07578103244304657,
|
| 932 |
+
"learning_rate": 5.888294711824124e-06,
|
| 933 |
+
"loss": 0.076,
|
| 934 |
+
"step": 2750
|
| 935 |
+
},
|
| 936 |
+
{
|
| 937 |
+
"epoch": 14.83957219251337,
|
| 938 |
+
"grad_norm": 0.08358561992645264,
|
| 939 |
+
"learning_rate": 5.739750445632799e-06,
|
| 940 |
+
"loss": 0.0187,
|
| 941 |
+
"step": 2775
|
| 942 |
+
},
|
| 943 |
+
{
|
| 944 |
+
"epoch": 14.973262032085561,
|
| 945 |
+
"grad_norm": 0.07746005803346634,
|
| 946 |
+
"learning_rate": 5.591206179441474e-06,
|
| 947 |
+
"loss": 0.0135,
|
| 948 |
+
"step": 2800
|
| 949 |
+
},
|
| 950 |
+
{
|
| 951 |
+
"epoch": 15.0,
|
| 952 |
+
"eval_accuracy": 0.8653846153846154,
|
| 953 |
+
"eval_loss": 0.874975860118866,
|
| 954 |
+
"eval_macro_f1": 0.8658226718172223,
|
| 955 |
+
"eval_runtime": 1.2051,
|
| 956 |
+
"eval_samples_per_second": 474.636,
|
| 957 |
+
"eval_steps_per_second": 29.872,
|
| 958 |
+
"eval_weighted_f1": 0.8651976025768445,
|
| 959 |
+
"step": 2805
|
| 960 |
+
},
|
| 961 |
+
{
|
| 962 |
+
"epoch": 15.106951871657754,
|
| 963 |
+
"grad_norm": 0.0651995912194252,
|
| 964 |
+
"learning_rate": 5.442661913250149e-06,
|
| 965 |
+
"loss": 0.0464,
|
| 966 |
+
"step": 2825
|
| 967 |
+
},
|
| 968 |
+
{
|
| 969 |
+
"epoch": 15.240641711229946,
|
| 970 |
+
"grad_norm": 0.06533762067556381,
|
| 971 |
+
"learning_rate": 5.294117647058824e-06,
|
| 972 |
+
"loss": 0.1026,
|
| 973 |
+
"step": 2850
|
| 974 |
+
},
|
| 975 |
+
{
|
| 976 |
+
"epoch": 15.37433155080214,
|
| 977 |
+
"grad_norm": 0.06515900790691376,
|
| 978 |
+
"learning_rate": 5.145573380867499e-06,
|
| 979 |
+
"loss": 0.0381,
|
| 980 |
+
"step": 2875
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 15.508021390374331,
|
| 984 |
+
"grad_norm": 0.06106071174144745,
|
| 985 |
+
"learning_rate": 4.997029114676174e-06,
|
| 986 |
+
"loss": 0.0096,
|
| 987 |
+
"step": 2900
|
| 988 |
+
},
|
| 989 |
+
{
|
| 990 |
+
"epoch": 15.641711229946525,
|
| 991 |
+
"grad_norm": 0.062081228941679,
|
| 992 |
+
"learning_rate": 4.848484848484849e-06,
|
| 993 |
+
"loss": 0.0978,
|
| 994 |
+
"step": 2925
|
| 995 |
+
},
|
| 996 |
+
{
|
| 997 |
+
"epoch": 15.775401069518717,
|
| 998 |
+
"grad_norm": 0.06622225046157837,
|
| 999 |
+
"learning_rate": 4.699940582293524e-06,
|
| 1000 |
+
"loss": 0.045,
|
| 1001 |
+
"step": 2950
|
| 1002 |
+
},
|
| 1003 |
+
{
|
| 1004 |
+
"epoch": 15.909090909090908,
|
| 1005 |
+
"grad_norm": 0.06503937393426895,
|
| 1006 |
+
"learning_rate": 4.551396316102199e-06,
|
| 1007 |
+
"loss": 0.0951,
|
| 1008 |
+
"step": 2975
|
| 1009 |
+
},
|
| 1010 |
+
{
|
| 1011 |
+
"epoch": 16.0,
|
| 1012 |
+
"eval_accuracy": 0.8618881118881119,
|
| 1013 |
+
"eval_loss": 0.8460711240768433,
|
| 1014 |
+
"eval_macro_f1": 0.8624659223982666,
|
| 1015 |
+
"eval_runtime": 1.1881,
|
| 1016 |
+
"eval_samples_per_second": 481.457,
|
| 1017 |
+
"eval_steps_per_second": 30.301,
|
| 1018 |
+
"eval_weighted_f1": 0.8619311378061421,
|
| 1019 |
+
"step": 2992
|
| 1020 |
+
},
|
| 1021 |
+
{
|
| 1022 |
+
"epoch": 16.0427807486631,
|
| 1023 |
+
"grad_norm": 0.06143380329012871,
|
| 1024 |
+
"learning_rate": 4.402852049910874e-06,
|
| 1025 |
+
"loss": 0.0094,
|
| 1026 |
+
"step": 3000
|
| 1027 |
+
},
|
| 1028 |
+
{
|
| 1029 |
+
"epoch": 16.176470588235293,
|
| 1030 |
+
"grad_norm": 0.059074513614177704,
|
| 1031 |
+
"learning_rate": 4.254307783719549e-06,
|
| 1032 |
+
"loss": 0.0092,
|
| 1033 |
+
"step": 3025
|
| 1034 |
+
},
|
| 1035 |
+
{
|
| 1036 |
+
"epoch": 16.310160427807485,
|
| 1037 |
+
"grad_norm": 0.07046302407979965,
|
| 1038 |
+
"learning_rate": 4.105763517528224e-06,
|
| 1039 |
+
"loss": 0.0093,
|
| 1040 |
+
"step": 3050
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 16.44385026737968,
|
| 1044 |
+
"grad_norm": 0.059618692845106125,
|
| 1045 |
+
"learning_rate": 3.957219251336899e-06,
|
| 1046 |
+
"loss": 0.0599,
|
| 1047 |
+
"step": 3075
|
| 1048 |
+
},
|
| 1049 |
+
{
|
| 1050 |
+
"epoch": 16.577540106951872,
|
| 1051 |
+
"grad_norm": 0.09442026168107986,
|
| 1052 |
+
"learning_rate": 3.8086749851455735e-06,
|
| 1053 |
+
"loss": 0.0198,
|
| 1054 |
+
"step": 3100
|
| 1055 |
+
},
|
| 1056 |
+
{
|
| 1057 |
+
"epoch": 16.711229946524064,
|
| 1058 |
+
"grad_norm": 0.05578425154089928,
|
| 1059 |
+
"learning_rate": 3.6601307189542484e-06,
|
| 1060 |
+
"loss": 0.0088,
|
| 1061 |
+
"step": 3125
|
| 1062 |
+
},
|
| 1063 |
+
{
|
| 1064 |
+
"epoch": 16.844919786096256,
|
| 1065 |
+
"grad_norm": 0.05314868688583374,
|
| 1066 |
+
"learning_rate": 3.5115864527629234e-06,
|
| 1067 |
+
"loss": 0.051,
|
| 1068 |
+
"step": 3150
|
| 1069 |
+
},
|
| 1070 |
+
{
|
| 1071 |
+
"epoch": 16.97860962566845,
|
| 1072 |
+
"grad_norm": 0.0874389111995697,
|
| 1073 |
+
"learning_rate": 3.3630421865715983e-06,
|
| 1074 |
+
"loss": 0.0087,
|
| 1075 |
+
"step": 3175
|
| 1076 |
+
},
|
| 1077 |
+
{
|
| 1078 |
+
"epoch": 17.0,
|
| 1079 |
+
"eval_accuracy": 0.8583916083916084,
|
| 1080 |
+
"eval_loss": 0.8884491920471191,
|
| 1081 |
+
"eval_macro_f1": 0.8590207768744355,
|
| 1082 |
+
"eval_runtime": 1.2582,
|
| 1083 |
+
"eval_samples_per_second": 454.616,
|
| 1084 |
+
"eval_steps_per_second": 28.612,
|
| 1085 |
+
"eval_weighted_f1": 0.8583370667273107,
|
| 1086 |
+
"step": 3179
|
| 1087 |
+
},
|
| 1088 |
+
{
|
| 1089 |
+
"epoch": 17.112299465240643,
|
| 1090 |
+
"grad_norm": 0.05569047853350639,
|
| 1091 |
+
"learning_rate": 3.2144979203802733e-06,
|
| 1092 |
+
"loss": 0.0084,
|
| 1093 |
+
"step": 3200
|
| 1094 |
+
},
|
| 1095 |
+
{
|
| 1096 |
+
"epoch": 17.245989304812834,
|
| 1097 |
+
"grad_norm": 0.05263025313615799,
|
| 1098 |
+
"learning_rate": 3.0659536541889482e-06,
|
| 1099 |
+
"loss": 0.0777,
|
| 1100 |
+
"step": 3225
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 17.379679144385026,
|
| 1104 |
+
"grad_norm": 0.058189645409584045,
|
| 1105 |
+
"learning_rate": 2.917409387997623e-06,
|
| 1106 |
+
"loss": 0.0216,
|
| 1107 |
+
"step": 3250
|
| 1108 |
+
},
|
| 1109 |
+
{
|
| 1110 |
+
"epoch": 17.513368983957218,
|
| 1111 |
+
"grad_norm": 0.06913315504789352,
|
| 1112 |
+
"learning_rate": 2.7688651218062986e-06,
|
| 1113 |
+
"loss": 0.0653,
|
| 1114 |
+
"step": 3275
|
| 1115 |
+
},
|
| 1116 |
+
{
|
| 1117 |
+
"epoch": 17.647058823529413,
|
| 1118 |
+
"grad_norm": 0.056622717529535294,
|
| 1119 |
+
"learning_rate": 2.6203208556149735e-06,
|
| 1120 |
+
"loss": 0.0415,
|
| 1121 |
+
"step": 3300
|
| 1122 |
+
},
|
| 1123 |
+
{
|
| 1124 |
+
"epoch": 17.780748663101605,
|
| 1125 |
+
"grad_norm": 0.05018749460577965,
|
| 1126 |
+
"learning_rate": 2.4717765894236485e-06,
|
| 1127 |
+
"loss": 0.008,
|
| 1128 |
+
"step": 3325
|
| 1129 |
+
},
|
| 1130 |
+
{
|
| 1131 |
+
"epoch": 17.914438502673796,
|
| 1132 |
+
"grad_norm": 0.056364480406045914,
|
| 1133 |
+
"learning_rate": 2.3232323232323234e-06,
|
| 1134 |
+
"loss": 0.0081,
|
| 1135 |
+
"step": 3350
|
| 1136 |
+
},
|
| 1137 |
+
{
|
| 1138 |
+
"epoch": 18.0,
|
| 1139 |
+
"eval_accuracy": 0.8636363636363636,
|
| 1140 |
+
"eval_loss": 0.8747503757476807,
|
| 1141 |
+
"eval_macro_f1": 0.8638818464963123,
|
| 1142 |
+
"eval_runtime": 1.2283,
|
| 1143 |
+
"eval_samples_per_second": 465.679,
|
| 1144 |
+
"eval_steps_per_second": 29.308,
|
| 1145 |
+
"eval_weighted_f1": 0.8633145228910446,
|
| 1146 |
+
"step": 3366
|
| 1147 |
+
}
|
| 1148 |
+
],
|
| 1149 |
+
"logging_steps": 25,
|
| 1150 |
+
"max_steps": 3740,
|
| 1151 |
+
"num_input_tokens_seen": 0,
|
| 1152 |
+
"num_train_epochs": 20,
|
| 1153 |
+
"save_steps": 500,
|
| 1154 |
+
"stateful_callbacks": {
|
| 1155 |
+
"EarlyStoppingCallback": {
|
| 1156 |
+
"args": {
|
| 1157 |
+
"early_stopping_patience": 3,
|
| 1158 |
+
"early_stopping_threshold": 0.0
|
| 1159 |
+
},
|
| 1160 |
+
"attributes": {
|
| 1161 |
+
"early_stopping_patience_counter": 3
|
| 1162 |
+
}
|
| 1163 |
+
},
|
| 1164 |
+
"TrainerControl": {
|
| 1165 |
+
"args": {
|
| 1166 |
+
"should_epoch_stop": false,
|
| 1167 |
+
"should_evaluate": false,
|
| 1168 |
+
"should_log": false,
|
| 1169 |
+
"should_save": true,
|
| 1170 |
+
"should_training_stop": true
|
| 1171 |
+
},
|
| 1172 |
+
"attributes": {}
|
| 1173 |
+
}
|
| 1174 |
+
},
|
| 1175 |
+
"total_flos": 0.0,
|
| 1176 |
+
"train_batch_size": 8,
|
| 1177 |
+
"trial_name": null,
|
| 1178 |
+
"trial_params": null
|
| 1179 |
+
}
|
ood-vifactcheck-phobert-topic-domain-00-s42-20260814T103527Z-00ca7096/trainer/checkpoint-3366/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-01-s42-20260814T105726Z-00ca7096/trainer/checkpoint-1146/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-topic-domain-01-s42-20260814T105726Z-00ca7096/trainer/checkpoint-1146/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-01-s42-20260814T105726Z-00ca7096/trainer/checkpoint-1146/special_tokens_map.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": "<mask>",
|
| 6 |
+
"pad_token": "<pad>",
|
| 7 |
+
"sep_token": "</s>",
|
| 8 |
+
"unk_token": "<unk>"
|
| 9 |
+
}
|
ood-vifactcheck-phobert-topic-domain-01-s42-20260814T105726Z-00ca7096/trainer/checkpoint-1146/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 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": "</s>",
|
| 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 |
+
"64000": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 256,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "PhobertTokenizer",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-phobert-topic-domain-01-s42-20260814T105726Z-00ca7096/trainer/checkpoint-1146/trainer_state.json
ADDED
|
@@ -0,0 +1,424 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 573,
|
| 3 |
+
"best_metric": 0.8500705366521903,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-topic-domain-01-s42-20260814T105726Z-00ca7096/trainer/checkpoint-573",
|
| 5 |
+
"epoch": 6.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1146,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.1310615989515072,
|
| 14 |
+
"grad_norm": 7.5195159912109375,
|
| 15 |
+
"learning_rate": 1.256544502617801e-06,
|
| 16 |
+
"loss": 4.456,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.2621231979030144,
|
| 21 |
+
"grad_norm": 5.693071365356445,
|
| 22 |
+
"learning_rate": 2.5654450261780104e-06,
|
| 23 |
+
"loss": 4.4009,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.3931847968545216,
|
| 28 |
+
"grad_norm": 6.818910121917725,
|
| 29 |
+
"learning_rate": 3.874345549738221e-06,
|
| 30 |
+
"loss": 4.3696,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5242463958060288,
|
| 35 |
+
"grad_norm": 6.5994439125061035,
|
| 36 |
+
"learning_rate": 5.18324607329843e-06,
|
| 37 |
+
"loss": 4.3679,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6553079947575361,
|
| 42 |
+
"grad_norm": 7.573624610900879,
|
| 43 |
+
"learning_rate": 6.492146596858639e-06,
|
| 44 |
+
"loss": 4.3334,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.7863695937090432,
|
| 49 |
+
"grad_norm": 15.265385627746582,
|
| 50 |
+
"learning_rate": 7.80104712041885e-06,
|
| 51 |
+
"loss": 4.291,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9174311926605505,
|
| 56 |
+
"grad_norm": 11.147649765014648,
|
| 57 |
+
"learning_rate": 9.109947643979057e-06,
|
| 58 |
+
"loss": 4.1856,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.5068493150684932,
|
| 64 |
+
"eval_loss": 1.0000033378601074,
|
| 65 |
+
"eval_macro_f1": 0.46955119519455774,
|
| 66 |
+
"eval_runtime": 1.1934,
|
| 67 |
+
"eval_samples_per_second": 489.353,
|
| 68 |
+
"eval_steps_per_second": 31.004,
|
| 69 |
+
"eval_weighted_f1": 0.4720473839327462,
|
| 70 |
+
"step": 191
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0471821756225426,
|
| 74 |
+
"grad_norm": 17.308910369873047,
|
| 75 |
+
"learning_rate": 1.0418848167539269e-05,
|
| 76 |
+
"loss": 3.8701,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.1782437745740497,
|
| 81 |
+
"grad_norm": 28.271278381347656,
|
| 82 |
+
"learning_rate": 1.1727748691099476e-05,
|
| 83 |
+
"loss": 3.6035,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.309305373525557,
|
| 88 |
+
"grad_norm": 23.250885009765625,
|
| 89 |
+
"learning_rate": 1.3036649214659686e-05,
|
| 90 |
+
"loss": 3.2413,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.4403669724770642,
|
| 95 |
+
"grad_norm": 22.559526443481445,
|
| 96 |
+
"learning_rate": 1.4345549738219897e-05,
|
| 97 |
+
"loss": 2.694,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.5714285714285714,
|
| 102 |
+
"grad_norm": 27.39093780517578,
|
| 103 |
+
"learning_rate": 1.5654450261780107e-05,
|
| 104 |
+
"loss": 2.4208,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7024901703800785,
|
| 109 |
+
"grad_norm": 37.660865783691406,
|
| 110 |
+
"learning_rate": 1.6963350785340316e-05,
|
| 111 |
+
"loss": 2.2826,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.833551769331586,
|
| 116 |
+
"grad_norm": 39.857200622558594,
|
| 117 |
+
"learning_rate": 1.8272251308900526e-05,
|
| 118 |
+
"loss": 2.2322,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 1.964613368283093,
|
| 123 |
+
"grad_norm": 63.87898635864258,
|
| 124 |
+
"learning_rate": 1.9581151832460736e-05,
|
| 125 |
+
"loss": 2.1782,
|
| 126 |
+
"step": 375
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"epoch": 2.0,
|
| 130 |
+
"eval_accuracy": 0.8133561643835616,
|
| 131 |
+
"eval_loss": 0.4772372245788574,
|
| 132 |
+
"eval_macro_f1": 0.8112302946882335,
|
| 133 |
+
"eval_runtime": 1.4355,
|
| 134 |
+
"eval_samples_per_second": 406.822,
|
| 135 |
+
"eval_steps_per_second": 25.775,
|
| 136 |
+
"eval_weighted_f1": 0.8117664384892669,
|
| 137 |
+
"step": 382
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.094364351245085,
|
| 141 |
+
"grad_norm": 25.855710983276367,
|
| 142 |
+
"learning_rate": 1.9901105293775453e-05,
|
| 143 |
+
"loss": 1.75,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.2254259501965925,
|
| 148 |
+
"grad_norm": 61.04750442504883,
|
| 149 |
+
"learning_rate": 1.9755671902268764e-05,
|
| 150 |
+
"loss": 1.7504,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.3564875491480994,
|
| 155 |
+
"grad_norm": 27.680606842041016,
|
| 156 |
+
"learning_rate": 1.961023851076207e-05,
|
| 157 |
+
"loss": 1.4448,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.487549148099607,
|
| 162 |
+
"grad_norm": 38.47274398803711,
|
| 163 |
+
"learning_rate": 1.9464805119255382e-05,
|
| 164 |
+
"loss": 1.372,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.618610747051114,
|
| 169 |
+
"grad_norm": 29.965795516967773,
|
| 170 |
+
"learning_rate": 1.9319371727748693e-05,
|
| 171 |
+
"loss": 1.4889,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.749672346002621,
|
| 176 |
+
"grad_norm": 30.392751693725586,
|
| 177 |
+
"learning_rate": 1.9173938336242e-05,
|
| 178 |
+
"loss": 1.4608,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 2.8807339449541285,
|
| 183 |
+
"grad_norm": 49.441959381103516,
|
| 184 |
+
"learning_rate": 1.9028504944735315e-05,
|
| 185 |
+
"loss": 1.6828,
|
| 186 |
+
"step": 550
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 3.0,
|
| 190 |
+
"eval_accuracy": 0.8527397260273972,
|
| 191 |
+
"eval_loss": 0.4306429624557495,
|
| 192 |
+
"eval_macro_f1": 0.8500705366521903,
|
| 193 |
+
"eval_runtime": 1.2378,
|
| 194 |
+
"eval_samples_per_second": 471.822,
|
| 195 |
+
"eval_steps_per_second": 29.893,
|
| 196 |
+
"eval_weighted_f1": 0.8504843057591035,
|
| 197 |
+
"step": 573
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.0104849279161208,
|
| 201 |
+
"grad_norm": 33.82920837402344,
|
| 202 |
+
"learning_rate": 1.8883071553228622e-05,
|
| 203 |
+
"loss": 1.5133,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.1415465268676277,
|
| 208 |
+
"grad_norm": 32.32925033569336,
|
| 209 |
+
"learning_rate": 1.8737638161721933e-05,
|
| 210 |
+
"loss": 1.1392,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.272608125819135,
|
| 215 |
+
"grad_norm": 49.39571762084961,
|
| 216 |
+
"learning_rate": 1.8592204770215244e-05,
|
| 217 |
+
"loss": 1.1462,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.4036697247706424,
|
| 222 |
+
"grad_norm": 25.576679229736328,
|
| 223 |
+
"learning_rate": 1.844677137870855e-05,
|
| 224 |
+
"loss": 0.7983,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.5347313237221494,
|
| 229 |
+
"grad_norm": 24.273832321166992,
|
| 230 |
+
"learning_rate": 1.8301337987201865e-05,
|
| 231 |
+
"loss": 0.9893,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.6657929226736568,
|
| 236 |
+
"grad_norm": 27.20750617980957,
|
| 237 |
+
"learning_rate": 1.8155904595695173e-05,
|
| 238 |
+
"loss": 0.8961,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.7968545216251637,
|
| 243 |
+
"grad_norm": 50.282962799072266,
|
| 244 |
+
"learning_rate": 1.8010471204188483e-05,
|
| 245 |
+
"loss": 1.1046,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 3.927916120576671,
|
| 250 |
+
"grad_norm": 67.69908905029297,
|
| 251 |
+
"learning_rate": 1.7865037812681794e-05,
|
| 252 |
+
"loss": 0.859,
|
| 253 |
+
"step": 750
|
| 254 |
+
},
|
| 255 |
+
{
|
| 256 |
+
"epoch": 4.0,
|
| 257 |
+
"eval_accuracy": 0.8356164383561644,
|
| 258 |
+
"eval_loss": 0.5322219729423523,
|
| 259 |
+
"eval_macro_f1": 0.8363672182821119,
|
| 260 |
+
"eval_runtime": 1.1909,
|
| 261 |
+
"eval_samples_per_second": 490.377,
|
| 262 |
+
"eval_steps_per_second": 31.068,
|
| 263 |
+
"eval_weighted_f1": 0.8363768661549921,
|
| 264 |
+
"step": 764
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.057667103538663,
|
| 268 |
+
"grad_norm": 21.964488983154297,
|
| 269 |
+
"learning_rate": 1.77196044211751e-05,
|
| 270 |
+
"loss": 0.8494,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.18872870249017,
|
| 275 |
+
"grad_norm": 87.52288818359375,
|
| 276 |
+
"learning_rate": 1.7574171029668416e-05,
|
| 277 |
+
"loss": 0.7407,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.319790301441677,
|
| 282 |
+
"grad_norm": 87.03518676757812,
|
| 283 |
+
"learning_rate": 1.7428737638161723e-05,
|
| 284 |
+
"loss": 0.6809,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.450851900393185,
|
| 289 |
+
"grad_norm": 39.1880989074707,
|
| 290 |
+
"learning_rate": 1.7283304246655034e-05,
|
| 291 |
+
"loss": 0.7907,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.581913499344692,
|
| 296 |
+
"grad_norm": 7.51291561126709,
|
| 297 |
+
"learning_rate": 1.7137870855148345e-05,
|
| 298 |
+
"loss": 0.6114,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.712975098296199,
|
| 303 |
+
"grad_norm": 76.20528411865234,
|
| 304 |
+
"learning_rate": 1.6992437463641652e-05,
|
| 305 |
+
"loss": 0.7657,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 4.844036697247707,
|
| 310 |
+
"grad_norm": 23.573692321777344,
|
| 311 |
+
"learning_rate": 1.6847004072134963e-05,
|
| 312 |
+
"loss": 0.7287,
|
| 313 |
+
"step": 925
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 4.975098296199214,
|
| 317 |
+
"grad_norm": 21.339487075805664,
|
| 318 |
+
"learning_rate": 1.6701570680628274e-05,
|
| 319 |
+
"loss": 0.6486,
|
| 320 |
+
"step": 950
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 5.0,
|
| 324 |
+
"eval_accuracy": 0.8373287671232876,
|
| 325 |
+
"eval_loss": 0.5619485974311829,
|
| 326 |
+
"eval_macro_f1": 0.8385890803338713,
|
| 327 |
+
"eval_runtime": 1.2284,
|
| 328 |
+
"eval_samples_per_second": 475.401,
|
| 329 |
+
"eval_steps_per_second": 30.12,
|
| 330 |
+
"eval_weighted_f1": 0.83888630356804,
|
| 331 |
+
"step": 955
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.104849279161206,
|
| 335 |
+
"grad_norm": 35.13225555419922,
|
| 336 |
+
"learning_rate": 1.6556137289121585e-05,
|
| 337 |
+
"loss": 0.5485,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.235910878112713,
|
| 342 |
+
"grad_norm": 1.1914633512496948,
|
| 343 |
+
"learning_rate": 1.6410703897614895e-05,
|
| 344 |
+
"loss": 0.5871,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.36697247706422,
|
| 349 |
+
"grad_norm": 19.375429153442383,
|
| 350 |
+
"learning_rate": 1.6265270506108203e-05,
|
| 351 |
+
"loss": 0.3756,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.498034076015728,
|
| 356 |
+
"grad_norm": 53.66272735595703,
|
| 357 |
+
"learning_rate": 1.6119837114601514e-05,
|
| 358 |
+
"loss": 0.4598,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.6290956749672345,
|
| 363 |
+
"grad_norm": 152.06411743164062,
|
| 364 |
+
"learning_rate": 1.5974403723094824e-05,
|
| 365 |
+
"loss": 0.3445,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 5.760157273918741,
|
| 370 |
+
"grad_norm": 124.9399642944336,
|
| 371 |
+
"learning_rate": 1.5828970331588135e-05,
|
| 372 |
+
"loss": 0.581,
|
| 373 |
+
"step": 1100
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 5.891218872870249,
|
| 377 |
+
"grad_norm": 16.140417098999023,
|
| 378 |
+
"learning_rate": 1.5683536940081443e-05,
|
| 379 |
+
"loss": 0.7217,
|
| 380 |
+
"step": 1125
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 6.0,
|
| 384 |
+
"eval_accuracy": 0.8424657534246576,
|
| 385 |
+
"eval_loss": 0.6397063136100769,
|
| 386 |
+
"eval_macro_f1": 0.8409629454504248,
|
| 387 |
+
"eval_runtime": 1.1877,
|
| 388 |
+
"eval_samples_per_second": 491.7,
|
| 389 |
+
"eval_steps_per_second": 31.152,
|
| 390 |
+
"eval_weighted_f1": 0.8410379276807625,
|
| 391 |
+
"step": 1146
|
| 392 |
+
}
|
| 393 |
+
],
|
| 394 |
+
"logging_steps": 25,
|
| 395 |
+
"max_steps": 3820,
|
| 396 |
+
"num_input_tokens_seen": 0,
|
| 397 |
+
"num_train_epochs": 20,
|
| 398 |
+
"save_steps": 500,
|
| 399 |
+
"stateful_callbacks": {
|
| 400 |
+
"EarlyStoppingCallback": {
|
| 401 |
+
"args": {
|
| 402 |
+
"early_stopping_patience": 3,
|
| 403 |
+
"early_stopping_threshold": 0.0
|
| 404 |
+
},
|
| 405 |
+
"attributes": {
|
| 406 |
+
"early_stopping_patience_counter": 3
|
| 407 |
+
}
|
| 408 |
+
},
|
| 409 |
+
"TrainerControl": {
|
| 410 |
+
"args": {
|
| 411 |
+
"should_epoch_stop": false,
|
| 412 |
+
"should_evaluate": false,
|
| 413 |
+
"should_log": false,
|
| 414 |
+
"should_save": true,
|
| 415 |
+
"should_training_stop": true
|
| 416 |
+
},
|
| 417 |
+
"attributes": {}
|
| 418 |
+
}
|
| 419 |
+
},
|
| 420 |
+
"total_flos": 0.0,
|
| 421 |
+
"train_batch_size": 8,
|
| 422 |
+
"trial_name": null,
|
| 423 |
+
"trial_params": null
|
| 424 |
+
}
|
ood-vifactcheck-phobert-topic-domain-01-s42-20260814T105726Z-00ca7096/trainer/checkpoint-1146/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-01-s42-20260814T105726Z-00ca7096/trainer/checkpoint-573/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-topic-domain-01-s42-20260814T105726Z-00ca7096/trainer/checkpoint-573/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-01-s42-20260814T105726Z-00ca7096/trainer/checkpoint-573/special_tokens_map.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": "<mask>",
|
| 6 |
+
"pad_token": "<pad>",
|
| 7 |
+
"sep_token": "</s>",
|
| 8 |
+
"unk_token": "<unk>"
|
| 9 |
+
}
|
ood-vifactcheck-phobert-topic-domain-01-s42-20260814T105726Z-00ca7096/trainer/checkpoint-573/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 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": "</s>",
|
| 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 |
+
"64000": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 256,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "PhobertTokenizer",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-phobert-topic-domain-01-s42-20260814T105726Z-00ca7096/trainer/checkpoint-573/trainer_state.json
ADDED
|
@@ -0,0 +1,230 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 573,
|
| 3 |
+
"best_metric": 0.8500705366521903,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-topic-domain-01-s42-20260814T105726Z-00ca7096/trainer/checkpoint-573",
|
| 5 |
+
"epoch": 3.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 573,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.1310615989515072,
|
| 14 |
+
"grad_norm": 7.5195159912109375,
|
| 15 |
+
"learning_rate": 1.256544502617801e-06,
|
| 16 |
+
"loss": 4.456,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.2621231979030144,
|
| 21 |
+
"grad_norm": 5.693071365356445,
|
| 22 |
+
"learning_rate": 2.5654450261780104e-06,
|
| 23 |
+
"loss": 4.4009,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.3931847968545216,
|
| 28 |
+
"grad_norm": 6.818910121917725,
|
| 29 |
+
"learning_rate": 3.874345549738221e-06,
|
| 30 |
+
"loss": 4.3696,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5242463958060288,
|
| 35 |
+
"grad_norm": 6.5994439125061035,
|
| 36 |
+
"learning_rate": 5.18324607329843e-06,
|
| 37 |
+
"loss": 4.3679,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6553079947575361,
|
| 42 |
+
"grad_norm": 7.573624610900879,
|
| 43 |
+
"learning_rate": 6.492146596858639e-06,
|
| 44 |
+
"loss": 4.3334,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.7863695937090432,
|
| 49 |
+
"grad_norm": 15.265385627746582,
|
| 50 |
+
"learning_rate": 7.80104712041885e-06,
|
| 51 |
+
"loss": 4.291,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9174311926605505,
|
| 56 |
+
"grad_norm": 11.147649765014648,
|
| 57 |
+
"learning_rate": 9.109947643979057e-06,
|
| 58 |
+
"loss": 4.1856,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.5068493150684932,
|
| 64 |
+
"eval_loss": 1.0000033378601074,
|
| 65 |
+
"eval_macro_f1": 0.46955119519455774,
|
| 66 |
+
"eval_runtime": 1.1934,
|
| 67 |
+
"eval_samples_per_second": 489.353,
|
| 68 |
+
"eval_steps_per_second": 31.004,
|
| 69 |
+
"eval_weighted_f1": 0.4720473839327462,
|
| 70 |
+
"step": 191
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0471821756225426,
|
| 74 |
+
"grad_norm": 17.308910369873047,
|
| 75 |
+
"learning_rate": 1.0418848167539269e-05,
|
| 76 |
+
"loss": 3.8701,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.1782437745740497,
|
| 81 |
+
"grad_norm": 28.271278381347656,
|
| 82 |
+
"learning_rate": 1.1727748691099476e-05,
|
| 83 |
+
"loss": 3.6035,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.309305373525557,
|
| 88 |
+
"grad_norm": 23.250885009765625,
|
| 89 |
+
"learning_rate": 1.3036649214659686e-05,
|
| 90 |
+
"loss": 3.2413,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.4403669724770642,
|
| 95 |
+
"grad_norm": 22.559526443481445,
|
| 96 |
+
"learning_rate": 1.4345549738219897e-05,
|
| 97 |
+
"loss": 2.694,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.5714285714285714,
|
| 102 |
+
"grad_norm": 27.39093780517578,
|
| 103 |
+
"learning_rate": 1.5654450261780107e-05,
|
| 104 |
+
"loss": 2.4208,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7024901703800785,
|
| 109 |
+
"grad_norm": 37.660865783691406,
|
| 110 |
+
"learning_rate": 1.6963350785340316e-05,
|
| 111 |
+
"loss": 2.2826,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.833551769331586,
|
| 116 |
+
"grad_norm": 39.857200622558594,
|
| 117 |
+
"learning_rate": 1.8272251308900526e-05,
|
| 118 |
+
"loss": 2.2322,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 1.964613368283093,
|
| 123 |
+
"grad_norm": 63.87898635864258,
|
| 124 |
+
"learning_rate": 1.9581151832460736e-05,
|
| 125 |
+
"loss": 2.1782,
|
| 126 |
+
"step": 375
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"epoch": 2.0,
|
| 130 |
+
"eval_accuracy": 0.8133561643835616,
|
| 131 |
+
"eval_loss": 0.4772372245788574,
|
| 132 |
+
"eval_macro_f1": 0.8112302946882335,
|
| 133 |
+
"eval_runtime": 1.4355,
|
| 134 |
+
"eval_samples_per_second": 406.822,
|
| 135 |
+
"eval_steps_per_second": 25.775,
|
| 136 |
+
"eval_weighted_f1": 0.8117664384892669,
|
| 137 |
+
"step": 382
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.094364351245085,
|
| 141 |
+
"grad_norm": 25.855710983276367,
|
| 142 |
+
"learning_rate": 1.9901105293775453e-05,
|
| 143 |
+
"loss": 1.75,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.2254259501965925,
|
| 148 |
+
"grad_norm": 61.04750442504883,
|
| 149 |
+
"learning_rate": 1.9755671902268764e-05,
|
| 150 |
+
"loss": 1.7504,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.3564875491480994,
|
| 155 |
+
"grad_norm": 27.680606842041016,
|
| 156 |
+
"learning_rate": 1.961023851076207e-05,
|
| 157 |
+
"loss": 1.4448,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.487549148099607,
|
| 162 |
+
"grad_norm": 38.47274398803711,
|
| 163 |
+
"learning_rate": 1.9464805119255382e-05,
|
| 164 |
+
"loss": 1.372,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.618610747051114,
|
| 169 |
+
"grad_norm": 29.965795516967773,
|
| 170 |
+
"learning_rate": 1.9319371727748693e-05,
|
| 171 |
+
"loss": 1.4889,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.749672346002621,
|
| 176 |
+
"grad_norm": 30.392751693725586,
|
| 177 |
+
"learning_rate": 1.9173938336242e-05,
|
| 178 |
+
"loss": 1.4608,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 2.8807339449541285,
|
| 183 |
+
"grad_norm": 49.441959381103516,
|
| 184 |
+
"learning_rate": 1.9028504944735315e-05,
|
| 185 |
+
"loss": 1.6828,
|
| 186 |
+
"step": 550
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 3.0,
|
| 190 |
+
"eval_accuracy": 0.8527397260273972,
|
| 191 |
+
"eval_loss": 0.4306429624557495,
|
| 192 |
+
"eval_macro_f1": 0.8500705366521903,
|
| 193 |
+
"eval_runtime": 1.2378,
|
| 194 |
+
"eval_samples_per_second": 471.822,
|
| 195 |
+
"eval_steps_per_second": 29.893,
|
| 196 |
+
"eval_weighted_f1": 0.8504843057591035,
|
| 197 |
+
"step": 573
|
| 198 |
+
}
|
| 199 |
+
],
|
| 200 |
+
"logging_steps": 25,
|
| 201 |
+
"max_steps": 3820,
|
| 202 |
+
"num_input_tokens_seen": 0,
|
| 203 |
+
"num_train_epochs": 20,
|
| 204 |
+
"save_steps": 500,
|
| 205 |
+
"stateful_callbacks": {
|
| 206 |
+
"EarlyStoppingCallback": {
|
| 207 |
+
"args": {
|
| 208 |
+
"early_stopping_patience": 3,
|
| 209 |
+
"early_stopping_threshold": 0.0
|
| 210 |
+
},
|
| 211 |
+
"attributes": {
|
| 212 |
+
"early_stopping_patience_counter": 0
|
| 213 |
+
}
|
| 214 |
+
},
|
| 215 |
+
"TrainerControl": {
|
| 216 |
+
"args": {
|
| 217 |
+
"should_epoch_stop": false,
|
| 218 |
+
"should_evaluate": false,
|
| 219 |
+
"should_log": false,
|
| 220 |
+
"should_save": true,
|
| 221 |
+
"should_training_stop": false
|
| 222 |
+
},
|
| 223 |
+
"attributes": {}
|
| 224 |
+
}
|
| 225 |
+
},
|
| 226 |
+
"total_flos": 0.0,
|
| 227 |
+
"train_batch_size": 8,
|
| 228 |
+
"trial_name": null,
|
| 229 |
+
"trial_params": null
|
| 230 |
+
}
|
ood-vifactcheck-phobert-topic-domain-01-s42-20260814T105726Z-00ca7096/trainer/checkpoint-573/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-02-s42-20260814T111300Z-00ca7096/trainer/checkpoint-2002/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-topic-domain-02-s42-20260814T111300Z-00ca7096/trainer/checkpoint-2002/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-02-s42-20260814T111300Z-00ca7096/trainer/checkpoint-2002/special_tokens_map.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": "<mask>",
|
| 6 |
+
"pad_token": "<pad>",
|
| 7 |
+
"sep_token": "</s>",
|
| 8 |
+
"unk_token": "<unk>"
|
| 9 |
+
}
|
ood-vifactcheck-phobert-topic-domain-02-s42-20260814T111300Z-00ca7096/trainer/checkpoint-2002/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 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": "</s>",
|
| 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 |
+
"64000": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 256,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "PhobertTokenizer",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-phobert-topic-domain-02-s42-20260814T111300Z-00ca7096/trainer/checkpoint-2002/trainer_state.json
ADDED
|
@@ -0,0 +1,724 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 2002,
|
| 3 |
+
"best_metric": 0.8861438281728137,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-topic-domain-02-s42-20260814T111300Z-00ca7096/trainer/checkpoint-2002",
|
| 5 |
+
"epoch": 11.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 2002,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.13793103448275862,
|
| 14 |
+
"grad_norm": 6.980297088623047,
|
| 15 |
+
"learning_rate": 1.3186813186813187e-06,
|
| 16 |
+
"loss": 4.399,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.27586206896551724,
|
| 21 |
+
"grad_norm": 6.4536871910095215,
|
| 22 |
+
"learning_rate": 2.6923076923076923e-06,
|
| 23 |
+
"loss": 4.4281,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.41379310344827586,
|
| 28 |
+
"grad_norm": 7.175063610076904,
|
| 29 |
+
"learning_rate": 4.065934065934066e-06,
|
| 30 |
+
"loss": 4.3793,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5517241379310345,
|
| 35 |
+
"grad_norm": 6.066775798797607,
|
| 36 |
+
"learning_rate": 5.43956043956044e-06,
|
| 37 |
+
"loss": 4.392,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6896551724137931,
|
| 42 |
+
"grad_norm": 6.8952250480651855,
|
| 43 |
+
"learning_rate": 6.813186813186814e-06,
|
| 44 |
+
"loss": 4.3561,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8275862068965517,
|
| 49 |
+
"grad_norm": 9.431622505187988,
|
| 50 |
+
"learning_rate": 8.186813186813188e-06,
|
| 51 |
+
"loss": 4.243,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9655172413793104,
|
| 56 |
+
"grad_norm": 13.363086700439453,
|
| 57 |
+
"learning_rate": 9.560439560439562e-06,
|
| 58 |
+
"loss": 4.1192,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.4970178926441352,
|
| 64 |
+
"eval_loss": 0.9695770740509033,
|
| 65 |
+
"eval_macro_f1": 0.448147915882771,
|
| 66 |
+
"eval_runtime": 1.1538,
|
| 67 |
+
"eval_samples_per_second": 435.953,
|
| 68 |
+
"eval_steps_per_second": 27.735,
|
| 69 |
+
"eval_weighted_f1": 0.4499735979374391,
|
| 70 |
+
"step": 182
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0993103448275863,
|
| 74 |
+
"grad_norm": 16.97210121154785,
|
| 75 |
+
"learning_rate": 1.0934065934065933e-05,
|
| 76 |
+
"loss": 3.6456,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.237241379310345,
|
| 81 |
+
"grad_norm": 35.87561798095703,
|
| 82 |
+
"learning_rate": 1.230769230769231e-05,
|
| 83 |
+
"loss": 3.4369,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.3751724137931034,
|
| 88 |
+
"grad_norm": 23.082792282104492,
|
| 89 |
+
"learning_rate": 1.3681318681318683e-05,
|
| 90 |
+
"loss": 3.0547,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.513103448275862,
|
| 95 |
+
"grad_norm": 30.00188636779785,
|
| 96 |
+
"learning_rate": 1.5054945054945056e-05,
|
| 97 |
+
"loss": 2.7245,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.6510344827586207,
|
| 102 |
+
"grad_norm": 41.87844467163086,
|
| 103 |
+
"learning_rate": 1.642857142857143e-05,
|
| 104 |
+
"loss": 2.462,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7889655172413792,
|
| 109 |
+
"grad_norm": 42.26472473144531,
|
| 110 |
+
"learning_rate": 1.78021978021978e-05,
|
| 111 |
+
"loss": 2.3049,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.926896551724138,
|
| 116 |
+
"grad_norm": 25.176780700683594,
|
| 117 |
+
"learning_rate": 1.9175824175824178e-05,
|
| 118 |
+
"loss": 2.2955,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.8409542743538767,
|
| 124 |
+
"eval_loss": 0.43624192476272583,
|
| 125 |
+
"eval_macro_f1": 0.8389594725457213,
|
| 126 |
+
"eval_runtime": 1.1094,
|
| 127 |
+
"eval_samples_per_second": 453.416,
|
| 128 |
+
"eval_steps_per_second": 28.846,
|
| 129 |
+
"eval_weighted_f1": 0.839276355898203,
|
| 130 |
+
"step": 364
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.0606896551724136,
|
| 134 |
+
"grad_norm": 27.015731811523438,
|
| 135 |
+
"learning_rate": 1.993894993894994e-05,
|
| 136 |
+
"loss": 1.7241,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.1986206896551725,
|
| 141 |
+
"grad_norm": 41.06747817993164,
|
| 142 |
+
"learning_rate": 1.9786324786324788e-05,
|
| 143 |
+
"loss": 1.6777,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.336551724137931,
|
| 148 |
+
"grad_norm": 39.86870574951172,
|
| 149 |
+
"learning_rate": 1.9633699633699634e-05,
|
| 150 |
+
"loss": 1.6905,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.47448275862069,
|
| 155 |
+
"grad_norm": 54.517059326171875,
|
| 156 |
+
"learning_rate": 1.948107448107448e-05,
|
| 157 |
+
"loss": 1.7093,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.6124137931034483,
|
| 162 |
+
"grad_norm": 33.706809997558594,
|
| 163 |
+
"learning_rate": 1.932844932844933e-05,
|
| 164 |
+
"loss": 1.7227,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.750344827586207,
|
| 169 |
+
"grad_norm": 44.67002868652344,
|
| 170 |
+
"learning_rate": 1.9175824175824178e-05,
|
| 171 |
+
"loss": 1.5474,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.8882758620689657,
|
| 176 |
+
"grad_norm": 31.4476375579834,
|
| 177 |
+
"learning_rate": 1.9023199023199025e-05,
|
| 178 |
+
"loss": 1.6935,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 3.0,
|
| 183 |
+
"eval_accuracy": 0.8409542743538767,
|
| 184 |
+
"eval_loss": 0.4332064390182495,
|
| 185 |
+
"eval_macro_f1": 0.8417257719991209,
|
| 186 |
+
"eval_runtime": 1.1482,
|
| 187 |
+
"eval_samples_per_second": 438.082,
|
| 188 |
+
"eval_steps_per_second": 27.87,
|
| 189 |
+
"eval_weighted_f1": 0.8416960520979491,
|
| 190 |
+
"step": 546
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.0220689655172412,
|
| 194 |
+
"grad_norm": 123.35029602050781,
|
| 195 |
+
"learning_rate": 1.887057387057387e-05,
|
| 196 |
+
"loss": 1.4182,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.16,
|
| 201 |
+
"grad_norm": 12.98510456085205,
|
| 202 |
+
"learning_rate": 1.8717948717948718e-05,
|
| 203 |
+
"loss": 0.9826,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.2979310344827586,
|
| 208 |
+
"grad_norm": 40.543701171875,
|
| 209 |
+
"learning_rate": 1.8565323565323568e-05,
|
| 210 |
+
"loss": 0.9675,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.435862068965517,
|
| 215 |
+
"grad_norm": 38.83287048339844,
|
| 216 |
+
"learning_rate": 1.8412698412698415e-05,
|
| 217 |
+
"loss": 1.0758,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.573793103448276,
|
| 222 |
+
"grad_norm": 13.0289306640625,
|
| 223 |
+
"learning_rate": 1.826007326007326e-05,
|
| 224 |
+
"loss": 1.3356,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.7117241379310344,
|
| 229 |
+
"grad_norm": 45.25871276855469,
|
| 230 |
+
"learning_rate": 1.8107448107448108e-05,
|
| 231 |
+
"loss": 1.2045,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.849655172413793,
|
| 236 |
+
"grad_norm": 55.038883209228516,
|
| 237 |
+
"learning_rate": 1.7954822954822958e-05,
|
| 238 |
+
"loss": 1.0059,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.9875862068965517,
|
| 243 |
+
"grad_norm": 25.700538635253906,
|
| 244 |
+
"learning_rate": 1.78021978021978e-05,
|
| 245 |
+
"loss": 1.1334,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 4.0,
|
| 250 |
+
"eval_accuracy": 0.8608349900596421,
|
| 251 |
+
"eval_loss": 0.38739511370658875,
|
| 252 |
+
"eval_macro_f1": 0.8595981063726925,
|
| 253 |
+
"eval_runtime": 1.1633,
|
| 254 |
+
"eval_samples_per_second": 432.377,
|
| 255 |
+
"eval_steps_per_second": 27.507,
|
| 256 |
+
"eval_weighted_f1": 0.8599470021984109,
|
| 257 |
+
"step": 728
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.121379310344827,
|
| 261 |
+
"grad_norm": 32.05792236328125,
|
| 262 |
+
"learning_rate": 1.764957264957265e-05,
|
| 263 |
+
"loss": 0.7435,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.259310344827586,
|
| 268 |
+
"grad_norm": 43.12069320678711,
|
| 269 |
+
"learning_rate": 1.74969474969475e-05,
|
| 270 |
+
"loss": 0.7742,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.397241379310345,
|
| 275 |
+
"grad_norm": 70.11759185791016,
|
| 276 |
+
"learning_rate": 1.7344322344322345e-05,
|
| 277 |
+
"loss": 0.8623,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.535172413793103,
|
| 282 |
+
"grad_norm": 81.32560729980469,
|
| 283 |
+
"learning_rate": 1.7191697191697195e-05,
|
| 284 |
+
"loss": 0.7702,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.673103448275862,
|
| 289 |
+
"grad_norm": 48.563751220703125,
|
| 290 |
+
"learning_rate": 1.703907203907204e-05,
|
| 291 |
+
"loss": 0.6305,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.811034482758621,
|
| 296 |
+
"grad_norm": 41.496421813964844,
|
| 297 |
+
"learning_rate": 1.688644688644689e-05,
|
| 298 |
+
"loss": 0.8142,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.94896551724138,
|
| 303 |
+
"grad_norm": 67.10346221923828,
|
| 304 |
+
"learning_rate": 1.6733821733821735e-05,
|
| 305 |
+
"loss": 0.7629,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 5.0,
|
| 310 |
+
"eval_accuracy": 0.8707753479125249,
|
| 311 |
+
"eval_loss": 0.4054329991340637,
|
| 312 |
+
"eval_macro_f1": 0.8702965801339321,
|
| 313 |
+
"eval_runtime": 1.1611,
|
| 314 |
+
"eval_samples_per_second": 433.199,
|
| 315 |
+
"eval_steps_per_second": 27.559,
|
| 316 |
+
"eval_weighted_f1": 0.8704710097579873,
|
| 317 |
+
"step": 910
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.082758620689655,
|
| 321 |
+
"grad_norm": 134.0499267578125,
|
| 322 |
+
"learning_rate": 1.6581196581196585e-05,
|
| 323 |
+
"loss": 0.5674,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.220689655172414,
|
| 328 |
+
"grad_norm": 46.29956817626953,
|
| 329 |
+
"learning_rate": 1.642857142857143e-05,
|
| 330 |
+
"loss": 0.6318,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.358620689655172,
|
| 335 |
+
"grad_norm": 148.27769470214844,
|
| 336 |
+
"learning_rate": 1.627594627594628e-05,
|
| 337 |
+
"loss": 0.5845,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.496551724137931,
|
| 342 |
+
"grad_norm": 32.090579986572266,
|
| 343 |
+
"learning_rate": 1.6123321123321125e-05,
|
| 344 |
+
"loss": 0.5006,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.63448275862069,
|
| 349 |
+
"grad_norm": 1.1076809167861938,
|
| 350 |
+
"learning_rate": 1.5970695970695972e-05,
|
| 351 |
+
"loss": 0.443,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.772413793103448,
|
| 356 |
+
"grad_norm": 54.46884536743164,
|
| 357 |
+
"learning_rate": 1.581807081807082e-05,
|
| 358 |
+
"loss": 0.7634,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.910344827586207,
|
| 363 |
+
"grad_norm": 23.109272003173828,
|
| 364 |
+
"learning_rate": 1.5665445665445665e-05,
|
| 365 |
+
"loss": 0.6585,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 6.0,
|
| 370 |
+
"eval_accuracy": 0.8687872763419483,
|
| 371 |
+
"eval_loss": 0.4956939220428467,
|
| 372 |
+
"eval_macro_f1": 0.867316765748193,
|
| 373 |
+
"eval_runtime": 1.1019,
|
| 374 |
+
"eval_samples_per_second": 456.499,
|
| 375 |
+
"eval_steps_per_second": 29.042,
|
| 376 |
+
"eval_weighted_f1": 0.8673513739052507,
|
| 377 |
+
"step": 1092
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 6.0441379310344825,
|
| 381 |
+
"grad_norm": 2.0426721572875977,
|
| 382 |
+
"learning_rate": 1.5512820512820516e-05,
|
| 383 |
+
"loss": 0.4618,
|
| 384 |
+
"step": 1100
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 6.182068965517241,
|
| 388 |
+
"grad_norm": 0.4476012885570526,
|
| 389 |
+
"learning_rate": 1.5360195360195362e-05,
|
| 390 |
+
"loss": 0.3361,
|
| 391 |
+
"step": 1125
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 6.32,
|
| 395 |
+
"grad_norm": 13.54196834564209,
|
| 396 |
+
"learning_rate": 1.5207570207570207e-05,
|
| 397 |
+
"loss": 0.4302,
|
| 398 |
+
"step": 1150
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.457931034482758,
|
| 402 |
+
"grad_norm": 0.7900168299674988,
|
| 403 |
+
"learning_rate": 1.5054945054945056e-05,
|
| 404 |
+
"loss": 0.3475,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.595862068965517,
|
| 409 |
+
"grad_norm": 0.5792030096054077,
|
| 410 |
+
"learning_rate": 1.4902319902319904e-05,
|
| 411 |
+
"loss": 0.3242,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.733793103448276,
|
| 416 |
+
"grad_norm": 89.2833023071289,
|
| 417 |
+
"learning_rate": 1.4749694749694752e-05,
|
| 418 |
+
"loss": 0.8001,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.871724137931034,
|
| 423 |
+
"grad_norm": 16.700443267822266,
|
| 424 |
+
"learning_rate": 1.4597069597069597e-05,
|
| 425 |
+
"loss": 0.6429,
|
| 426 |
+
"step": 1250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 7.0,
|
| 430 |
+
"eval_accuracy": 0.8727634194831014,
|
| 431 |
+
"eval_loss": 0.5787009596824646,
|
| 432 |
+
"eval_macro_f1": 0.8721713817707942,
|
| 433 |
+
"eval_runtime": 1.1044,
|
| 434 |
+
"eval_samples_per_second": 455.469,
|
| 435 |
+
"eval_steps_per_second": 28.976,
|
| 436 |
+
"eval_weighted_f1": 0.8721134073440937,
|
| 437 |
+
"step": 1274
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 7.0055172413793105,
|
| 441 |
+
"grad_norm": 0.33309298753738403,
|
| 442 |
+
"learning_rate": 1.4444444444444446e-05,
|
| 443 |
+
"loss": 0.2602,
|
| 444 |
+
"step": 1275
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 7.1434482758620685,
|
| 448 |
+
"grad_norm": 0.4772789478302002,
|
| 449 |
+
"learning_rate": 1.4291819291819293e-05,
|
| 450 |
+
"loss": 0.3047,
|
| 451 |
+
"step": 1300
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 7.281379310344827,
|
| 455 |
+
"grad_norm": 0.4417325556278229,
|
| 456 |
+
"learning_rate": 1.4139194139194141e-05,
|
| 457 |
+
"loss": 0.2396,
|
| 458 |
+
"step": 1325
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 7.419310344827586,
|
| 462 |
+
"grad_norm": 0.24772858619689941,
|
| 463 |
+
"learning_rate": 1.3986568986568988e-05,
|
| 464 |
+
"loss": 0.2161,
|
| 465 |
+
"step": 1350
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 7.557241379310344,
|
| 469 |
+
"grad_norm": 94.29884338378906,
|
| 470 |
+
"learning_rate": 1.3833943833943834e-05,
|
| 471 |
+
"loss": 0.2722,
|
| 472 |
+
"step": 1375
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 7.695172413793103,
|
| 476 |
+
"grad_norm": 0.24893832206726074,
|
| 477 |
+
"learning_rate": 1.3681318681318683e-05,
|
| 478 |
+
"loss": 0.4054,
|
| 479 |
+
"step": 1400
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 7.833103448275862,
|
| 483 |
+
"grad_norm": 3.7579987049102783,
|
| 484 |
+
"learning_rate": 1.3528693528693531e-05,
|
| 485 |
+
"loss": 0.4109,
|
| 486 |
+
"step": 1425
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 7.971034482758621,
|
| 490 |
+
"grad_norm": 101.4913558959961,
|
| 491 |
+
"learning_rate": 1.3376068376068376e-05,
|
| 492 |
+
"loss": 0.46,
|
| 493 |
+
"step": 1450
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 8.0,
|
| 497 |
+
"eval_accuracy": 0.8727634194831014,
|
| 498 |
+
"eval_loss": 0.5464478135108948,
|
| 499 |
+
"eval_macro_f1": 0.87217750287882,
|
| 500 |
+
"eval_runtime": 1.0988,
|
| 501 |
+
"eval_samples_per_second": 457.777,
|
| 502 |
+
"eval_steps_per_second": 29.123,
|
| 503 |
+
"eval_weighted_f1": 0.8722226283191905,
|
| 504 |
+
"step": 1456
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 8.104827586206897,
|
| 508 |
+
"grad_norm": 169.70433044433594,
|
| 509 |
+
"learning_rate": 1.3223443223443224e-05,
|
| 510 |
+
"loss": 0.3612,
|
| 511 |
+
"step": 1475
|
| 512 |
+
},
|
| 513 |
+
{
|
| 514 |
+
"epoch": 8.242758620689655,
|
| 515 |
+
"grad_norm": 24.66989517211914,
|
| 516 |
+
"learning_rate": 1.3070818070818073e-05,
|
| 517 |
+
"loss": 0.3827,
|
| 518 |
+
"step": 1500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 8.380689655172414,
|
| 522 |
+
"grad_norm": 0.23338928818702698,
|
| 523 |
+
"learning_rate": 1.291819291819292e-05,
|
| 524 |
+
"loss": 0.1306,
|
| 525 |
+
"step": 1525
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 8.518620689655172,
|
| 529 |
+
"grad_norm": 0.19539080560207367,
|
| 530 |
+
"learning_rate": 1.2765567765567766e-05,
|
| 531 |
+
"loss": 0.3355,
|
| 532 |
+
"step": 1550
|
| 533 |
+
},
|
| 534 |
+
{
|
| 535 |
+
"epoch": 8.65655172413793,
|
| 536 |
+
"grad_norm": 103.85006713867188,
|
| 537 |
+
"learning_rate": 1.2612942612942613e-05,
|
| 538 |
+
"loss": 0.1294,
|
| 539 |
+
"step": 1575
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 8.79448275862069,
|
| 543 |
+
"grad_norm": 1.6145248413085938,
|
| 544 |
+
"learning_rate": 1.2460317460317461e-05,
|
| 545 |
+
"loss": 0.2027,
|
| 546 |
+
"step": 1600
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 8.932413793103448,
|
| 550 |
+
"grad_norm": 0.3407277762889862,
|
| 551 |
+
"learning_rate": 1.230769230769231e-05,
|
| 552 |
+
"loss": 0.3451,
|
| 553 |
+
"step": 1625
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 9.0,
|
| 557 |
+
"eval_accuracy": 0.8588469184890656,
|
| 558 |
+
"eval_loss": 0.6675087213516235,
|
| 559 |
+
"eval_macro_f1": 0.8575896178741614,
|
| 560 |
+
"eval_runtime": 1.1135,
|
| 561 |
+
"eval_samples_per_second": 451.726,
|
| 562 |
+
"eval_steps_per_second": 28.738,
|
| 563 |
+
"eval_weighted_f1": 0.8576990957746697,
|
| 564 |
+
"step": 1638
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"epoch": 9.066206896551725,
|
| 568 |
+
"grad_norm": 0.20673930644989014,
|
| 569 |
+
"learning_rate": 1.2155067155067155e-05,
|
| 570 |
+
"loss": 0.229,
|
| 571 |
+
"step": 1650
|
| 572 |
+
},
|
| 573 |
+
{
|
| 574 |
+
"epoch": 9.204137931034483,
|
| 575 |
+
"grad_norm": 0.19027166068553925,
|
| 576 |
+
"learning_rate": 1.2002442002442003e-05,
|
| 577 |
+
"loss": 0.1645,
|
| 578 |
+
"step": 1675
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 9.34206896551724,
|
| 582 |
+
"grad_norm": 115.7024917602539,
|
| 583 |
+
"learning_rate": 1.1849816849816852e-05,
|
| 584 |
+
"loss": 0.2544,
|
| 585 |
+
"step": 1700
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 9.48,
|
| 589 |
+
"grad_norm": 2.5617661476135254,
|
| 590 |
+
"learning_rate": 1.1697191697191698e-05,
|
| 591 |
+
"loss": 0.1958,
|
| 592 |
+
"step": 1725
|
| 593 |
+
},
|
| 594 |
+
{
|
| 595 |
+
"epoch": 9.617931034482758,
|
| 596 |
+
"grad_norm": 0.16203327476978302,
|
| 597 |
+
"learning_rate": 1.1544566544566545e-05,
|
| 598 |
+
"loss": 0.2091,
|
| 599 |
+
"step": 1750
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 9.755862068965516,
|
| 603 |
+
"grad_norm": 0.16939979791641235,
|
| 604 |
+
"learning_rate": 1.1391941391941392e-05,
|
| 605 |
+
"loss": 0.2617,
|
| 606 |
+
"step": 1775
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"epoch": 9.893793103448276,
|
| 610 |
+
"grad_norm": 0.1613183617591858,
|
| 611 |
+
"learning_rate": 1.123931623931624e-05,
|
| 612 |
+
"loss": 0.1845,
|
| 613 |
+
"step": 1800
|
| 614 |
+
},
|
| 615 |
+
{
|
| 616 |
+
"epoch": 10.0,
|
| 617 |
+
"eval_accuracy": 0.8767395626242545,
|
| 618 |
+
"eval_loss": 0.685275673866272,
|
| 619 |
+
"eval_macro_f1": 0.8751112507855575,
|
| 620 |
+
"eval_runtime": 1.1147,
|
| 621 |
+
"eval_samples_per_second": 451.231,
|
| 622 |
+
"eval_steps_per_second": 28.707,
|
| 623 |
+
"eval_weighted_f1": 0.8752190573327314,
|
| 624 |
+
"step": 1820
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"epoch": 10.027586206896551,
|
| 628 |
+
"grad_norm": 0.1641041785478592,
|
| 629 |
+
"learning_rate": 1.1086691086691088e-05,
|
| 630 |
+
"loss": 0.079,
|
| 631 |
+
"step": 1825
|
| 632 |
+
},
|
| 633 |
+
{
|
| 634 |
+
"epoch": 10.16551724137931,
|
| 635 |
+
"grad_norm": 0.14107441902160645,
|
| 636 |
+
"learning_rate": 1.0934065934065933e-05,
|
| 637 |
+
"loss": 0.1898,
|
| 638 |
+
"step": 1850
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 10.303448275862069,
|
| 642 |
+
"grad_norm": 1.0427343845367432,
|
| 643 |
+
"learning_rate": 1.0781440781440782e-05,
|
| 644 |
+
"loss": 0.0603,
|
| 645 |
+
"step": 1875
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 10.441379310344828,
|
| 649 |
+
"grad_norm": 27.041128158569336,
|
| 650 |
+
"learning_rate": 1.062881562881563e-05,
|
| 651 |
+
"loss": 0.1245,
|
| 652 |
+
"step": 1900
|
| 653 |
+
},
|
| 654 |
+
{
|
| 655 |
+
"epoch": 10.579310344827586,
|
| 656 |
+
"grad_norm": 101.82066345214844,
|
| 657 |
+
"learning_rate": 1.0476190476190477e-05,
|
| 658 |
+
"loss": 0.1608,
|
| 659 |
+
"step": 1925
|
| 660 |
+
},
|
| 661 |
+
{
|
| 662 |
+
"epoch": 10.717241379310344,
|
| 663 |
+
"grad_norm": 0.14787974953651428,
|
| 664 |
+
"learning_rate": 1.0323565323565325e-05,
|
| 665 |
+
"loss": 0.125,
|
| 666 |
+
"step": 1950
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"epoch": 10.855172413793104,
|
| 670 |
+
"grad_norm": 0.13074085116386414,
|
| 671 |
+
"learning_rate": 1.0170940170940172e-05,
|
| 672 |
+
"loss": 0.1203,
|
| 673 |
+
"step": 1975
|
| 674 |
+
},
|
| 675 |
+
{
|
| 676 |
+
"epoch": 10.993103448275862,
|
| 677 |
+
"grad_norm": 151.71702575683594,
|
| 678 |
+
"learning_rate": 1.0018315018315019e-05,
|
| 679 |
+
"loss": 0.2005,
|
| 680 |
+
"step": 2000
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 11.0,
|
| 684 |
+
"eval_accuracy": 0.8866799204771372,
|
| 685 |
+
"eval_loss": 0.6383737325668335,
|
| 686 |
+
"eval_macro_f1": 0.8861438281728137,
|
| 687 |
+
"eval_runtime": 1.121,
|
| 688 |
+
"eval_samples_per_second": 448.693,
|
| 689 |
+
"eval_steps_per_second": 28.545,
|
| 690 |
+
"eval_weighted_f1": 0.8863145449713001,
|
| 691 |
+
"step": 2002
|
| 692 |
+
}
|
| 693 |
+
],
|
| 694 |
+
"logging_steps": 25,
|
| 695 |
+
"max_steps": 3640,
|
| 696 |
+
"num_input_tokens_seen": 0,
|
| 697 |
+
"num_train_epochs": 20,
|
| 698 |
+
"save_steps": 500,
|
| 699 |
+
"stateful_callbacks": {
|
| 700 |
+
"EarlyStoppingCallback": {
|
| 701 |
+
"args": {
|
| 702 |
+
"early_stopping_patience": 3,
|
| 703 |
+
"early_stopping_threshold": 0.0
|
| 704 |
+
},
|
| 705 |
+
"attributes": {
|
| 706 |
+
"early_stopping_patience_counter": 0
|
| 707 |
+
}
|
| 708 |
+
},
|
| 709 |
+
"TrainerControl": {
|
| 710 |
+
"args": {
|
| 711 |
+
"should_epoch_stop": false,
|
| 712 |
+
"should_evaluate": false,
|
| 713 |
+
"should_log": false,
|
| 714 |
+
"should_save": true,
|
| 715 |
+
"should_training_stop": false
|
| 716 |
+
},
|
| 717 |
+
"attributes": {}
|
| 718 |
+
}
|
| 719 |
+
},
|
| 720 |
+
"total_flos": 0.0,
|
| 721 |
+
"train_batch_size": 8,
|
| 722 |
+
"trial_name": null,
|
| 723 |
+
"trial_params": null
|
| 724 |
+
}
|
ood-vifactcheck-phobert-topic-domain-02-s42-20260814T111300Z-00ca7096/trainer/checkpoint-2002/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-02-s42-20260814T111300Z-00ca7096/trainer/checkpoint-2548/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-topic-domain-02-s42-20260814T111300Z-00ca7096/trainer/checkpoint-2548/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-02-s42-20260814T111300Z-00ca7096/trainer/checkpoint-2548/special_tokens_map.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": "<mask>",
|
| 6 |
+
"pad_token": "<pad>",
|
| 7 |
+
"sep_token": "</s>",
|
| 8 |
+
"unk_token": "<unk>"
|
| 9 |
+
}
|
ood-vifactcheck-phobert-topic-domain-02-s42-20260814T111300Z-00ca7096/trainer/checkpoint-2548/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 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": "</s>",
|
| 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 |
+
"64000": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 256,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "PhobertTokenizer",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-phobert-topic-domain-02-s42-20260814T111300Z-00ca7096/trainer/checkpoint-2548/trainer_state.json
ADDED
|
@@ -0,0 +1,904 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 2002,
|
| 3 |
+
"best_metric": 0.8861438281728137,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-topic-domain-02-s42-20260814T111300Z-00ca7096/trainer/checkpoint-2002",
|
| 5 |
+
"epoch": 14.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 2548,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.13793103448275862,
|
| 14 |
+
"grad_norm": 6.980297088623047,
|
| 15 |
+
"learning_rate": 1.3186813186813187e-06,
|
| 16 |
+
"loss": 4.399,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.27586206896551724,
|
| 21 |
+
"grad_norm": 6.4536871910095215,
|
| 22 |
+
"learning_rate": 2.6923076923076923e-06,
|
| 23 |
+
"loss": 4.4281,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.41379310344827586,
|
| 28 |
+
"grad_norm": 7.175063610076904,
|
| 29 |
+
"learning_rate": 4.065934065934066e-06,
|
| 30 |
+
"loss": 4.3793,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5517241379310345,
|
| 35 |
+
"grad_norm": 6.066775798797607,
|
| 36 |
+
"learning_rate": 5.43956043956044e-06,
|
| 37 |
+
"loss": 4.392,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6896551724137931,
|
| 42 |
+
"grad_norm": 6.8952250480651855,
|
| 43 |
+
"learning_rate": 6.813186813186814e-06,
|
| 44 |
+
"loss": 4.3561,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8275862068965517,
|
| 49 |
+
"grad_norm": 9.431622505187988,
|
| 50 |
+
"learning_rate": 8.186813186813188e-06,
|
| 51 |
+
"loss": 4.243,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9655172413793104,
|
| 56 |
+
"grad_norm": 13.363086700439453,
|
| 57 |
+
"learning_rate": 9.560439560439562e-06,
|
| 58 |
+
"loss": 4.1192,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.4970178926441352,
|
| 64 |
+
"eval_loss": 0.9695770740509033,
|
| 65 |
+
"eval_macro_f1": 0.448147915882771,
|
| 66 |
+
"eval_runtime": 1.1538,
|
| 67 |
+
"eval_samples_per_second": 435.953,
|
| 68 |
+
"eval_steps_per_second": 27.735,
|
| 69 |
+
"eval_weighted_f1": 0.4499735979374391,
|
| 70 |
+
"step": 182
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0993103448275863,
|
| 74 |
+
"grad_norm": 16.97210121154785,
|
| 75 |
+
"learning_rate": 1.0934065934065933e-05,
|
| 76 |
+
"loss": 3.6456,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.237241379310345,
|
| 81 |
+
"grad_norm": 35.87561798095703,
|
| 82 |
+
"learning_rate": 1.230769230769231e-05,
|
| 83 |
+
"loss": 3.4369,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.3751724137931034,
|
| 88 |
+
"grad_norm": 23.082792282104492,
|
| 89 |
+
"learning_rate": 1.3681318681318683e-05,
|
| 90 |
+
"loss": 3.0547,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.513103448275862,
|
| 95 |
+
"grad_norm": 30.00188636779785,
|
| 96 |
+
"learning_rate": 1.5054945054945056e-05,
|
| 97 |
+
"loss": 2.7245,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.6510344827586207,
|
| 102 |
+
"grad_norm": 41.87844467163086,
|
| 103 |
+
"learning_rate": 1.642857142857143e-05,
|
| 104 |
+
"loss": 2.462,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7889655172413792,
|
| 109 |
+
"grad_norm": 42.26472473144531,
|
| 110 |
+
"learning_rate": 1.78021978021978e-05,
|
| 111 |
+
"loss": 2.3049,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.926896551724138,
|
| 116 |
+
"grad_norm": 25.176780700683594,
|
| 117 |
+
"learning_rate": 1.9175824175824178e-05,
|
| 118 |
+
"loss": 2.2955,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.8409542743538767,
|
| 124 |
+
"eval_loss": 0.43624192476272583,
|
| 125 |
+
"eval_macro_f1": 0.8389594725457213,
|
| 126 |
+
"eval_runtime": 1.1094,
|
| 127 |
+
"eval_samples_per_second": 453.416,
|
| 128 |
+
"eval_steps_per_second": 28.846,
|
| 129 |
+
"eval_weighted_f1": 0.839276355898203,
|
| 130 |
+
"step": 364
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.0606896551724136,
|
| 134 |
+
"grad_norm": 27.015731811523438,
|
| 135 |
+
"learning_rate": 1.993894993894994e-05,
|
| 136 |
+
"loss": 1.7241,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.1986206896551725,
|
| 141 |
+
"grad_norm": 41.06747817993164,
|
| 142 |
+
"learning_rate": 1.9786324786324788e-05,
|
| 143 |
+
"loss": 1.6777,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.336551724137931,
|
| 148 |
+
"grad_norm": 39.86870574951172,
|
| 149 |
+
"learning_rate": 1.9633699633699634e-05,
|
| 150 |
+
"loss": 1.6905,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.47448275862069,
|
| 155 |
+
"grad_norm": 54.517059326171875,
|
| 156 |
+
"learning_rate": 1.948107448107448e-05,
|
| 157 |
+
"loss": 1.7093,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.6124137931034483,
|
| 162 |
+
"grad_norm": 33.706809997558594,
|
| 163 |
+
"learning_rate": 1.932844932844933e-05,
|
| 164 |
+
"loss": 1.7227,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.750344827586207,
|
| 169 |
+
"grad_norm": 44.67002868652344,
|
| 170 |
+
"learning_rate": 1.9175824175824178e-05,
|
| 171 |
+
"loss": 1.5474,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.8882758620689657,
|
| 176 |
+
"grad_norm": 31.4476375579834,
|
| 177 |
+
"learning_rate": 1.9023199023199025e-05,
|
| 178 |
+
"loss": 1.6935,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 3.0,
|
| 183 |
+
"eval_accuracy": 0.8409542743538767,
|
| 184 |
+
"eval_loss": 0.4332064390182495,
|
| 185 |
+
"eval_macro_f1": 0.8417257719991209,
|
| 186 |
+
"eval_runtime": 1.1482,
|
| 187 |
+
"eval_samples_per_second": 438.082,
|
| 188 |
+
"eval_steps_per_second": 27.87,
|
| 189 |
+
"eval_weighted_f1": 0.8416960520979491,
|
| 190 |
+
"step": 546
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.0220689655172412,
|
| 194 |
+
"grad_norm": 123.35029602050781,
|
| 195 |
+
"learning_rate": 1.887057387057387e-05,
|
| 196 |
+
"loss": 1.4182,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.16,
|
| 201 |
+
"grad_norm": 12.98510456085205,
|
| 202 |
+
"learning_rate": 1.8717948717948718e-05,
|
| 203 |
+
"loss": 0.9826,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.2979310344827586,
|
| 208 |
+
"grad_norm": 40.543701171875,
|
| 209 |
+
"learning_rate": 1.8565323565323568e-05,
|
| 210 |
+
"loss": 0.9675,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.435862068965517,
|
| 215 |
+
"grad_norm": 38.83287048339844,
|
| 216 |
+
"learning_rate": 1.8412698412698415e-05,
|
| 217 |
+
"loss": 1.0758,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.573793103448276,
|
| 222 |
+
"grad_norm": 13.0289306640625,
|
| 223 |
+
"learning_rate": 1.826007326007326e-05,
|
| 224 |
+
"loss": 1.3356,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.7117241379310344,
|
| 229 |
+
"grad_norm": 45.25871276855469,
|
| 230 |
+
"learning_rate": 1.8107448107448108e-05,
|
| 231 |
+
"loss": 1.2045,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.849655172413793,
|
| 236 |
+
"grad_norm": 55.038883209228516,
|
| 237 |
+
"learning_rate": 1.7954822954822958e-05,
|
| 238 |
+
"loss": 1.0059,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.9875862068965517,
|
| 243 |
+
"grad_norm": 25.700538635253906,
|
| 244 |
+
"learning_rate": 1.78021978021978e-05,
|
| 245 |
+
"loss": 1.1334,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 4.0,
|
| 250 |
+
"eval_accuracy": 0.8608349900596421,
|
| 251 |
+
"eval_loss": 0.38739511370658875,
|
| 252 |
+
"eval_macro_f1": 0.8595981063726925,
|
| 253 |
+
"eval_runtime": 1.1633,
|
| 254 |
+
"eval_samples_per_second": 432.377,
|
| 255 |
+
"eval_steps_per_second": 27.507,
|
| 256 |
+
"eval_weighted_f1": 0.8599470021984109,
|
| 257 |
+
"step": 728
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.121379310344827,
|
| 261 |
+
"grad_norm": 32.05792236328125,
|
| 262 |
+
"learning_rate": 1.764957264957265e-05,
|
| 263 |
+
"loss": 0.7435,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.259310344827586,
|
| 268 |
+
"grad_norm": 43.12069320678711,
|
| 269 |
+
"learning_rate": 1.74969474969475e-05,
|
| 270 |
+
"loss": 0.7742,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.397241379310345,
|
| 275 |
+
"grad_norm": 70.11759185791016,
|
| 276 |
+
"learning_rate": 1.7344322344322345e-05,
|
| 277 |
+
"loss": 0.8623,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.535172413793103,
|
| 282 |
+
"grad_norm": 81.32560729980469,
|
| 283 |
+
"learning_rate": 1.7191697191697195e-05,
|
| 284 |
+
"loss": 0.7702,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.673103448275862,
|
| 289 |
+
"grad_norm": 48.563751220703125,
|
| 290 |
+
"learning_rate": 1.703907203907204e-05,
|
| 291 |
+
"loss": 0.6305,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.811034482758621,
|
| 296 |
+
"grad_norm": 41.496421813964844,
|
| 297 |
+
"learning_rate": 1.688644688644689e-05,
|
| 298 |
+
"loss": 0.8142,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.94896551724138,
|
| 303 |
+
"grad_norm": 67.10346221923828,
|
| 304 |
+
"learning_rate": 1.6733821733821735e-05,
|
| 305 |
+
"loss": 0.7629,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 5.0,
|
| 310 |
+
"eval_accuracy": 0.8707753479125249,
|
| 311 |
+
"eval_loss": 0.4054329991340637,
|
| 312 |
+
"eval_macro_f1": 0.8702965801339321,
|
| 313 |
+
"eval_runtime": 1.1611,
|
| 314 |
+
"eval_samples_per_second": 433.199,
|
| 315 |
+
"eval_steps_per_second": 27.559,
|
| 316 |
+
"eval_weighted_f1": 0.8704710097579873,
|
| 317 |
+
"step": 910
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.082758620689655,
|
| 321 |
+
"grad_norm": 134.0499267578125,
|
| 322 |
+
"learning_rate": 1.6581196581196585e-05,
|
| 323 |
+
"loss": 0.5674,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.220689655172414,
|
| 328 |
+
"grad_norm": 46.29956817626953,
|
| 329 |
+
"learning_rate": 1.642857142857143e-05,
|
| 330 |
+
"loss": 0.6318,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.358620689655172,
|
| 335 |
+
"grad_norm": 148.27769470214844,
|
| 336 |
+
"learning_rate": 1.627594627594628e-05,
|
| 337 |
+
"loss": 0.5845,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.496551724137931,
|
| 342 |
+
"grad_norm": 32.090579986572266,
|
| 343 |
+
"learning_rate": 1.6123321123321125e-05,
|
| 344 |
+
"loss": 0.5006,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.63448275862069,
|
| 349 |
+
"grad_norm": 1.1076809167861938,
|
| 350 |
+
"learning_rate": 1.5970695970695972e-05,
|
| 351 |
+
"loss": 0.443,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.772413793103448,
|
| 356 |
+
"grad_norm": 54.46884536743164,
|
| 357 |
+
"learning_rate": 1.581807081807082e-05,
|
| 358 |
+
"loss": 0.7634,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.910344827586207,
|
| 363 |
+
"grad_norm": 23.109272003173828,
|
| 364 |
+
"learning_rate": 1.5665445665445665e-05,
|
| 365 |
+
"loss": 0.6585,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 6.0,
|
| 370 |
+
"eval_accuracy": 0.8687872763419483,
|
| 371 |
+
"eval_loss": 0.4956939220428467,
|
| 372 |
+
"eval_macro_f1": 0.867316765748193,
|
| 373 |
+
"eval_runtime": 1.1019,
|
| 374 |
+
"eval_samples_per_second": 456.499,
|
| 375 |
+
"eval_steps_per_second": 29.042,
|
| 376 |
+
"eval_weighted_f1": 0.8673513739052507,
|
| 377 |
+
"step": 1092
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 6.0441379310344825,
|
| 381 |
+
"grad_norm": 2.0426721572875977,
|
| 382 |
+
"learning_rate": 1.5512820512820516e-05,
|
| 383 |
+
"loss": 0.4618,
|
| 384 |
+
"step": 1100
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 6.182068965517241,
|
| 388 |
+
"grad_norm": 0.4476012885570526,
|
| 389 |
+
"learning_rate": 1.5360195360195362e-05,
|
| 390 |
+
"loss": 0.3361,
|
| 391 |
+
"step": 1125
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 6.32,
|
| 395 |
+
"grad_norm": 13.54196834564209,
|
| 396 |
+
"learning_rate": 1.5207570207570207e-05,
|
| 397 |
+
"loss": 0.4302,
|
| 398 |
+
"step": 1150
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.457931034482758,
|
| 402 |
+
"grad_norm": 0.7900168299674988,
|
| 403 |
+
"learning_rate": 1.5054945054945056e-05,
|
| 404 |
+
"loss": 0.3475,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.595862068965517,
|
| 409 |
+
"grad_norm": 0.5792030096054077,
|
| 410 |
+
"learning_rate": 1.4902319902319904e-05,
|
| 411 |
+
"loss": 0.3242,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.733793103448276,
|
| 416 |
+
"grad_norm": 89.2833023071289,
|
| 417 |
+
"learning_rate": 1.4749694749694752e-05,
|
| 418 |
+
"loss": 0.8001,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.871724137931034,
|
| 423 |
+
"grad_norm": 16.700443267822266,
|
| 424 |
+
"learning_rate": 1.4597069597069597e-05,
|
| 425 |
+
"loss": 0.6429,
|
| 426 |
+
"step": 1250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 7.0,
|
| 430 |
+
"eval_accuracy": 0.8727634194831014,
|
| 431 |
+
"eval_loss": 0.5787009596824646,
|
| 432 |
+
"eval_macro_f1": 0.8721713817707942,
|
| 433 |
+
"eval_runtime": 1.1044,
|
| 434 |
+
"eval_samples_per_second": 455.469,
|
| 435 |
+
"eval_steps_per_second": 28.976,
|
| 436 |
+
"eval_weighted_f1": 0.8721134073440937,
|
| 437 |
+
"step": 1274
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 7.0055172413793105,
|
| 441 |
+
"grad_norm": 0.33309298753738403,
|
| 442 |
+
"learning_rate": 1.4444444444444446e-05,
|
| 443 |
+
"loss": 0.2602,
|
| 444 |
+
"step": 1275
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 7.1434482758620685,
|
| 448 |
+
"grad_norm": 0.4772789478302002,
|
| 449 |
+
"learning_rate": 1.4291819291819293e-05,
|
| 450 |
+
"loss": 0.3047,
|
| 451 |
+
"step": 1300
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 7.281379310344827,
|
| 455 |
+
"grad_norm": 0.4417325556278229,
|
| 456 |
+
"learning_rate": 1.4139194139194141e-05,
|
| 457 |
+
"loss": 0.2396,
|
| 458 |
+
"step": 1325
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 7.419310344827586,
|
| 462 |
+
"grad_norm": 0.24772858619689941,
|
| 463 |
+
"learning_rate": 1.3986568986568988e-05,
|
| 464 |
+
"loss": 0.2161,
|
| 465 |
+
"step": 1350
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 7.557241379310344,
|
| 469 |
+
"grad_norm": 94.29884338378906,
|
| 470 |
+
"learning_rate": 1.3833943833943834e-05,
|
| 471 |
+
"loss": 0.2722,
|
| 472 |
+
"step": 1375
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 7.695172413793103,
|
| 476 |
+
"grad_norm": 0.24893832206726074,
|
| 477 |
+
"learning_rate": 1.3681318681318683e-05,
|
| 478 |
+
"loss": 0.4054,
|
| 479 |
+
"step": 1400
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 7.833103448275862,
|
| 483 |
+
"grad_norm": 3.7579987049102783,
|
| 484 |
+
"learning_rate": 1.3528693528693531e-05,
|
| 485 |
+
"loss": 0.4109,
|
| 486 |
+
"step": 1425
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 7.971034482758621,
|
| 490 |
+
"grad_norm": 101.4913558959961,
|
| 491 |
+
"learning_rate": 1.3376068376068376e-05,
|
| 492 |
+
"loss": 0.46,
|
| 493 |
+
"step": 1450
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 8.0,
|
| 497 |
+
"eval_accuracy": 0.8727634194831014,
|
| 498 |
+
"eval_loss": 0.5464478135108948,
|
| 499 |
+
"eval_macro_f1": 0.87217750287882,
|
| 500 |
+
"eval_runtime": 1.0988,
|
| 501 |
+
"eval_samples_per_second": 457.777,
|
| 502 |
+
"eval_steps_per_second": 29.123,
|
| 503 |
+
"eval_weighted_f1": 0.8722226283191905,
|
| 504 |
+
"step": 1456
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 8.104827586206897,
|
| 508 |
+
"grad_norm": 169.70433044433594,
|
| 509 |
+
"learning_rate": 1.3223443223443224e-05,
|
| 510 |
+
"loss": 0.3612,
|
| 511 |
+
"step": 1475
|
| 512 |
+
},
|
| 513 |
+
{
|
| 514 |
+
"epoch": 8.242758620689655,
|
| 515 |
+
"grad_norm": 24.66989517211914,
|
| 516 |
+
"learning_rate": 1.3070818070818073e-05,
|
| 517 |
+
"loss": 0.3827,
|
| 518 |
+
"step": 1500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 8.380689655172414,
|
| 522 |
+
"grad_norm": 0.23338928818702698,
|
| 523 |
+
"learning_rate": 1.291819291819292e-05,
|
| 524 |
+
"loss": 0.1306,
|
| 525 |
+
"step": 1525
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 8.518620689655172,
|
| 529 |
+
"grad_norm": 0.19539080560207367,
|
| 530 |
+
"learning_rate": 1.2765567765567766e-05,
|
| 531 |
+
"loss": 0.3355,
|
| 532 |
+
"step": 1550
|
| 533 |
+
},
|
| 534 |
+
{
|
| 535 |
+
"epoch": 8.65655172413793,
|
| 536 |
+
"grad_norm": 103.85006713867188,
|
| 537 |
+
"learning_rate": 1.2612942612942613e-05,
|
| 538 |
+
"loss": 0.1294,
|
| 539 |
+
"step": 1575
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 8.79448275862069,
|
| 543 |
+
"grad_norm": 1.6145248413085938,
|
| 544 |
+
"learning_rate": 1.2460317460317461e-05,
|
| 545 |
+
"loss": 0.2027,
|
| 546 |
+
"step": 1600
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 8.932413793103448,
|
| 550 |
+
"grad_norm": 0.3407277762889862,
|
| 551 |
+
"learning_rate": 1.230769230769231e-05,
|
| 552 |
+
"loss": 0.3451,
|
| 553 |
+
"step": 1625
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 9.0,
|
| 557 |
+
"eval_accuracy": 0.8588469184890656,
|
| 558 |
+
"eval_loss": 0.6675087213516235,
|
| 559 |
+
"eval_macro_f1": 0.8575896178741614,
|
| 560 |
+
"eval_runtime": 1.1135,
|
| 561 |
+
"eval_samples_per_second": 451.726,
|
| 562 |
+
"eval_steps_per_second": 28.738,
|
| 563 |
+
"eval_weighted_f1": 0.8576990957746697,
|
| 564 |
+
"step": 1638
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"epoch": 9.066206896551725,
|
| 568 |
+
"grad_norm": 0.20673930644989014,
|
| 569 |
+
"learning_rate": 1.2155067155067155e-05,
|
| 570 |
+
"loss": 0.229,
|
| 571 |
+
"step": 1650
|
| 572 |
+
},
|
| 573 |
+
{
|
| 574 |
+
"epoch": 9.204137931034483,
|
| 575 |
+
"grad_norm": 0.19027166068553925,
|
| 576 |
+
"learning_rate": 1.2002442002442003e-05,
|
| 577 |
+
"loss": 0.1645,
|
| 578 |
+
"step": 1675
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 9.34206896551724,
|
| 582 |
+
"grad_norm": 115.7024917602539,
|
| 583 |
+
"learning_rate": 1.1849816849816852e-05,
|
| 584 |
+
"loss": 0.2544,
|
| 585 |
+
"step": 1700
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 9.48,
|
| 589 |
+
"grad_norm": 2.5617661476135254,
|
| 590 |
+
"learning_rate": 1.1697191697191698e-05,
|
| 591 |
+
"loss": 0.1958,
|
| 592 |
+
"step": 1725
|
| 593 |
+
},
|
| 594 |
+
{
|
| 595 |
+
"epoch": 9.617931034482758,
|
| 596 |
+
"grad_norm": 0.16203327476978302,
|
| 597 |
+
"learning_rate": 1.1544566544566545e-05,
|
| 598 |
+
"loss": 0.2091,
|
| 599 |
+
"step": 1750
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 9.755862068965516,
|
| 603 |
+
"grad_norm": 0.16939979791641235,
|
| 604 |
+
"learning_rate": 1.1391941391941392e-05,
|
| 605 |
+
"loss": 0.2617,
|
| 606 |
+
"step": 1775
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"epoch": 9.893793103448276,
|
| 610 |
+
"grad_norm": 0.1613183617591858,
|
| 611 |
+
"learning_rate": 1.123931623931624e-05,
|
| 612 |
+
"loss": 0.1845,
|
| 613 |
+
"step": 1800
|
| 614 |
+
},
|
| 615 |
+
{
|
| 616 |
+
"epoch": 10.0,
|
| 617 |
+
"eval_accuracy": 0.8767395626242545,
|
| 618 |
+
"eval_loss": 0.685275673866272,
|
| 619 |
+
"eval_macro_f1": 0.8751112507855575,
|
| 620 |
+
"eval_runtime": 1.1147,
|
| 621 |
+
"eval_samples_per_second": 451.231,
|
| 622 |
+
"eval_steps_per_second": 28.707,
|
| 623 |
+
"eval_weighted_f1": 0.8752190573327314,
|
| 624 |
+
"step": 1820
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"epoch": 10.027586206896551,
|
| 628 |
+
"grad_norm": 0.1641041785478592,
|
| 629 |
+
"learning_rate": 1.1086691086691088e-05,
|
| 630 |
+
"loss": 0.079,
|
| 631 |
+
"step": 1825
|
| 632 |
+
},
|
| 633 |
+
{
|
| 634 |
+
"epoch": 10.16551724137931,
|
| 635 |
+
"grad_norm": 0.14107441902160645,
|
| 636 |
+
"learning_rate": 1.0934065934065933e-05,
|
| 637 |
+
"loss": 0.1898,
|
| 638 |
+
"step": 1850
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 10.303448275862069,
|
| 642 |
+
"grad_norm": 1.0427343845367432,
|
| 643 |
+
"learning_rate": 1.0781440781440782e-05,
|
| 644 |
+
"loss": 0.0603,
|
| 645 |
+
"step": 1875
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 10.441379310344828,
|
| 649 |
+
"grad_norm": 27.041128158569336,
|
| 650 |
+
"learning_rate": 1.062881562881563e-05,
|
| 651 |
+
"loss": 0.1245,
|
| 652 |
+
"step": 1900
|
| 653 |
+
},
|
| 654 |
+
{
|
| 655 |
+
"epoch": 10.579310344827586,
|
| 656 |
+
"grad_norm": 101.82066345214844,
|
| 657 |
+
"learning_rate": 1.0476190476190477e-05,
|
| 658 |
+
"loss": 0.1608,
|
| 659 |
+
"step": 1925
|
| 660 |
+
},
|
| 661 |
+
{
|
| 662 |
+
"epoch": 10.717241379310344,
|
| 663 |
+
"grad_norm": 0.14787974953651428,
|
| 664 |
+
"learning_rate": 1.0323565323565325e-05,
|
| 665 |
+
"loss": 0.125,
|
| 666 |
+
"step": 1950
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"epoch": 10.855172413793104,
|
| 670 |
+
"grad_norm": 0.13074085116386414,
|
| 671 |
+
"learning_rate": 1.0170940170940172e-05,
|
| 672 |
+
"loss": 0.1203,
|
| 673 |
+
"step": 1975
|
| 674 |
+
},
|
| 675 |
+
{
|
| 676 |
+
"epoch": 10.993103448275862,
|
| 677 |
+
"grad_norm": 151.71702575683594,
|
| 678 |
+
"learning_rate": 1.0018315018315019e-05,
|
| 679 |
+
"loss": 0.2005,
|
| 680 |
+
"step": 2000
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 11.0,
|
| 684 |
+
"eval_accuracy": 0.8866799204771372,
|
| 685 |
+
"eval_loss": 0.6383737325668335,
|
| 686 |
+
"eval_macro_f1": 0.8861438281728137,
|
| 687 |
+
"eval_runtime": 1.121,
|
| 688 |
+
"eval_samples_per_second": 448.693,
|
| 689 |
+
"eval_steps_per_second": 28.545,
|
| 690 |
+
"eval_weighted_f1": 0.8863145449713001,
|
| 691 |
+
"step": 2002
|
| 692 |
+
},
|
| 693 |
+
{
|
| 694 |
+
"epoch": 11.126896551724139,
|
| 695 |
+
"grad_norm": 110.28417205810547,
|
| 696 |
+
"learning_rate": 9.865689865689867e-06,
|
| 697 |
+
"loss": 0.1009,
|
| 698 |
+
"step": 2025
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 11.264827586206897,
|
| 702 |
+
"grad_norm": 0.12746545672416687,
|
| 703 |
+
"learning_rate": 9.713064713064714e-06,
|
| 704 |
+
"loss": 0.2261,
|
| 705 |
+
"step": 2050
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"epoch": 11.402758620689655,
|
| 709 |
+
"grad_norm": 24.0385799407959,
|
| 710 |
+
"learning_rate": 9.560439560439562e-06,
|
| 711 |
+
"loss": 0.1079,
|
| 712 |
+
"step": 2075
|
| 713 |
+
},
|
| 714 |
+
{
|
| 715 |
+
"epoch": 11.540689655172415,
|
| 716 |
+
"grad_norm": 4.590055465698242,
|
| 717 |
+
"learning_rate": 9.407814407814409e-06,
|
| 718 |
+
"loss": 0.0809,
|
| 719 |
+
"step": 2100
|
| 720 |
+
},
|
| 721 |
+
{
|
| 722 |
+
"epoch": 11.678620689655173,
|
| 723 |
+
"grad_norm": 0.1077900156378746,
|
| 724 |
+
"learning_rate": 9.255189255189256e-06,
|
| 725 |
+
"loss": 0.1809,
|
| 726 |
+
"step": 2125
|
| 727 |
+
},
|
| 728 |
+
{
|
| 729 |
+
"epoch": 11.81655172413793,
|
| 730 |
+
"grad_norm": 19.242076873779297,
|
| 731 |
+
"learning_rate": 9.102564102564104e-06,
|
| 732 |
+
"loss": 0.1259,
|
| 733 |
+
"step": 2150
|
| 734 |
+
},
|
| 735 |
+
{
|
| 736 |
+
"epoch": 11.95448275862069,
|
| 737 |
+
"grad_norm": 188.9077606201172,
|
| 738 |
+
"learning_rate": 8.94993894993895e-06,
|
| 739 |
+
"loss": 0.0935,
|
| 740 |
+
"step": 2175
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 12.0,
|
| 744 |
+
"eval_accuracy": 0.878727634194831,
|
| 745 |
+
"eval_loss": 0.6814209222793579,
|
| 746 |
+
"eval_macro_f1": 0.877549288305877,
|
| 747 |
+
"eval_runtime": 1.1891,
|
| 748 |
+
"eval_samples_per_second": 423.004,
|
| 749 |
+
"eval_steps_per_second": 26.911,
|
| 750 |
+
"eval_weighted_f1": 0.8776707970083399,
|
| 751 |
+
"step": 2184
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 12.088275862068965,
|
| 755 |
+
"grad_norm": 0.12463584542274475,
|
| 756 |
+
"learning_rate": 8.797313797313797e-06,
|
| 757 |
+
"loss": 0.0992,
|
| 758 |
+
"step": 2200
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 12.226206896551725,
|
| 762 |
+
"grad_norm": 226.091552734375,
|
| 763 |
+
"learning_rate": 8.644688644688646e-06,
|
| 764 |
+
"loss": 0.0833,
|
| 765 |
+
"step": 2225
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 12.364137931034483,
|
| 769 |
+
"grad_norm": 0.09647581726312637,
|
| 770 |
+
"learning_rate": 8.492063492063492e-06,
|
| 771 |
+
"loss": 0.0939,
|
| 772 |
+
"step": 2250
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 12.50206896551724,
|
| 776 |
+
"grad_norm": 287.18804931640625,
|
| 777 |
+
"learning_rate": 8.33943833943834e-06,
|
| 778 |
+
"loss": 0.1438,
|
| 779 |
+
"step": 2275
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 12.64,
|
| 783 |
+
"grad_norm": 0.10999006032943726,
|
| 784 |
+
"learning_rate": 8.186813186813188e-06,
|
| 785 |
+
"loss": 0.1107,
|
| 786 |
+
"step": 2300
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 12.777931034482759,
|
| 790 |
+
"grad_norm": 0.08639182895421982,
|
| 791 |
+
"learning_rate": 8.034188034188036e-06,
|
| 792 |
+
"loss": 0.0447,
|
| 793 |
+
"step": 2325
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 12.915862068965517,
|
| 797 |
+
"grad_norm": 313.2142639160156,
|
| 798 |
+
"learning_rate": 7.881562881562883e-06,
|
| 799 |
+
"loss": 0.0562,
|
| 800 |
+
"step": 2350
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 13.0,
|
| 804 |
+
"eval_accuracy": 0.8727634194831014,
|
| 805 |
+
"eval_loss": 0.7390493750572205,
|
| 806 |
+
"eval_macro_f1": 0.8709769431984329,
|
| 807 |
+
"eval_runtime": 1.1704,
|
| 808 |
+
"eval_samples_per_second": 429.78,
|
| 809 |
+
"eval_steps_per_second": 27.342,
|
| 810 |
+
"eval_weighted_f1": 0.8711033443052347,
|
| 811 |
+
"step": 2366
|
| 812 |
+
},
|
| 813 |
+
{
|
| 814 |
+
"epoch": 13.049655172413793,
|
| 815 |
+
"grad_norm": 14.408998489379883,
|
| 816 |
+
"learning_rate": 7.72893772893773e-06,
|
| 817 |
+
"loss": 0.1664,
|
| 818 |
+
"step": 2375
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 13.187586206896551,
|
| 822 |
+
"grad_norm": 0.09482051432132721,
|
| 823 |
+
"learning_rate": 7.576312576312577e-06,
|
| 824 |
+
"loss": 0.0756,
|
| 825 |
+
"step": 2400
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 13.32551724137931,
|
| 829 |
+
"grad_norm": 0.0959722101688385,
|
| 830 |
+
"learning_rate": 7.423687423687424e-06,
|
| 831 |
+
"loss": 0.1513,
|
| 832 |
+
"step": 2425
|
| 833 |
+
},
|
| 834 |
+
{
|
| 835 |
+
"epoch": 13.463448275862069,
|
| 836 |
+
"grad_norm": 0.1113782674074173,
|
| 837 |
+
"learning_rate": 7.271062271062271e-06,
|
| 838 |
+
"loss": 0.076,
|
| 839 |
+
"step": 2450
|
| 840 |
+
},
|
| 841 |
+
{
|
| 842 |
+
"epoch": 13.601379310344827,
|
| 843 |
+
"grad_norm": 0.08026798069477081,
|
| 844 |
+
"learning_rate": 7.1184371184371195e-06,
|
| 845 |
+
"loss": 0.0129,
|
| 846 |
+
"step": 2475
|
| 847 |
+
},
|
| 848 |
+
{
|
| 849 |
+
"epoch": 13.739310344827587,
|
| 850 |
+
"grad_norm": 9.052029609680176,
|
| 851 |
+
"learning_rate": 6.965811965811966e-06,
|
| 852 |
+
"loss": 0.07,
|
| 853 |
+
"step": 2500
|
| 854 |
+
},
|
| 855 |
+
{
|
| 856 |
+
"epoch": 13.877241379310345,
|
| 857 |
+
"grad_norm": 0.08703938871622086,
|
| 858 |
+
"learning_rate": 6.813186813186814e-06,
|
| 859 |
+
"loss": 0.0121,
|
| 860 |
+
"step": 2525
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 14.0,
|
| 864 |
+
"eval_accuracy": 0.8687872763419483,
|
| 865 |
+
"eval_loss": 0.7567536234855652,
|
| 866 |
+
"eval_macro_f1": 0.8673273341723067,
|
| 867 |
+
"eval_runtime": 1.1601,
|
| 868 |
+
"eval_samples_per_second": 433.575,
|
| 869 |
+
"eval_steps_per_second": 27.583,
|
| 870 |
+
"eval_weighted_f1": 0.8674994989197677,
|
| 871 |
+
"step": 2548
|
| 872 |
+
}
|
| 873 |
+
],
|
| 874 |
+
"logging_steps": 25,
|
| 875 |
+
"max_steps": 3640,
|
| 876 |
+
"num_input_tokens_seen": 0,
|
| 877 |
+
"num_train_epochs": 20,
|
| 878 |
+
"save_steps": 500,
|
| 879 |
+
"stateful_callbacks": {
|
| 880 |
+
"EarlyStoppingCallback": {
|
| 881 |
+
"args": {
|
| 882 |
+
"early_stopping_patience": 3,
|
| 883 |
+
"early_stopping_threshold": 0.0
|
| 884 |
+
},
|
| 885 |
+
"attributes": {
|
| 886 |
+
"early_stopping_patience_counter": 3
|
| 887 |
+
}
|
| 888 |
+
},
|
| 889 |
+
"TrainerControl": {
|
| 890 |
+
"args": {
|
| 891 |
+
"should_epoch_stop": false,
|
| 892 |
+
"should_evaluate": false,
|
| 893 |
+
"should_log": false,
|
| 894 |
+
"should_save": true,
|
| 895 |
+
"should_training_stop": true
|
| 896 |
+
},
|
| 897 |
+
"attributes": {}
|
| 898 |
+
}
|
| 899 |
+
},
|
| 900 |
+
"total_flos": 0.0,
|
| 901 |
+
"train_batch_size": 8,
|
| 902 |
+
"trial_name": null,
|
| 903 |
+
"trial_params": null
|
| 904 |
+
}
|
ood-vifactcheck-phobert-topic-domain-02-s42-20260814T111300Z-00ca7096/trainer/checkpoint-2548/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-03-s42-20260814T112759Z-00ca7096/trainer/checkpoint-1560/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-topic-domain-03-s42-20260814T112759Z-00ca7096/trainer/checkpoint-1560/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-03-s42-20260814T112759Z-00ca7096/trainer/checkpoint-1560/special_tokens_map.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": "<mask>",
|
| 6 |
+
"pad_token": "<pad>",
|
| 7 |
+
"sep_token": "</s>",
|
| 8 |
+
"unk_token": "<unk>"
|
| 9 |
+
}
|
ood-vifactcheck-phobert-topic-domain-03-s42-20260814T112759Z-00ca7096/trainer/checkpoint-1560/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 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": "</s>",
|
| 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 |
+
"64000": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 256,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "PhobertTokenizer",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-phobert-topic-domain-03-s42-20260814T112759Z-00ca7096/trainer/checkpoint-1560/trainer_state.json
ADDED
|
@@ -0,0 +1,565 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1560,
|
| 3 |
+
"best_metric": 0.855665823868191,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-topic-domain-03-s42-20260814T112759Z-00ca7096/trainer/checkpoint-1560",
|
| 5 |
+
"epoch": 8.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1560,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.12836970474967907,
|
| 14 |
+
"grad_norm": 7.738715171813965,
|
| 15 |
+
"learning_rate": 1.230769230769231e-06,
|
| 16 |
+
"loss": 4.432,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.25673940949935814,
|
| 21 |
+
"grad_norm": 6.359468460083008,
|
| 22 |
+
"learning_rate": 2.512820512820513e-06,
|
| 23 |
+
"loss": 4.4004,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.3851091142490372,
|
| 28 |
+
"grad_norm": 7.449909687042236,
|
| 29 |
+
"learning_rate": 3.794871794871795e-06,
|
| 30 |
+
"loss": 4.4099,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5134788189987163,
|
| 35 |
+
"grad_norm": 10.274007797241211,
|
| 36 |
+
"learning_rate": 5.076923076923077e-06,
|
| 37 |
+
"loss": 4.346,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6418485237483954,
|
| 42 |
+
"grad_norm": 8.241455078125,
|
| 43 |
+
"learning_rate": 6.358974358974359e-06,
|
| 44 |
+
"loss": 4.3432,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.7702182284980744,
|
| 49 |
+
"grad_norm": 8.314175605773926,
|
| 50 |
+
"learning_rate": 7.641025641025641e-06,
|
| 51 |
+
"loss": 4.2779,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.8985879332477535,
|
| 56 |
+
"grad_norm": 10.666259765625,
|
| 57 |
+
"learning_rate": 8.923076923076925e-06,
|
| 58 |
+
"loss": 4.1999,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.5674740484429066,
|
| 64 |
+
"eval_loss": 0.9280925393104553,
|
| 65 |
+
"eval_macro_f1": 0.5527504669799055,
|
| 66 |
+
"eval_runtime": 1.2061,
|
| 67 |
+
"eval_samples_per_second": 479.249,
|
| 68 |
+
"eval_steps_per_second": 30.679,
|
| 69 |
+
"eval_weighted_f1": 0.5548835500832098,
|
| 70 |
+
"step": 195
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0256739409499358,
|
| 74 |
+
"grad_norm": 20.324617385864258,
|
| 75 |
+
"learning_rate": 1.0205128205128205e-05,
|
| 76 |
+
"loss": 3.8373,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.1540436456996148,
|
| 81 |
+
"grad_norm": 30.23924446105957,
|
| 82 |
+
"learning_rate": 1.1487179487179487e-05,
|
| 83 |
+
"loss": 3.6217,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.282413350449294,
|
| 88 |
+
"grad_norm": 33.74998474121094,
|
| 89 |
+
"learning_rate": 1.2769230769230769e-05,
|
| 90 |
+
"loss": 3.2352,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.410783055198973,
|
| 95 |
+
"grad_norm": 44.562416076660156,
|
| 96 |
+
"learning_rate": 1.405128205128205e-05,
|
| 97 |
+
"loss": 2.8012,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.539152759948652,
|
| 102 |
+
"grad_norm": 37.261165618896484,
|
| 103 |
+
"learning_rate": 1.5333333333333334e-05,
|
| 104 |
+
"loss": 2.3632,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.6675224646983313,
|
| 109 |
+
"grad_norm": 18.699153900146484,
|
| 110 |
+
"learning_rate": 1.6615384615384618e-05,
|
| 111 |
+
"loss": 2.3893,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.7958921694480101,
|
| 116 |
+
"grad_norm": 42.164276123046875,
|
| 117 |
+
"learning_rate": 1.78974358974359e-05,
|
| 118 |
+
"loss": 2.1621,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 1.9242618741976893,
|
| 123 |
+
"grad_norm": 34.17839813232422,
|
| 124 |
+
"learning_rate": 1.9179487179487182e-05,
|
| 125 |
+
"loss": 2.2728,
|
| 126 |
+
"step": 375
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"epoch": 2.0,
|
| 130 |
+
"eval_accuracy": 0.8373702422145328,
|
| 131 |
+
"eval_loss": 0.4474438726902008,
|
| 132 |
+
"eval_macro_f1": 0.8368192149663433,
|
| 133 |
+
"eval_runtime": 1.2377,
|
| 134 |
+
"eval_samples_per_second": 466.99,
|
| 135 |
+
"eval_steps_per_second": 29.894,
|
| 136 |
+
"eval_weighted_f1": 0.8368650047143401,
|
| 137 |
+
"step": 390
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.0513478818998716,
|
| 141 |
+
"grad_norm": 28.649656295776367,
|
| 142 |
+
"learning_rate": 1.994871794871795e-05,
|
| 143 |
+
"loss": 2.0455,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.179717586649551,
|
| 148 |
+
"grad_norm": 47.43769454956055,
|
| 149 |
+
"learning_rate": 1.9806267806267806e-05,
|
| 150 |
+
"loss": 1.7031,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.3080872913992296,
|
| 155 |
+
"grad_norm": 27.74767303466797,
|
| 156 |
+
"learning_rate": 1.9663817663817667e-05,
|
| 157 |
+
"loss": 1.8297,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.436456996148909,
|
| 162 |
+
"grad_norm": 65.58319854736328,
|
| 163 |
+
"learning_rate": 1.952136752136752e-05,
|
| 164 |
+
"loss": 1.5858,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.564826700898588,
|
| 169 |
+
"grad_norm": 39.27559280395508,
|
| 170 |
+
"learning_rate": 1.9378917378917382e-05,
|
| 171 |
+
"loss": 1.3903,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.693196405648267,
|
| 176 |
+
"grad_norm": 40.734222412109375,
|
| 177 |
+
"learning_rate": 1.923646723646724e-05,
|
| 178 |
+
"loss": 1.6601,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 2.821566110397946,
|
| 183 |
+
"grad_norm": 20.323850631713867,
|
| 184 |
+
"learning_rate": 1.9094017094017097e-05,
|
| 185 |
+
"loss": 1.5183,
|
| 186 |
+
"step": 550
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 2.949935815147625,
|
| 190 |
+
"grad_norm": 29.32554054260254,
|
| 191 |
+
"learning_rate": 1.8951566951566954e-05,
|
| 192 |
+
"loss": 1.4179,
|
| 193 |
+
"step": 575
|
| 194 |
+
},
|
| 195 |
+
{
|
| 196 |
+
"epoch": 3.0,
|
| 197 |
+
"eval_accuracy": 0.8442906574394463,
|
| 198 |
+
"eval_loss": 0.4606327414512634,
|
| 199 |
+
"eval_macro_f1": 0.8443810927459473,
|
| 200 |
+
"eval_runtime": 1.2066,
|
| 201 |
+
"eval_samples_per_second": 479.019,
|
| 202 |
+
"eval_steps_per_second": 30.664,
|
| 203 |
+
"eval_weighted_f1": 0.8442724333779262,
|
| 204 |
+
"step": 585
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.0770218228498076,
|
| 208 |
+
"grad_norm": 28.768789291381836,
|
| 209 |
+
"learning_rate": 1.880911680911681e-05,
|
| 210 |
+
"loss": 1.1603,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.2053915275994864,
|
| 215 |
+
"grad_norm": 49.38003921508789,
|
| 216 |
+
"learning_rate": 1.866666666666667e-05,
|
| 217 |
+
"loss": 0.858,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.3337612323491657,
|
| 222 |
+
"grad_norm": 78.51256561279297,
|
| 223 |
+
"learning_rate": 1.8524216524216526e-05,
|
| 224 |
+
"loss": 1.0086,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.462130937098845,
|
| 229 |
+
"grad_norm": 58.57139205932617,
|
| 230 |
+
"learning_rate": 1.8381766381766383e-05,
|
| 231 |
+
"loss": 1.0527,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.5905006418485237,
|
| 236 |
+
"grad_norm": 79.29813385009766,
|
| 237 |
+
"learning_rate": 1.823931623931624e-05,
|
| 238 |
+
"loss": 1.0231,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.718870346598203,
|
| 243 |
+
"grad_norm": 11.536803245544434,
|
| 244 |
+
"learning_rate": 1.8096866096866098e-05,
|
| 245 |
+
"loss": 1.0351,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 3.8472400513478817,
|
| 250 |
+
"grad_norm": 48.34772491455078,
|
| 251 |
+
"learning_rate": 1.7954415954415955e-05,
|
| 252 |
+
"loss": 1.0577,
|
| 253 |
+
"step": 750
|
| 254 |
+
},
|
| 255 |
+
{
|
| 256 |
+
"epoch": 3.975609756097561,
|
| 257 |
+
"grad_norm": 124.57843017578125,
|
| 258 |
+
"learning_rate": 1.7811965811965813e-05,
|
| 259 |
+
"loss": 1.058,
|
| 260 |
+
"step": 775
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 4.0,
|
| 264 |
+
"eval_accuracy": 0.8494809688581315,
|
| 265 |
+
"eval_loss": 0.4874649941921234,
|
| 266 |
+
"eval_macro_f1": 0.8471943822953413,
|
| 267 |
+
"eval_runtime": 1.4863,
|
| 268 |
+
"eval_samples_per_second": 388.892,
|
| 269 |
+
"eval_steps_per_second": 24.894,
|
| 270 |
+
"eval_weighted_f1": 0.8474832962637345,
|
| 271 |
+
"step": 780
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.102695763799743,
|
| 275 |
+
"grad_norm": 74.82074737548828,
|
| 276 |
+
"learning_rate": 1.766951566951567e-05,
|
| 277 |
+
"loss": 0.613,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.2310654685494224,
|
| 282 |
+
"grad_norm": 96.63713073730469,
|
| 283 |
+
"learning_rate": 1.7527065527065528e-05,
|
| 284 |
+
"loss": 0.5871,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.359435173299102,
|
| 289 |
+
"grad_norm": 80.54320526123047,
|
| 290 |
+
"learning_rate": 1.7384615384615385e-05,
|
| 291 |
+
"loss": 0.663,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.487804878048781,
|
| 296 |
+
"grad_norm": 21.244792938232422,
|
| 297 |
+
"learning_rate": 1.7242165242165242e-05,
|
| 298 |
+
"loss": 0.8173,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.616174582798459,
|
| 303 |
+
"grad_norm": 101.47552490234375,
|
| 304 |
+
"learning_rate": 1.7099715099715103e-05,
|
| 305 |
+
"loss": 0.6848,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 4.7445442875481385,
|
| 310 |
+
"grad_norm": 88.3195571899414,
|
| 311 |
+
"learning_rate": 1.6957264957264957e-05,
|
| 312 |
+
"loss": 0.7748,
|
| 313 |
+
"step": 925
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 4.872913992297818,
|
| 317 |
+
"grad_norm": 83.55799865722656,
|
| 318 |
+
"learning_rate": 1.6814814814814818e-05,
|
| 319 |
+
"loss": 0.8949,
|
| 320 |
+
"step": 950
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 5.0,
|
| 324 |
+
"grad_norm": 26.083642959594727,
|
| 325 |
+
"learning_rate": 1.6672364672364675e-05,
|
| 326 |
+
"loss": 0.6674,
|
| 327 |
+
"step": 975
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 5.0,
|
| 331 |
+
"eval_accuracy": 0.8339100346020761,
|
| 332 |
+
"eval_loss": 0.5811436176300049,
|
| 333 |
+
"eval_macro_f1": 0.8331711480743936,
|
| 334 |
+
"eval_runtime": 1.2472,
|
| 335 |
+
"eval_samples_per_second": 463.447,
|
| 336 |
+
"eval_steps_per_second": 29.667,
|
| 337 |
+
"eval_weighted_f1": 0.8325282788815155,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.128369704749679,
|
| 342 |
+
"grad_norm": 56.67775344848633,
|
| 343 |
+
"learning_rate": 1.652991452991453e-05,
|
| 344 |
+
"loss": 0.5335,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.2567394094993585,
|
| 349 |
+
"grad_norm": 10.136966705322266,
|
| 350 |
+
"learning_rate": 1.638746438746439e-05,
|
| 351 |
+
"loss": 0.4871,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.385109114249037,
|
| 356 |
+
"grad_norm": 62.851436614990234,
|
| 357 |
+
"learning_rate": 1.6245014245014247e-05,
|
| 358 |
+
"loss": 0.5509,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.513478818998716,
|
| 363 |
+
"grad_norm": 14.664275169372559,
|
| 364 |
+
"learning_rate": 1.6102564102564105e-05,
|
| 365 |
+
"loss": 0.4572,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 5.641848523748395,
|
| 370 |
+
"grad_norm": 41.357601165771484,
|
| 371 |
+
"learning_rate": 1.5960113960113962e-05,
|
| 372 |
+
"loss": 0.3681,
|
| 373 |
+
"step": 1100
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 5.7702182284980745,
|
| 377 |
+
"grad_norm": 0.611042320728302,
|
| 378 |
+
"learning_rate": 1.581766381766382e-05,
|
| 379 |
+
"loss": 0.5537,
|
| 380 |
+
"step": 1125
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 5.898587933247754,
|
| 384 |
+
"grad_norm": 34.371280670166016,
|
| 385 |
+
"learning_rate": 1.5675213675213677e-05,
|
| 386 |
+
"loss": 0.6049,
|
| 387 |
+
"step": 1150
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 6.0,
|
| 391 |
+
"eval_accuracy": 0.8512110726643599,
|
| 392 |
+
"eval_loss": 0.6086398363113403,
|
| 393 |
+
"eval_macro_f1": 0.851572829229505,
|
| 394 |
+
"eval_runtime": 1.2503,
|
| 395 |
+
"eval_samples_per_second": 462.298,
|
| 396 |
+
"eval_steps_per_second": 29.593,
|
| 397 |
+
"eval_weighted_f1": 0.8517077434004167,
|
| 398 |
+
"step": 1170
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.025673940949936,
|
| 402 |
+
"grad_norm": 49.769004821777344,
|
| 403 |
+
"learning_rate": 1.5532763532763534e-05,
|
| 404 |
+
"loss": 0.5111,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.154043645699615,
|
| 409 |
+
"grad_norm": 0.7132329940795898,
|
| 410 |
+
"learning_rate": 1.539031339031339e-05,
|
| 411 |
+
"loss": 0.4221,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.282413350449294,
|
| 416 |
+
"grad_norm": 95.32638549804688,
|
| 417 |
+
"learning_rate": 1.5247863247863249e-05,
|
| 418 |
+
"loss": 0.2003,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.410783055198973,
|
| 423 |
+
"grad_norm": 0.970021665096283,
|
| 424 |
+
"learning_rate": 1.5105413105413106e-05,
|
| 425 |
+
"loss": 0.3006,
|
| 426 |
+
"step": 1250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 6.539152759948652,
|
| 430 |
+
"grad_norm": 45.701297760009766,
|
| 431 |
+
"learning_rate": 1.4962962962962964e-05,
|
| 432 |
+
"loss": 0.3727,
|
| 433 |
+
"step": 1275
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 6.667522464698331,
|
| 437 |
+
"grad_norm": 104.99532318115234,
|
| 438 |
+
"learning_rate": 1.4820512820512823e-05,
|
| 439 |
+
"loss": 0.2674,
|
| 440 |
+
"step": 1300
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 6.7958921694480106,
|
| 444 |
+
"grad_norm": 53.279571533203125,
|
| 445 |
+
"learning_rate": 1.4678062678062678e-05,
|
| 446 |
+
"loss": 0.6969,
|
| 447 |
+
"step": 1325
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 6.92426187419769,
|
| 451 |
+
"grad_norm": 1.6106728315353394,
|
| 452 |
+
"learning_rate": 1.4535612535612537e-05,
|
| 453 |
+
"loss": 0.4367,
|
| 454 |
+
"step": 1350
|
| 455 |
+
},
|
| 456 |
+
{
|
| 457 |
+
"epoch": 7.0,
|
| 458 |
+
"eval_accuracy": 0.8477508650519031,
|
| 459 |
+
"eval_loss": 0.6865471005439758,
|
| 460 |
+
"eval_macro_f1": 0.8474582690605778,
|
| 461 |
+
"eval_runtime": 1.283,
|
| 462 |
+
"eval_samples_per_second": 450.512,
|
| 463 |
+
"eval_steps_per_second": 28.839,
|
| 464 |
+
"eval_weighted_f1": 0.8475714330649988,
|
| 465 |
+
"step": 1365
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 7.051347881899872,
|
| 469 |
+
"grad_norm": 1.957703948020935,
|
| 470 |
+
"learning_rate": 1.4393162393162395e-05,
|
| 471 |
+
"loss": 0.3854,
|
| 472 |
+
"step": 1375
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 7.17971758664955,
|
| 476 |
+
"grad_norm": 54.61171340942383,
|
| 477 |
+
"learning_rate": 1.425071225071225e-05,
|
| 478 |
+
"loss": 0.1821,
|
| 479 |
+
"step": 1400
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 7.30808729139923,
|
| 483 |
+
"grad_norm": 0.3422224521636963,
|
| 484 |
+
"learning_rate": 1.410826210826211e-05,
|
| 485 |
+
"loss": 0.4178,
|
| 486 |
+
"step": 1425
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 7.436456996148909,
|
| 490 |
+
"grad_norm": 6.124192714691162,
|
| 491 |
+
"learning_rate": 1.3965811965811967e-05,
|
| 492 |
+
"loss": 0.1619,
|
| 493 |
+
"step": 1450
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 7.564826700898588,
|
| 497 |
+
"grad_norm": 0.22317878901958466,
|
| 498 |
+
"learning_rate": 1.3823361823361824e-05,
|
| 499 |
+
"loss": 0.1837,
|
| 500 |
+
"step": 1475
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 7.693196405648267,
|
| 504 |
+
"grad_norm": 146.4649658203125,
|
| 505 |
+
"learning_rate": 1.3680911680911682e-05,
|
| 506 |
+
"loss": 0.4807,
|
| 507 |
+
"step": 1500
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 7.821566110397946,
|
| 511 |
+
"grad_norm": 136.75704956054688,
|
| 512 |
+
"learning_rate": 1.353846153846154e-05,
|
| 513 |
+
"loss": 0.3685,
|
| 514 |
+
"step": 1525
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 7.949935815147625,
|
| 518 |
+
"grad_norm": 0.3173060417175293,
|
| 519 |
+
"learning_rate": 1.3396011396011396e-05,
|
| 520 |
+
"loss": 0.3258,
|
| 521 |
+
"step": 1550
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 8.0,
|
| 525 |
+
"eval_accuracy": 0.856401384083045,
|
| 526 |
+
"eval_loss": 0.7225363850593567,
|
| 527 |
+
"eval_macro_f1": 0.855665823868191,
|
| 528 |
+
"eval_runtime": 1.1883,
|
| 529 |
+
"eval_samples_per_second": 486.419,
|
| 530 |
+
"eval_steps_per_second": 31.138,
|
| 531 |
+
"eval_weighted_f1": 0.855944421347496,
|
| 532 |
+
"step": 1560
|
| 533 |
+
}
|
| 534 |
+
],
|
| 535 |
+
"logging_steps": 25,
|
| 536 |
+
"max_steps": 3900,
|
| 537 |
+
"num_input_tokens_seen": 0,
|
| 538 |
+
"num_train_epochs": 20,
|
| 539 |
+
"save_steps": 500,
|
| 540 |
+
"stateful_callbacks": {
|
| 541 |
+
"EarlyStoppingCallback": {
|
| 542 |
+
"args": {
|
| 543 |
+
"early_stopping_patience": 3,
|
| 544 |
+
"early_stopping_threshold": 0.0
|
| 545 |
+
},
|
| 546 |
+
"attributes": {
|
| 547 |
+
"early_stopping_patience_counter": 0
|
| 548 |
+
}
|
| 549 |
+
},
|
| 550 |
+
"TrainerControl": {
|
| 551 |
+
"args": {
|
| 552 |
+
"should_epoch_stop": false,
|
| 553 |
+
"should_evaluate": false,
|
| 554 |
+
"should_log": false,
|
| 555 |
+
"should_save": true,
|
| 556 |
+
"should_training_stop": false
|
| 557 |
+
},
|
| 558 |
+
"attributes": {}
|
| 559 |
+
}
|
| 560 |
+
},
|
| 561 |
+
"total_flos": 0.0,
|
| 562 |
+
"train_batch_size": 8,
|
| 563 |
+
"trial_name": null,
|
| 564 |
+
"trial_params": null
|
| 565 |
+
}
|
ood-vifactcheck-phobert-topic-domain-03-s42-20260814T112759Z-00ca7096/trainer/checkpoint-1560/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-03-s42-20260814T112759Z-00ca7096/trainer/checkpoint-2145/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-topic-domain-03-s42-20260814T112759Z-00ca7096/trainer/checkpoint-2145/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-03-s42-20260814T112759Z-00ca7096/trainer/checkpoint-2145/special_tokens_map.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": "<mask>",
|
| 6 |
+
"pad_token": "<pad>",
|
| 7 |
+
"sep_token": "</s>",
|
| 8 |
+
"unk_token": "<unk>"
|
| 9 |
+
}
|
ood-vifactcheck-phobert-topic-domain-03-s42-20260814T112759Z-00ca7096/trainer/checkpoint-2145/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 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": "</s>",
|
| 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 |
+
"64000": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 256,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "PhobertTokenizer",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-phobert-topic-domain-03-s42-20260814T112759Z-00ca7096/trainer/checkpoint-2145/trainer_state.json
ADDED
|
@@ -0,0 +1,759 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1560,
|
| 3 |
+
"best_metric": 0.855665823868191,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-topic-domain-03-s42-20260814T112759Z-00ca7096/trainer/checkpoint-1560",
|
| 5 |
+
"epoch": 11.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 2145,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.12836970474967907,
|
| 14 |
+
"grad_norm": 7.738715171813965,
|
| 15 |
+
"learning_rate": 1.230769230769231e-06,
|
| 16 |
+
"loss": 4.432,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.25673940949935814,
|
| 21 |
+
"grad_norm": 6.359468460083008,
|
| 22 |
+
"learning_rate": 2.512820512820513e-06,
|
| 23 |
+
"loss": 4.4004,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.3851091142490372,
|
| 28 |
+
"grad_norm": 7.449909687042236,
|
| 29 |
+
"learning_rate": 3.794871794871795e-06,
|
| 30 |
+
"loss": 4.4099,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5134788189987163,
|
| 35 |
+
"grad_norm": 10.274007797241211,
|
| 36 |
+
"learning_rate": 5.076923076923077e-06,
|
| 37 |
+
"loss": 4.346,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6418485237483954,
|
| 42 |
+
"grad_norm": 8.241455078125,
|
| 43 |
+
"learning_rate": 6.358974358974359e-06,
|
| 44 |
+
"loss": 4.3432,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.7702182284980744,
|
| 49 |
+
"grad_norm": 8.314175605773926,
|
| 50 |
+
"learning_rate": 7.641025641025641e-06,
|
| 51 |
+
"loss": 4.2779,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.8985879332477535,
|
| 56 |
+
"grad_norm": 10.666259765625,
|
| 57 |
+
"learning_rate": 8.923076923076925e-06,
|
| 58 |
+
"loss": 4.1999,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.5674740484429066,
|
| 64 |
+
"eval_loss": 0.9280925393104553,
|
| 65 |
+
"eval_macro_f1": 0.5527504669799055,
|
| 66 |
+
"eval_runtime": 1.2061,
|
| 67 |
+
"eval_samples_per_second": 479.249,
|
| 68 |
+
"eval_steps_per_second": 30.679,
|
| 69 |
+
"eval_weighted_f1": 0.5548835500832098,
|
| 70 |
+
"step": 195
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0256739409499358,
|
| 74 |
+
"grad_norm": 20.324617385864258,
|
| 75 |
+
"learning_rate": 1.0205128205128205e-05,
|
| 76 |
+
"loss": 3.8373,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.1540436456996148,
|
| 81 |
+
"grad_norm": 30.23924446105957,
|
| 82 |
+
"learning_rate": 1.1487179487179487e-05,
|
| 83 |
+
"loss": 3.6217,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.282413350449294,
|
| 88 |
+
"grad_norm": 33.74998474121094,
|
| 89 |
+
"learning_rate": 1.2769230769230769e-05,
|
| 90 |
+
"loss": 3.2352,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.410783055198973,
|
| 95 |
+
"grad_norm": 44.562416076660156,
|
| 96 |
+
"learning_rate": 1.405128205128205e-05,
|
| 97 |
+
"loss": 2.8012,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.539152759948652,
|
| 102 |
+
"grad_norm": 37.261165618896484,
|
| 103 |
+
"learning_rate": 1.5333333333333334e-05,
|
| 104 |
+
"loss": 2.3632,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.6675224646983313,
|
| 109 |
+
"grad_norm": 18.699153900146484,
|
| 110 |
+
"learning_rate": 1.6615384615384618e-05,
|
| 111 |
+
"loss": 2.3893,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.7958921694480101,
|
| 116 |
+
"grad_norm": 42.164276123046875,
|
| 117 |
+
"learning_rate": 1.78974358974359e-05,
|
| 118 |
+
"loss": 2.1621,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 1.9242618741976893,
|
| 123 |
+
"grad_norm": 34.17839813232422,
|
| 124 |
+
"learning_rate": 1.9179487179487182e-05,
|
| 125 |
+
"loss": 2.2728,
|
| 126 |
+
"step": 375
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"epoch": 2.0,
|
| 130 |
+
"eval_accuracy": 0.8373702422145328,
|
| 131 |
+
"eval_loss": 0.4474438726902008,
|
| 132 |
+
"eval_macro_f1": 0.8368192149663433,
|
| 133 |
+
"eval_runtime": 1.2377,
|
| 134 |
+
"eval_samples_per_second": 466.99,
|
| 135 |
+
"eval_steps_per_second": 29.894,
|
| 136 |
+
"eval_weighted_f1": 0.8368650047143401,
|
| 137 |
+
"step": 390
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.0513478818998716,
|
| 141 |
+
"grad_norm": 28.649656295776367,
|
| 142 |
+
"learning_rate": 1.994871794871795e-05,
|
| 143 |
+
"loss": 2.0455,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.179717586649551,
|
| 148 |
+
"grad_norm": 47.43769454956055,
|
| 149 |
+
"learning_rate": 1.9806267806267806e-05,
|
| 150 |
+
"loss": 1.7031,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.3080872913992296,
|
| 155 |
+
"grad_norm": 27.74767303466797,
|
| 156 |
+
"learning_rate": 1.9663817663817667e-05,
|
| 157 |
+
"loss": 1.8297,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.436456996148909,
|
| 162 |
+
"grad_norm": 65.58319854736328,
|
| 163 |
+
"learning_rate": 1.952136752136752e-05,
|
| 164 |
+
"loss": 1.5858,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.564826700898588,
|
| 169 |
+
"grad_norm": 39.27559280395508,
|
| 170 |
+
"learning_rate": 1.9378917378917382e-05,
|
| 171 |
+
"loss": 1.3903,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.693196405648267,
|
| 176 |
+
"grad_norm": 40.734222412109375,
|
| 177 |
+
"learning_rate": 1.923646723646724e-05,
|
| 178 |
+
"loss": 1.6601,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 2.821566110397946,
|
| 183 |
+
"grad_norm": 20.323850631713867,
|
| 184 |
+
"learning_rate": 1.9094017094017097e-05,
|
| 185 |
+
"loss": 1.5183,
|
| 186 |
+
"step": 550
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 2.949935815147625,
|
| 190 |
+
"grad_norm": 29.32554054260254,
|
| 191 |
+
"learning_rate": 1.8951566951566954e-05,
|
| 192 |
+
"loss": 1.4179,
|
| 193 |
+
"step": 575
|
| 194 |
+
},
|
| 195 |
+
{
|
| 196 |
+
"epoch": 3.0,
|
| 197 |
+
"eval_accuracy": 0.8442906574394463,
|
| 198 |
+
"eval_loss": 0.4606327414512634,
|
| 199 |
+
"eval_macro_f1": 0.8443810927459473,
|
| 200 |
+
"eval_runtime": 1.2066,
|
| 201 |
+
"eval_samples_per_second": 479.019,
|
| 202 |
+
"eval_steps_per_second": 30.664,
|
| 203 |
+
"eval_weighted_f1": 0.8442724333779262,
|
| 204 |
+
"step": 585
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.0770218228498076,
|
| 208 |
+
"grad_norm": 28.768789291381836,
|
| 209 |
+
"learning_rate": 1.880911680911681e-05,
|
| 210 |
+
"loss": 1.1603,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.2053915275994864,
|
| 215 |
+
"grad_norm": 49.38003921508789,
|
| 216 |
+
"learning_rate": 1.866666666666667e-05,
|
| 217 |
+
"loss": 0.858,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.3337612323491657,
|
| 222 |
+
"grad_norm": 78.51256561279297,
|
| 223 |
+
"learning_rate": 1.8524216524216526e-05,
|
| 224 |
+
"loss": 1.0086,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.462130937098845,
|
| 229 |
+
"grad_norm": 58.57139205932617,
|
| 230 |
+
"learning_rate": 1.8381766381766383e-05,
|
| 231 |
+
"loss": 1.0527,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.5905006418485237,
|
| 236 |
+
"grad_norm": 79.29813385009766,
|
| 237 |
+
"learning_rate": 1.823931623931624e-05,
|
| 238 |
+
"loss": 1.0231,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.718870346598203,
|
| 243 |
+
"grad_norm": 11.536803245544434,
|
| 244 |
+
"learning_rate": 1.8096866096866098e-05,
|
| 245 |
+
"loss": 1.0351,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 3.8472400513478817,
|
| 250 |
+
"grad_norm": 48.34772491455078,
|
| 251 |
+
"learning_rate": 1.7954415954415955e-05,
|
| 252 |
+
"loss": 1.0577,
|
| 253 |
+
"step": 750
|
| 254 |
+
},
|
| 255 |
+
{
|
| 256 |
+
"epoch": 3.975609756097561,
|
| 257 |
+
"grad_norm": 124.57843017578125,
|
| 258 |
+
"learning_rate": 1.7811965811965813e-05,
|
| 259 |
+
"loss": 1.058,
|
| 260 |
+
"step": 775
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 4.0,
|
| 264 |
+
"eval_accuracy": 0.8494809688581315,
|
| 265 |
+
"eval_loss": 0.4874649941921234,
|
| 266 |
+
"eval_macro_f1": 0.8471943822953413,
|
| 267 |
+
"eval_runtime": 1.4863,
|
| 268 |
+
"eval_samples_per_second": 388.892,
|
| 269 |
+
"eval_steps_per_second": 24.894,
|
| 270 |
+
"eval_weighted_f1": 0.8474832962637345,
|
| 271 |
+
"step": 780
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.102695763799743,
|
| 275 |
+
"grad_norm": 74.82074737548828,
|
| 276 |
+
"learning_rate": 1.766951566951567e-05,
|
| 277 |
+
"loss": 0.613,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.2310654685494224,
|
| 282 |
+
"grad_norm": 96.63713073730469,
|
| 283 |
+
"learning_rate": 1.7527065527065528e-05,
|
| 284 |
+
"loss": 0.5871,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.359435173299102,
|
| 289 |
+
"grad_norm": 80.54320526123047,
|
| 290 |
+
"learning_rate": 1.7384615384615385e-05,
|
| 291 |
+
"loss": 0.663,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.487804878048781,
|
| 296 |
+
"grad_norm": 21.244792938232422,
|
| 297 |
+
"learning_rate": 1.7242165242165242e-05,
|
| 298 |
+
"loss": 0.8173,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.616174582798459,
|
| 303 |
+
"grad_norm": 101.47552490234375,
|
| 304 |
+
"learning_rate": 1.7099715099715103e-05,
|
| 305 |
+
"loss": 0.6848,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 4.7445442875481385,
|
| 310 |
+
"grad_norm": 88.3195571899414,
|
| 311 |
+
"learning_rate": 1.6957264957264957e-05,
|
| 312 |
+
"loss": 0.7748,
|
| 313 |
+
"step": 925
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 4.872913992297818,
|
| 317 |
+
"grad_norm": 83.55799865722656,
|
| 318 |
+
"learning_rate": 1.6814814814814818e-05,
|
| 319 |
+
"loss": 0.8949,
|
| 320 |
+
"step": 950
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 5.0,
|
| 324 |
+
"grad_norm": 26.083642959594727,
|
| 325 |
+
"learning_rate": 1.6672364672364675e-05,
|
| 326 |
+
"loss": 0.6674,
|
| 327 |
+
"step": 975
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 5.0,
|
| 331 |
+
"eval_accuracy": 0.8339100346020761,
|
| 332 |
+
"eval_loss": 0.5811436176300049,
|
| 333 |
+
"eval_macro_f1": 0.8331711480743936,
|
| 334 |
+
"eval_runtime": 1.2472,
|
| 335 |
+
"eval_samples_per_second": 463.447,
|
| 336 |
+
"eval_steps_per_second": 29.667,
|
| 337 |
+
"eval_weighted_f1": 0.8325282788815155,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.128369704749679,
|
| 342 |
+
"grad_norm": 56.67775344848633,
|
| 343 |
+
"learning_rate": 1.652991452991453e-05,
|
| 344 |
+
"loss": 0.5335,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.2567394094993585,
|
| 349 |
+
"grad_norm": 10.136966705322266,
|
| 350 |
+
"learning_rate": 1.638746438746439e-05,
|
| 351 |
+
"loss": 0.4871,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.385109114249037,
|
| 356 |
+
"grad_norm": 62.851436614990234,
|
| 357 |
+
"learning_rate": 1.6245014245014247e-05,
|
| 358 |
+
"loss": 0.5509,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.513478818998716,
|
| 363 |
+
"grad_norm": 14.664275169372559,
|
| 364 |
+
"learning_rate": 1.6102564102564105e-05,
|
| 365 |
+
"loss": 0.4572,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 5.641848523748395,
|
| 370 |
+
"grad_norm": 41.357601165771484,
|
| 371 |
+
"learning_rate": 1.5960113960113962e-05,
|
| 372 |
+
"loss": 0.3681,
|
| 373 |
+
"step": 1100
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 5.7702182284980745,
|
| 377 |
+
"grad_norm": 0.611042320728302,
|
| 378 |
+
"learning_rate": 1.581766381766382e-05,
|
| 379 |
+
"loss": 0.5537,
|
| 380 |
+
"step": 1125
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 5.898587933247754,
|
| 384 |
+
"grad_norm": 34.371280670166016,
|
| 385 |
+
"learning_rate": 1.5675213675213677e-05,
|
| 386 |
+
"loss": 0.6049,
|
| 387 |
+
"step": 1150
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 6.0,
|
| 391 |
+
"eval_accuracy": 0.8512110726643599,
|
| 392 |
+
"eval_loss": 0.6086398363113403,
|
| 393 |
+
"eval_macro_f1": 0.851572829229505,
|
| 394 |
+
"eval_runtime": 1.2503,
|
| 395 |
+
"eval_samples_per_second": 462.298,
|
| 396 |
+
"eval_steps_per_second": 29.593,
|
| 397 |
+
"eval_weighted_f1": 0.8517077434004167,
|
| 398 |
+
"step": 1170
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.025673940949936,
|
| 402 |
+
"grad_norm": 49.769004821777344,
|
| 403 |
+
"learning_rate": 1.5532763532763534e-05,
|
| 404 |
+
"loss": 0.5111,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.154043645699615,
|
| 409 |
+
"grad_norm": 0.7132329940795898,
|
| 410 |
+
"learning_rate": 1.539031339031339e-05,
|
| 411 |
+
"loss": 0.4221,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.282413350449294,
|
| 416 |
+
"grad_norm": 95.32638549804688,
|
| 417 |
+
"learning_rate": 1.5247863247863249e-05,
|
| 418 |
+
"loss": 0.2003,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.410783055198973,
|
| 423 |
+
"grad_norm": 0.970021665096283,
|
| 424 |
+
"learning_rate": 1.5105413105413106e-05,
|
| 425 |
+
"loss": 0.3006,
|
| 426 |
+
"step": 1250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 6.539152759948652,
|
| 430 |
+
"grad_norm": 45.701297760009766,
|
| 431 |
+
"learning_rate": 1.4962962962962964e-05,
|
| 432 |
+
"loss": 0.3727,
|
| 433 |
+
"step": 1275
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 6.667522464698331,
|
| 437 |
+
"grad_norm": 104.99532318115234,
|
| 438 |
+
"learning_rate": 1.4820512820512823e-05,
|
| 439 |
+
"loss": 0.2674,
|
| 440 |
+
"step": 1300
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 6.7958921694480106,
|
| 444 |
+
"grad_norm": 53.279571533203125,
|
| 445 |
+
"learning_rate": 1.4678062678062678e-05,
|
| 446 |
+
"loss": 0.6969,
|
| 447 |
+
"step": 1325
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 6.92426187419769,
|
| 451 |
+
"grad_norm": 1.6106728315353394,
|
| 452 |
+
"learning_rate": 1.4535612535612537e-05,
|
| 453 |
+
"loss": 0.4367,
|
| 454 |
+
"step": 1350
|
| 455 |
+
},
|
| 456 |
+
{
|
| 457 |
+
"epoch": 7.0,
|
| 458 |
+
"eval_accuracy": 0.8477508650519031,
|
| 459 |
+
"eval_loss": 0.6865471005439758,
|
| 460 |
+
"eval_macro_f1": 0.8474582690605778,
|
| 461 |
+
"eval_runtime": 1.283,
|
| 462 |
+
"eval_samples_per_second": 450.512,
|
| 463 |
+
"eval_steps_per_second": 28.839,
|
| 464 |
+
"eval_weighted_f1": 0.8475714330649988,
|
| 465 |
+
"step": 1365
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 7.051347881899872,
|
| 469 |
+
"grad_norm": 1.957703948020935,
|
| 470 |
+
"learning_rate": 1.4393162393162395e-05,
|
| 471 |
+
"loss": 0.3854,
|
| 472 |
+
"step": 1375
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 7.17971758664955,
|
| 476 |
+
"grad_norm": 54.61171340942383,
|
| 477 |
+
"learning_rate": 1.425071225071225e-05,
|
| 478 |
+
"loss": 0.1821,
|
| 479 |
+
"step": 1400
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 7.30808729139923,
|
| 483 |
+
"grad_norm": 0.3422224521636963,
|
| 484 |
+
"learning_rate": 1.410826210826211e-05,
|
| 485 |
+
"loss": 0.4178,
|
| 486 |
+
"step": 1425
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 7.436456996148909,
|
| 490 |
+
"grad_norm": 6.124192714691162,
|
| 491 |
+
"learning_rate": 1.3965811965811967e-05,
|
| 492 |
+
"loss": 0.1619,
|
| 493 |
+
"step": 1450
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 7.564826700898588,
|
| 497 |
+
"grad_norm": 0.22317878901958466,
|
| 498 |
+
"learning_rate": 1.3823361823361824e-05,
|
| 499 |
+
"loss": 0.1837,
|
| 500 |
+
"step": 1475
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 7.693196405648267,
|
| 504 |
+
"grad_norm": 146.4649658203125,
|
| 505 |
+
"learning_rate": 1.3680911680911682e-05,
|
| 506 |
+
"loss": 0.4807,
|
| 507 |
+
"step": 1500
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 7.821566110397946,
|
| 511 |
+
"grad_norm": 136.75704956054688,
|
| 512 |
+
"learning_rate": 1.353846153846154e-05,
|
| 513 |
+
"loss": 0.3685,
|
| 514 |
+
"step": 1525
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 7.949935815147625,
|
| 518 |
+
"grad_norm": 0.3173060417175293,
|
| 519 |
+
"learning_rate": 1.3396011396011396e-05,
|
| 520 |
+
"loss": 0.3258,
|
| 521 |
+
"step": 1550
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 8.0,
|
| 525 |
+
"eval_accuracy": 0.856401384083045,
|
| 526 |
+
"eval_loss": 0.7225363850593567,
|
| 527 |
+
"eval_macro_f1": 0.855665823868191,
|
| 528 |
+
"eval_runtime": 1.1883,
|
| 529 |
+
"eval_samples_per_second": 486.419,
|
| 530 |
+
"eval_steps_per_second": 31.138,
|
| 531 |
+
"eval_weighted_f1": 0.855944421347496,
|
| 532 |
+
"step": 1560
|
| 533 |
+
},
|
| 534 |
+
{
|
| 535 |
+
"epoch": 8.077021822849808,
|
| 536 |
+
"grad_norm": 24.3969783782959,
|
| 537 |
+
"learning_rate": 1.3253561253561255e-05,
|
| 538 |
+
"loss": 0.2562,
|
| 539 |
+
"step": 1575
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 8.205391527599486,
|
| 543 |
+
"grad_norm": 52.96929931640625,
|
| 544 |
+
"learning_rate": 1.3111111111111113e-05,
|
| 545 |
+
"loss": 0.2434,
|
| 546 |
+
"step": 1600
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 8.333761232349165,
|
| 550 |
+
"grad_norm": 0.1785208284854889,
|
| 551 |
+
"learning_rate": 1.2968660968660968e-05,
|
| 552 |
+
"loss": 0.2622,
|
| 553 |
+
"step": 1625
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 8.462130937098845,
|
| 557 |
+
"grad_norm": 55.70664978027344,
|
| 558 |
+
"learning_rate": 1.2826210826210828e-05,
|
| 559 |
+
"loss": 0.4072,
|
| 560 |
+
"step": 1650
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 8.590500641848523,
|
| 564 |
+
"grad_norm": 0.30701741576194763,
|
| 565 |
+
"learning_rate": 1.2683760683760685e-05,
|
| 566 |
+
"loss": 0.2225,
|
| 567 |
+
"step": 1675
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 8.718870346598203,
|
| 571 |
+
"grad_norm": 28.206430435180664,
|
| 572 |
+
"learning_rate": 1.2541310541310542e-05,
|
| 573 |
+
"loss": 0.2043,
|
| 574 |
+
"step": 1700
|
| 575 |
+
},
|
| 576 |
+
{
|
| 577 |
+
"epoch": 8.847240051347882,
|
| 578 |
+
"grad_norm": 17.189531326293945,
|
| 579 |
+
"learning_rate": 1.23988603988604e-05,
|
| 580 |
+
"loss": 0.2392,
|
| 581 |
+
"step": 1725
|
| 582 |
+
},
|
| 583 |
+
{
|
| 584 |
+
"epoch": 8.975609756097562,
|
| 585 |
+
"grad_norm": 0.7421823740005493,
|
| 586 |
+
"learning_rate": 1.2256410256410259e-05,
|
| 587 |
+
"loss": 0.1808,
|
| 588 |
+
"step": 1750
|
| 589 |
+
},
|
| 590 |
+
{
|
| 591 |
+
"epoch": 9.0,
|
| 592 |
+
"eval_accuracy": 0.8460207612456747,
|
| 593 |
+
"eval_loss": 0.8177312612533569,
|
| 594 |
+
"eval_macro_f1": 0.8459505929286951,
|
| 595 |
+
"eval_runtime": 1.2245,
|
| 596 |
+
"eval_samples_per_second": 472.012,
|
| 597 |
+
"eval_steps_per_second": 30.215,
|
| 598 |
+
"eval_weighted_f1": 0.8460869818875526,
|
| 599 |
+
"step": 1755
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 9.102695763799744,
|
| 603 |
+
"grad_norm": 0.14991885423660278,
|
| 604 |
+
"learning_rate": 1.2113960113960114e-05,
|
| 605 |
+
"loss": 0.2549,
|
| 606 |
+
"step": 1775
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"epoch": 9.231065468549422,
|
| 610 |
+
"grad_norm": 0.17786766588687897,
|
| 611 |
+
"learning_rate": 1.1971509971509972e-05,
|
| 612 |
+
"loss": 0.1443,
|
| 613 |
+
"step": 1800
|
| 614 |
+
},
|
| 615 |
+
{
|
| 616 |
+
"epoch": 9.3594351732991,
|
| 617 |
+
"grad_norm": 0.13415665924549103,
|
| 618 |
+
"learning_rate": 1.182905982905983e-05,
|
| 619 |
+
"loss": 0.177,
|
| 620 |
+
"step": 1825
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 9.487804878048781,
|
| 624 |
+
"grad_norm": 0.14083965122699738,
|
| 625 |
+
"learning_rate": 1.1686609686609687e-05,
|
| 626 |
+
"loss": 0.1419,
|
| 627 |
+
"step": 1850
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 9.61617458279846,
|
| 631 |
+
"grad_norm": 0.13508036732673645,
|
| 632 |
+
"learning_rate": 1.1544159544159546e-05,
|
| 633 |
+
"loss": 0.2433,
|
| 634 |
+
"step": 1875
|
| 635 |
+
},
|
| 636 |
+
{
|
| 637 |
+
"epoch": 9.74454428754814,
|
| 638 |
+
"grad_norm": 4.5024542808532715,
|
| 639 |
+
"learning_rate": 1.1401709401709403e-05,
|
| 640 |
+
"loss": 0.3815,
|
| 641 |
+
"step": 1900
|
| 642 |
+
},
|
| 643 |
+
{
|
| 644 |
+
"epoch": 9.872913992297818,
|
| 645 |
+
"grad_norm": 0.13423407077789307,
|
| 646 |
+
"learning_rate": 1.125925925925926e-05,
|
| 647 |
+
"loss": 0.2225,
|
| 648 |
+
"step": 1925
|
| 649 |
+
},
|
| 650 |
+
{
|
| 651 |
+
"epoch": 10.0,
|
| 652 |
+
"grad_norm": 0.09093524515628815,
|
| 653 |
+
"learning_rate": 1.1116809116809118e-05,
|
| 654 |
+
"loss": 0.1477,
|
| 655 |
+
"step": 1950
|
| 656 |
+
},
|
| 657 |
+
{
|
| 658 |
+
"epoch": 10.0,
|
| 659 |
+
"eval_accuracy": 0.8529411764705882,
|
| 660 |
+
"eval_loss": 0.8117770552635193,
|
| 661 |
+
"eval_macro_f1": 0.8527336685616255,
|
| 662 |
+
"eval_runtime": 1.2038,
|
| 663 |
+
"eval_samples_per_second": 480.135,
|
| 664 |
+
"eval_steps_per_second": 30.735,
|
| 665 |
+
"eval_weighted_f1": 0.8528748100093116,
|
| 666 |
+
"step": 1950
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"epoch": 10.128369704749678,
|
| 670 |
+
"grad_norm": 0.13830582797527313,
|
| 671 |
+
"learning_rate": 1.0974358974358977e-05,
|
| 672 |
+
"loss": 0.1483,
|
| 673 |
+
"step": 1975
|
| 674 |
+
},
|
| 675 |
+
{
|
| 676 |
+
"epoch": 10.256739409499358,
|
| 677 |
+
"grad_norm": 0.15360301733016968,
|
| 678 |
+
"learning_rate": 1.0831908831908832e-05,
|
| 679 |
+
"loss": 0.0726,
|
| 680 |
+
"step": 2000
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 10.385109114249037,
|
| 684 |
+
"grad_norm": 0.26155582070350647,
|
| 685 |
+
"learning_rate": 1.068945868945869e-05,
|
| 686 |
+
"loss": 0.0944,
|
| 687 |
+
"step": 2025
|
| 688 |
+
},
|
| 689 |
+
{
|
| 690 |
+
"epoch": 10.513478818998717,
|
| 691 |
+
"grad_norm": 0.1234959065914154,
|
| 692 |
+
"learning_rate": 1.0547008547008549e-05,
|
| 693 |
+
"loss": 0.1216,
|
| 694 |
+
"step": 2050
|
| 695 |
+
},
|
| 696 |
+
{
|
| 697 |
+
"epoch": 10.641848523748395,
|
| 698 |
+
"grad_norm": 0.11160784959793091,
|
| 699 |
+
"learning_rate": 1.0404558404558405e-05,
|
| 700 |
+
"loss": 0.1355,
|
| 701 |
+
"step": 2075
|
| 702 |
+
},
|
| 703 |
+
{
|
| 704 |
+
"epoch": 10.770218228498074,
|
| 705 |
+
"grad_norm": 2.9147608280181885,
|
| 706 |
+
"learning_rate": 1.0262108262108264e-05,
|
| 707 |
+
"loss": 0.1659,
|
| 708 |
+
"step": 2100
|
| 709 |
+
},
|
| 710 |
+
{
|
| 711 |
+
"epoch": 10.898587933247754,
|
| 712 |
+
"grad_norm": 0.10950982570648193,
|
| 713 |
+
"learning_rate": 1.011965811965812e-05,
|
| 714 |
+
"loss": 0.1211,
|
| 715 |
+
"step": 2125
|
| 716 |
+
},
|
| 717 |
+
{
|
| 718 |
+
"epoch": 11.0,
|
| 719 |
+
"eval_accuracy": 0.8546712802768166,
|
| 720 |
+
"eval_loss": 0.8427116870880127,
|
| 721 |
+
"eval_macro_f1": 0.8538993720544736,
|
| 722 |
+
"eval_runtime": 1.1813,
|
| 723 |
+
"eval_samples_per_second": 489.284,
|
| 724 |
+
"eval_steps_per_second": 31.321,
|
| 725 |
+
"eval_weighted_f1": 0.8540993447704491,
|
| 726 |
+
"step": 2145
|
| 727 |
+
}
|
| 728 |
+
],
|
| 729 |
+
"logging_steps": 25,
|
| 730 |
+
"max_steps": 3900,
|
| 731 |
+
"num_input_tokens_seen": 0,
|
| 732 |
+
"num_train_epochs": 20,
|
| 733 |
+
"save_steps": 500,
|
| 734 |
+
"stateful_callbacks": {
|
| 735 |
+
"EarlyStoppingCallback": {
|
| 736 |
+
"args": {
|
| 737 |
+
"early_stopping_patience": 3,
|
| 738 |
+
"early_stopping_threshold": 0.0
|
| 739 |
+
},
|
| 740 |
+
"attributes": {
|
| 741 |
+
"early_stopping_patience_counter": 3
|
| 742 |
+
}
|
| 743 |
+
},
|
| 744 |
+
"TrainerControl": {
|
| 745 |
+
"args": {
|
| 746 |
+
"should_epoch_stop": false,
|
| 747 |
+
"should_evaluate": false,
|
| 748 |
+
"should_log": false,
|
| 749 |
+
"should_save": true,
|
| 750 |
+
"should_training_stop": true
|
| 751 |
+
},
|
| 752 |
+
"attributes": {}
|
| 753 |
+
}
|
| 754 |
+
},
|
| 755 |
+
"total_flos": 0.0,
|
| 756 |
+
"train_batch_size": 8,
|
| 757 |
+
"trial_name": null,
|
| 758 |
+
"trial_params": null
|
| 759 |
+
}
|
ood-vifactcheck-phobert-topic-domain-03-s42-20260814T112759Z-00ca7096/trainer/checkpoint-2145/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-04-s42-20260814T114226Z-00ca7096/trainer/checkpoint-752/special_tokens_map.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": "<mask>",
|
| 6 |
+
"pad_token": "<pad>",
|
| 7 |
+
"sep_token": "</s>",
|
| 8 |
+
"unk_token": "<unk>"
|
| 9 |
+
}
|
ood-vifactcheck-phobert-topic-domain-04-s42-20260814T114226Z-00ca7096/trainer/checkpoint-752/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 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": "</s>",
|
| 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 |
+
"64000": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 256,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "PhobertTokenizer",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|