Add files using upload-large-folder tool
Browse filesThis view is limited to 50 files because it contains too many changes. See raw diff
- ood-vifactcheck-phobert-source-source-00-s43-20260826T153605Z-bdd2c308/trainer/checkpoint-1456/added_tokens.json +3 -0
- ood-vifactcheck-phobert-source-source-00-s43-20260826T153605Z-bdd2c308/trainer/checkpoint-1456/bpe.codes +0 -0
- ood-vifactcheck-phobert-source-source-00-s43-20260826T153605Z-bdd2c308/trainer/checkpoint-1456/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-source-source-00-s43-20260826T153605Z-bdd2c308/trainer/checkpoint-1456/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-source-source-00-s43-20260826T153605Z-bdd2c308/trainer/checkpoint-1456/trainer_state.json +537 -0
- ood-vifactcheck-phobert-source-source-00-s43-20260826T153605Z-bdd2c308/trainer/checkpoint-1456/vocab.txt +0 -0
- ood-vifactcheck-phobert-source-source-00-s43-20260826T153605Z-bdd2c308/trainer/checkpoint-2002/added_tokens.json +3 -0
- ood-vifactcheck-phobert-source-source-00-s43-20260826T153605Z-bdd2c308/trainer/checkpoint-2002/bpe.codes +0 -0
- ood-vifactcheck-phobert-source-source-00-s43-20260826T153605Z-bdd2c308/trainer/checkpoint-2002/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-source-source-00-s43-20260826T153605Z-bdd2c308/trainer/checkpoint-2002/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-source-source-00-s43-20260826T153605Z-bdd2c308/trainer/checkpoint-2002/trainer_state.json +724 -0
- ood-vifactcheck-phobert-source-source-00-s43-20260826T153605Z-bdd2c308/trainer/checkpoint-2002/vocab.txt +0 -0
- ood-vifactcheck-phobert-source-source-01-s43-20260826T153943Z-bdd2c308/trainer/checkpoint-1026/added_tokens.json +3 -0
- ood-vifactcheck-phobert-source-source-01-s43-20260826T153943Z-bdd2c308/trainer/checkpoint-1026/bpe.codes +0 -0
- ood-vifactcheck-phobert-source-source-01-s43-20260826T153943Z-bdd2c308/trainer/checkpoint-1026/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-source-source-01-s43-20260826T153943Z-bdd2c308/trainer/checkpoint-1026/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-source-source-01-s43-20260826T153943Z-bdd2c308/trainer/checkpoint-1026/trainer_state.json +396 -0
- ood-vifactcheck-phobert-source-source-01-s43-20260826T153943Z-bdd2c308/trainer/checkpoint-1026/vocab.txt +0 -0
- ood-vifactcheck-phobert-source-source-01-s43-20260826T153943Z-bdd2c308/trainer/checkpoint-513/added_tokens.json +3 -0
- ood-vifactcheck-phobert-source-source-01-s43-20260826T153943Z-bdd2c308/trainer/checkpoint-513/bpe.codes +0 -0
- ood-vifactcheck-phobert-source-source-01-s43-20260826T153943Z-bdd2c308/trainer/checkpoint-513/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-source-source-01-s43-20260826T153943Z-bdd2c308/trainer/checkpoint-513/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-source-source-01-s43-20260826T153943Z-bdd2c308/trainer/checkpoint-513/trainer_state.json +216 -0
- ood-vifactcheck-phobert-source-source-01-s43-20260826T153943Z-bdd2c308/trainer/checkpoint-513/vocab.txt +0 -0
- ood-vifactcheck-phobert-source-source-02-s43-20260826T154150Z-bdd2c308/trainer/checkpoint-1448/added_tokens.json +3 -0
- ood-vifactcheck-phobert-source-source-02-s43-20260826T154150Z-bdd2c308/trainer/checkpoint-1448/bpe.codes +0 -0
- ood-vifactcheck-phobert-source-source-02-s43-20260826T154150Z-bdd2c308/trainer/checkpoint-1448/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-source-source-02-s43-20260826T154150Z-bdd2c308/trainer/checkpoint-1448/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-source-source-02-s43-20260826T154150Z-bdd2c308/trainer/checkpoint-1448/trainer_state.json +530 -0
- ood-vifactcheck-phobert-source-source-02-s43-20260826T154150Z-bdd2c308/trainer/checkpoint-1448/vocab.txt +0 -0
- ood-vifactcheck-phobert-source-source-02-s43-20260826T154150Z-bdd2c308/trainer/checkpoint-905/added_tokens.json +3 -0
- ood-vifactcheck-phobert-source-source-02-s43-20260826T154150Z-bdd2c308/trainer/checkpoint-905/bpe.codes +0 -0
- ood-vifactcheck-phobert-source-source-02-s43-20260826T154150Z-bdd2c308/trainer/checkpoint-905/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-source-source-02-s43-20260826T154150Z-bdd2c308/trainer/checkpoint-905/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-source-source-02-s43-20260826T154150Z-bdd2c308/trainer/checkpoint-905/trainer_state.json +350 -0
- ood-vifactcheck-phobert-source-source-02-s43-20260826T154150Z-bdd2c308/trainer/checkpoint-905/vocab.txt +0 -0
- ood-vifactcheck-phobert-source-source-03-s43-20260826T154435Z-bdd2c308/trainer/checkpoint-1464/added_tokens.json +3 -0
- ood-vifactcheck-phobert-source-source-03-s43-20260826T154435Z-bdd2c308/trainer/checkpoint-1464/bpe.codes +0 -0
- ood-vifactcheck-phobert-source-source-03-s43-20260826T154435Z-bdd2c308/trainer/checkpoint-1464/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-source-source-03-s43-20260826T154435Z-bdd2c308/trainer/checkpoint-1464/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-source-source-03-s43-20260826T154435Z-bdd2c308/trainer/checkpoint-1464/trainer_state.json +537 -0
- ood-vifactcheck-phobert-source-source-03-s43-20260826T154435Z-bdd2c308/trainer/checkpoint-1464/vocab.txt +0 -0
- ood-vifactcheck-phobert-source-source-03-s43-20260826T154435Z-bdd2c308/trainer/checkpoint-915/added_tokens.json +3 -0
- ood-vifactcheck-phobert-source-source-03-s43-20260826T154435Z-bdd2c308/trainer/checkpoint-915/bpe.codes +0 -0
- ood-vifactcheck-phobert-source-source-03-s43-20260826T154435Z-bdd2c308/trainer/checkpoint-915/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-source-source-03-s43-20260826T154435Z-bdd2c308/trainer/checkpoint-915/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-source-source-03-s43-20260826T154435Z-bdd2c308/trainer/checkpoint-915/trainer_state.json +350 -0
- ood-vifactcheck-phobert-source-source-03-s43-20260826T154435Z-bdd2c308/trainer/checkpoint-915/vocab.txt +0 -0
- ood-vifactcheck-phobert-source-source-04-s43-20260826T154723Z-bdd2c308/trainer/checkpoint-1408/added_tokens.json +3 -0
- ood-vifactcheck-phobert-source-source-04-s43-20260826T154723Z-bdd2c308/trainer/checkpoint-1408/bpe.codes +0 -0
ood-vifactcheck-phobert-source-source-00-s43-20260826T153605Z-bdd2c308/trainer/checkpoint-1456/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-source-source-00-s43-20260826T153605Z-bdd2c308/trainer/checkpoint-1456/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-source-source-00-s43-20260826T153605Z-bdd2c308/trainer/checkpoint-1456/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-source-source-00-s43-20260826T153605Z-bdd2c308/trainer/checkpoint-1456/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-source-source-00-s43-20260826T153605Z-bdd2c308/trainer/checkpoint-1456/trainer_state.json
ADDED
|
@@ -0,0 +1,537 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1456,
|
| 3 |
+
"best_metric": 0.8325881485992662,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-source-source-00-s43-20260826T153605Z-bdd2c308/trainer/checkpoint-1456",
|
| 5 |
+
"epoch": 8.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1456,
|
| 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": 6.686661720275879,
|
| 15 |
+
"learning_rate": 1.3186813186813187e-06,
|
| 16 |
+
"loss": 4.4622,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.27586206896551724,
|
| 21 |
+
"grad_norm": 9.171771049499512,
|
| 22 |
+
"learning_rate": 2.6923076923076923e-06,
|
| 23 |
+
"loss": 4.4141,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.41379310344827586,
|
| 28 |
+
"grad_norm": 6.143097400665283,
|
| 29 |
+
"learning_rate": 4.065934065934066e-06,
|
| 30 |
+
"loss": 4.4003,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5517241379310345,
|
| 35 |
+
"grad_norm": 6.021666049957275,
|
| 36 |
+
"learning_rate": 5.43956043956044e-06,
|
| 37 |
+
"loss": 4.3802,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6896551724137931,
|
| 42 |
+
"grad_norm": 5.948333263397217,
|
| 43 |
+
"learning_rate": 6.813186813186814e-06,
|
| 44 |
+
"loss": 4.35,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8275862068965517,
|
| 49 |
+
"grad_norm": 11.873772621154785,
|
| 50 |
+
"learning_rate": 8.186813186813188e-06,
|
| 51 |
+
"loss": 4.28,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9655172413793104,
|
| 56 |
+
"grad_norm": 10.159364700317383,
|
| 57 |
+
"learning_rate": 9.560439560439562e-06,
|
| 58 |
+
"loss": 4.1552,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.5342902711323764,
|
| 64 |
+
"eval_loss": 0.9755438566207886,
|
| 65 |
+
"eval_macro_f1": 0.5229715051741568,
|
| 66 |
+
"eval_runtime": 0.4665,
|
| 67 |
+
"eval_samples_per_second": 1343.937,
|
| 68 |
+
"eval_steps_per_second": 85.738,
|
| 69 |
+
"eval_weighted_f1": 0.5271983731472104,
|
| 70 |
+
"step": 182
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0993103448275863,
|
| 74 |
+
"grad_norm": 17.15261459350586,
|
| 75 |
+
"learning_rate": 1.0934065934065933e-05,
|
| 76 |
+
"loss": 3.7594,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.237241379310345,
|
| 81 |
+
"grad_norm": 28.597841262817383,
|
| 82 |
+
"learning_rate": 1.230769230769231e-05,
|
| 83 |
+
"loss": 3.569,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.3751724137931034,
|
| 88 |
+
"grad_norm": 45.464900970458984,
|
| 89 |
+
"learning_rate": 1.3681318681318683e-05,
|
| 90 |
+
"loss": 3.1891,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.513103448275862,
|
| 95 |
+
"grad_norm": 41.03870391845703,
|
| 96 |
+
"learning_rate": 1.5054945054945056e-05,
|
| 97 |
+
"loss": 2.8787,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.6510344827586207,
|
| 102 |
+
"grad_norm": 38.41599655151367,
|
| 103 |
+
"learning_rate": 1.642857142857143e-05,
|
| 104 |
+
"loss": 2.491,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7889655172413792,
|
| 109 |
+
"grad_norm": 50.78264236450195,
|
| 110 |
+
"learning_rate": 1.78021978021978e-05,
|
| 111 |
+
"loss": 2.3184,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.926896551724138,
|
| 116 |
+
"grad_norm": 64.83228302001953,
|
| 117 |
+
"learning_rate": 1.9175824175824178e-05,
|
| 118 |
+
"loss": 2.1515,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.7894736842105263,
|
| 124 |
+
"eval_loss": 0.5627413988113403,
|
| 125 |
+
"eval_macro_f1": 0.7897006323624473,
|
| 126 |
+
"eval_runtime": 0.4586,
|
| 127 |
+
"eval_samples_per_second": 1367.263,
|
| 128 |
+
"eval_steps_per_second": 87.226,
|
| 129 |
+
"eval_weighted_f1": 0.7907496468211039,
|
| 130 |
+
"step": 364
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.0606896551724136,
|
| 134 |
+
"grad_norm": 25.707651138305664,
|
| 135 |
+
"learning_rate": 1.993894993894994e-05,
|
| 136 |
+
"loss": 1.8722,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.1986206896551725,
|
| 141 |
+
"grad_norm": 56.51182174682617,
|
| 142 |
+
"learning_rate": 1.9786324786324788e-05,
|
| 143 |
+
"loss": 1.5633,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.336551724137931,
|
| 148 |
+
"grad_norm": 36.8007926940918,
|
| 149 |
+
"learning_rate": 1.9633699633699634e-05,
|
| 150 |
+
"loss": 1.4867,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.47448275862069,
|
| 155 |
+
"grad_norm": 43.388999938964844,
|
| 156 |
+
"learning_rate": 1.948107448107448e-05,
|
| 157 |
+
"loss": 1.5222,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.6124137931034483,
|
| 162 |
+
"grad_norm": 45.3205451965332,
|
| 163 |
+
"learning_rate": 1.932844932844933e-05,
|
| 164 |
+
"loss": 1.3935,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.750344827586207,
|
| 169 |
+
"grad_norm": 34.28248596191406,
|
| 170 |
+
"learning_rate": 1.9175824175824178e-05,
|
| 171 |
+
"loss": 1.6564,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.8882758620689657,
|
| 176 |
+
"grad_norm": 48.68022537231445,
|
| 177 |
+
"learning_rate": 1.9023199023199025e-05,
|
| 178 |
+
"loss": 1.7533,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 3.0,
|
| 183 |
+
"eval_accuracy": 0.8149920255183413,
|
| 184 |
+
"eval_loss": 0.5148630142211914,
|
| 185 |
+
"eval_macro_f1": 0.8147536795927627,
|
| 186 |
+
"eval_runtime": 0.4699,
|
| 187 |
+
"eval_samples_per_second": 1334.289,
|
| 188 |
+
"eval_steps_per_second": 85.122,
|
| 189 |
+
"eval_weighted_f1": 0.8157419112240982,
|
| 190 |
+
"step": 546
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.0220689655172412,
|
| 194 |
+
"grad_norm": 49.46675491333008,
|
| 195 |
+
"learning_rate": 1.887057387057387e-05,
|
| 196 |
+
"loss": 1.5452,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.16,
|
| 201 |
+
"grad_norm": 11.307979583740234,
|
| 202 |
+
"learning_rate": 1.8717948717948718e-05,
|
| 203 |
+
"loss": 1.1415,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.2979310344827586,
|
| 208 |
+
"grad_norm": 44.298988342285156,
|
| 209 |
+
"learning_rate": 1.8565323565323568e-05,
|
| 210 |
+
"loss": 1.0489,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.435862068965517,
|
| 215 |
+
"grad_norm": 20.450481414794922,
|
| 216 |
+
"learning_rate": 1.8412698412698415e-05,
|
| 217 |
+
"loss": 1.0247,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.573793103448276,
|
| 222 |
+
"grad_norm": 104.5062484741211,
|
| 223 |
+
"learning_rate": 1.826007326007326e-05,
|
| 224 |
+
"loss": 0.9099,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.7117241379310344,
|
| 229 |
+
"grad_norm": 42.8377571105957,
|
| 230 |
+
"learning_rate": 1.8107448107448108e-05,
|
| 231 |
+
"loss": 0.8576,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.849655172413793,
|
| 236 |
+
"grad_norm": 77.01546478271484,
|
| 237 |
+
"learning_rate": 1.7954822954822958e-05,
|
| 238 |
+
"loss": 1.2518,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.9875862068965517,
|
| 243 |
+
"grad_norm": 33.022186279296875,
|
| 244 |
+
"learning_rate": 1.78021978021978e-05,
|
| 245 |
+
"loss": 1.0455,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 4.0,
|
| 250 |
+
"eval_accuracy": 0.8229665071770335,
|
| 251 |
+
"eval_loss": 0.5552509427070618,
|
| 252 |
+
"eval_macro_f1": 0.8209902209748191,
|
| 253 |
+
"eval_runtime": 0.4353,
|
| 254 |
+
"eval_samples_per_second": 1440.462,
|
| 255 |
+
"eval_steps_per_second": 91.896,
|
| 256 |
+
"eval_weighted_f1": 0.8216399364269156,
|
| 257 |
+
"step": 728
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.121379310344827,
|
| 261 |
+
"grad_norm": 7.092617034912109,
|
| 262 |
+
"learning_rate": 1.764957264957265e-05,
|
| 263 |
+
"loss": 0.679,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.259310344827586,
|
| 268 |
+
"grad_norm": 32.29753112792969,
|
| 269 |
+
"learning_rate": 1.74969474969475e-05,
|
| 270 |
+
"loss": 0.748,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.397241379310345,
|
| 275 |
+
"grad_norm": 40.31268310546875,
|
| 276 |
+
"learning_rate": 1.7344322344322345e-05,
|
| 277 |
+
"loss": 0.7101,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.535172413793103,
|
| 282 |
+
"grad_norm": 43.25869369506836,
|
| 283 |
+
"learning_rate": 1.7191697191697195e-05,
|
| 284 |
+
"loss": 0.7234,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.673103448275862,
|
| 289 |
+
"grad_norm": 41.71521759033203,
|
| 290 |
+
"learning_rate": 1.703907203907204e-05,
|
| 291 |
+
"loss": 0.6363,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.811034482758621,
|
| 296 |
+
"grad_norm": 13.076591491699219,
|
| 297 |
+
"learning_rate": 1.688644688644689e-05,
|
| 298 |
+
"loss": 0.6994,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.94896551724138,
|
| 303 |
+
"grad_norm": 55.197872161865234,
|
| 304 |
+
"learning_rate": 1.6733821733821735e-05,
|
| 305 |
+
"loss": 0.7616,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 5.0,
|
| 310 |
+
"eval_accuracy": 0.8229665071770335,
|
| 311 |
+
"eval_loss": 0.6477085947990417,
|
| 312 |
+
"eval_macro_f1": 0.8216170421915824,
|
| 313 |
+
"eval_runtime": 0.4515,
|
| 314 |
+
"eval_samples_per_second": 1388.572,
|
| 315 |
+
"eval_steps_per_second": 88.585,
|
| 316 |
+
"eval_weighted_f1": 0.8225339518347661,
|
| 317 |
+
"step": 910
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.082758620689655,
|
| 321 |
+
"grad_norm": 105.48500061035156,
|
| 322 |
+
"learning_rate": 1.6581196581196585e-05,
|
| 323 |
+
"loss": 0.6473,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.220689655172414,
|
| 328 |
+
"grad_norm": 24.407155990600586,
|
| 329 |
+
"learning_rate": 1.642857142857143e-05,
|
| 330 |
+
"loss": 0.4367,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.358620689655172,
|
| 335 |
+
"grad_norm": 39.0447883605957,
|
| 336 |
+
"learning_rate": 1.627594627594628e-05,
|
| 337 |
+
"loss": 0.6359,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.496551724137931,
|
| 342 |
+
"grad_norm": 18.864765167236328,
|
| 343 |
+
"learning_rate": 1.6123321123321125e-05,
|
| 344 |
+
"loss": 0.6158,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.63448275862069,
|
| 349 |
+
"grad_norm": 39.232460021972656,
|
| 350 |
+
"learning_rate": 1.5970695970695972e-05,
|
| 351 |
+
"loss": 0.5719,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.772413793103448,
|
| 356 |
+
"grad_norm": 116.5931625366211,
|
| 357 |
+
"learning_rate": 1.581807081807082e-05,
|
| 358 |
+
"loss": 0.6208,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.910344827586207,
|
| 363 |
+
"grad_norm": 15.676041603088379,
|
| 364 |
+
"learning_rate": 1.5665445665445665e-05,
|
| 365 |
+
"loss": 0.715,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 6.0,
|
| 370 |
+
"eval_accuracy": 0.8309409888357256,
|
| 371 |
+
"eval_loss": 0.6961700320243835,
|
| 372 |
+
"eval_macro_f1": 0.8276455899292472,
|
| 373 |
+
"eval_runtime": 0.46,
|
| 374 |
+
"eval_samples_per_second": 1362.959,
|
| 375 |
+
"eval_steps_per_second": 86.951,
|
| 376 |
+
"eval_weighted_f1": 0.8288980583340427,
|
| 377 |
+
"step": 1092
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 6.0441379310344825,
|
| 381 |
+
"grad_norm": 45.342185974121094,
|
| 382 |
+
"learning_rate": 1.5512820512820516e-05,
|
| 383 |
+
"loss": 0.6092,
|
| 384 |
+
"step": 1100
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 6.182068965517241,
|
| 388 |
+
"grad_norm": 18.596826553344727,
|
| 389 |
+
"learning_rate": 1.5360195360195362e-05,
|
| 390 |
+
"loss": 0.4873,
|
| 391 |
+
"step": 1125
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 6.32,
|
| 395 |
+
"grad_norm": 22.515487670898438,
|
| 396 |
+
"learning_rate": 1.5207570207570207e-05,
|
| 397 |
+
"loss": 0.47,
|
| 398 |
+
"step": 1150
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.457931034482758,
|
| 402 |
+
"grad_norm": 50.00104904174805,
|
| 403 |
+
"learning_rate": 1.5054945054945056e-05,
|
| 404 |
+
"loss": 0.249,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.595862068965517,
|
| 409 |
+
"grad_norm": 1.878219723701477,
|
| 410 |
+
"learning_rate": 1.4902319902319904e-05,
|
| 411 |
+
"loss": 0.3166,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.733793103448276,
|
| 416 |
+
"grad_norm": 230.46878051757812,
|
| 417 |
+
"learning_rate": 1.4749694749694752e-05,
|
| 418 |
+
"loss": 0.4405,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.871724137931034,
|
| 423 |
+
"grad_norm": 2.7776849269866943,
|
| 424 |
+
"learning_rate": 1.4597069597069597e-05,
|
| 425 |
+
"loss": 0.598,
|
| 426 |
+
"step": 1250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 7.0,
|
| 430 |
+
"eval_accuracy": 0.8165869218500797,
|
| 431 |
+
"eval_loss": 0.8390244841575623,
|
| 432 |
+
"eval_macro_f1": 0.816228369025199,
|
| 433 |
+
"eval_runtime": 0.4453,
|
| 434 |
+
"eval_samples_per_second": 1408.127,
|
| 435 |
+
"eval_steps_per_second": 89.833,
|
| 436 |
+
"eval_weighted_f1": 0.8164647288261327,
|
| 437 |
+
"step": 1274
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 7.0055172413793105,
|
| 441 |
+
"grad_norm": 97.53742980957031,
|
| 442 |
+
"learning_rate": 1.4444444444444446e-05,
|
| 443 |
+
"loss": 0.309,
|
| 444 |
+
"step": 1275
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 7.1434482758620685,
|
| 448 |
+
"grad_norm": 2.1324219703674316,
|
| 449 |
+
"learning_rate": 1.4291819291819293e-05,
|
| 450 |
+
"loss": 0.2612,
|
| 451 |
+
"step": 1300
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 7.281379310344827,
|
| 455 |
+
"grad_norm": 1.8065520524978638,
|
| 456 |
+
"learning_rate": 1.4139194139194141e-05,
|
| 457 |
+
"loss": 0.3383,
|
| 458 |
+
"step": 1325
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 7.419310344827586,
|
| 462 |
+
"grad_norm": 0.28863993287086487,
|
| 463 |
+
"learning_rate": 1.3986568986568988e-05,
|
| 464 |
+
"loss": 0.4322,
|
| 465 |
+
"step": 1350
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 7.557241379310344,
|
| 469 |
+
"grad_norm": 5.9753851890563965,
|
| 470 |
+
"learning_rate": 1.3833943833943834e-05,
|
| 471 |
+
"loss": 0.4495,
|
| 472 |
+
"step": 1375
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 7.695172413793103,
|
| 476 |
+
"grad_norm": 64.97172546386719,
|
| 477 |
+
"learning_rate": 1.3681318681318683e-05,
|
| 478 |
+
"loss": 0.3148,
|
| 479 |
+
"step": 1400
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 7.833103448275862,
|
| 483 |
+
"grad_norm": 4.758451461791992,
|
| 484 |
+
"learning_rate": 1.3528693528693531e-05,
|
| 485 |
+
"loss": 0.3551,
|
| 486 |
+
"step": 1425
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 7.971034482758621,
|
| 490 |
+
"grad_norm": 0.2615467309951782,
|
| 491 |
+
"learning_rate": 1.3376068376068376e-05,
|
| 492 |
+
"loss": 0.3473,
|
| 493 |
+
"step": 1450
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 8.0,
|
| 497 |
+
"eval_accuracy": 0.8341307814992025,
|
| 498 |
+
"eval_loss": 0.8162709474563599,
|
| 499 |
+
"eval_macro_f1": 0.8325881485992662,
|
| 500 |
+
"eval_runtime": 0.4417,
|
| 501 |
+
"eval_samples_per_second": 1419.381,
|
| 502 |
+
"eval_steps_per_second": 90.551,
|
| 503 |
+
"eval_weighted_f1": 0.8335870769388658,
|
| 504 |
+
"step": 1456
|
| 505 |
+
}
|
| 506 |
+
],
|
| 507 |
+
"logging_steps": 25,
|
| 508 |
+
"max_steps": 3640,
|
| 509 |
+
"num_input_tokens_seen": 0,
|
| 510 |
+
"num_train_epochs": 20,
|
| 511 |
+
"save_steps": 500,
|
| 512 |
+
"stateful_callbacks": {
|
| 513 |
+
"EarlyStoppingCallback": {
|
| 514 |
+
"args": {
|
| 515 |
+
"early_stopping_patience": 3,
|
| 516 |
+
"early_stopping_threshold": 0.0
|
| 517 |
+
},
|
| 518 |
+
"attributes": {
|
| 519 |
+
"early_stopping_patience_counter": 0
|
| 520 |
+
}
|
| 521 |
+
},
|
| 522 |
+
"TrainerControl": {
|
| 523 |
+
"args": {
|
| 524 |
+
"should_epoch_stop": false,
|
| 525 |
+
"should_evaluate": false,
|
| 526 |
+
"should_log": false,
|
| 527 |
+
"should_save": true,
|
| 528 |
+
"should_training_stop": false
|
| 529 |
+
},
|
| 530 |
+
"attributes": {}
|
| 531 |
+
}
|
| 532 |
+
},
|
| 533 |
+
"total_flos": 0.0,
|
| 534 |
+
"train_batch_size": 8,
|
| 535 |
+
"trial_name": null,
|
| 536 |
+
"trial_params": null
|
| 537 |
+
}
|
ood-vifactcheck-phobert-source-source-00-s43-20260826T153605Z-bdd2c308/trainer/checkpoint-1456/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-source-source-00-s43-20260826T153605Z-bdd2c308/trainer/checkpoint-2002/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-source-source-00-s43-20260826T153605Z-bdd2c308/trainer/checkpoint-2002/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-source-source-00-s43-20260826T153605Z-bdd2c308/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-source-source-00-s43-20260826T153605Z-bdd2c308/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-source-source-00-s43-20260826T153605Z-bdd2c308/trainer/checkpoint-2002/trainer_state.json
ADDED
|
@@ -0,0 +1,724 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1456,
|
| 3 |
+
"best_metric": 0.8325881485992662,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-source-source-00-s43-20260826T153605Z-bdd2c308/trainer/checkpoint-1456",
|
| 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": 6.686661720275879,
|
| 15 |
+
"learning_rate": 1.3186813186813187e-06,
|
| 16 |
+
"loss": 4.4622,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.27586206896551724,
|
| 21 |
+
"grad_norm": 9.171771049499512,
|
| 22 |
+
"learning_rate": 2.6923076923076923e-06,
|
| 23 |
+
"loss": 4.4141,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.41379310344827586,
|
| 28 |
+
"grad_norm": 6.143097400665283,
|
| 29 |
+
"learning_rate": 4.065934065934066e-06,
|
| 30 |
+
"loss": 4.4003,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5517241379310345,
|
| 35 |
+
"grad_norm": 6.021666049957275,
|
| 36 |
+
"learning_rate": 5.43956043956044e-06,
|
| 37 |
+
"loss": 4.3802,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6896551724137931,
|
| 42 |
+
"grad_norm": 5.948333263397217,
|
| 43 |
+
"learning_rate": 6.813186813186814e-06,
|
| 44 |
+
"loss": 4.35,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8275862068965517,
|
| 49 |
+
"grad_norm": 11.873772621154785,
|
| 50 |
+
"learning_rate": 8.186813186813188e-06,
|
| 51 |
+
"loss": 4.28,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9655172413793104,
|
| 56 |
+
"grad_norm": 10.159364700317383,
|
| 57 |
+
"learning_rate": 9.560439560439562e-06,
|
| 58 |
+
"loss": 4.1552,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.5342902711323764,
|
| 64 |
+
"eval_loss": 0.9755438566207886,
|
| 65 |
+
"eval_macro_f1": 0.5229715051741568,
|
| 66 |
+
"eval_runtime": 0.4665,
|
| 67 |
+
"eval_samples_per_second": 1343.937,
|
| 68 |
+
"eval_steps_per_second": 85.738,
|
| 69 |
+
"eval_weighted_f1": 0.5271983731472104,
|
| 70 |
+
"step": 182
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0993103448275863,
|
| 74 |
+
"grad_norm": 17.15261459350586,
|
| 75 |
+
"learning_rate": 1.0934065934065933e-05,
|
| 76 |
+
"loss": 3.7594,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.237241379310345,
|
| 81 |
+
"grad_norm": 28.597841262817383,
|
| 82 |
+
"learning_rate": 1.230769230769231e-05,
|
| 83 |
+
"loss": 3.569,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.3751724137931034,
|
| 88 |
+
"grad_norm": 45.464900970458984,
|
| 89 |
+
"learning_rate": 1.3681318681318683e-05,
|
| 90 |
+
"loss": 3.1891,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.513103448275862,
|
| 95 |
+
"grad_norm": 41.03870391845703,
|
| 96 |
+
"learning_rate": 1.5054945054945056e-05,
|
| 97 |
+
"loss": 2.8787,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.6510344827586207,
|
| 102 |
+
"grad_norm": 38.41599655151367,
|
| 103 |
+
"learning_rate": 1.642857142857143e-05,
|
| 104 |
+
"loss": 2.491,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7889655172413792,
|
| 109 |
+
"grad_norm": 50.78264236450195,
|
| 110 |
+
"learning_rate": 1.78021978021978e-05,
|
| 111 |
+
"loss": 2.3184,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.926896551724138,
|
| 116 |
+
"grad_norm": 64.83228302001953,
|
| 117 |
+
"learning_rate": 1.9175824175824178e-05,
|
| 118 |
+
"loss": 2.1515,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.7894736842105263,
|
| 124 |
+
"eval_loss": 0.5627413988113403,
|
| 125 |
+
"eval_macro_f1": 0.7897006323624473,
|
| 126 |
+
"eval_runtime": 0.4586,
|
| 127 |
+
"eval_samples_per_second": 1367.263,
|
| 128 |
+
"eval_steps_per_second": 87.226,
|
| 129 |
+
"eval_weighted_f1": 0.7907496468211039,
|
| 130 |
+
"step": 364
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.0606896551724136,
|
| 134 |
+
"grad_norm": 25.707651138305664,
|
| 135 |
+
"learning_rate": 1.993894993894994e-05,
|
| 136 |
+
"loss": 1.8722,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.1986206896551725,
|
| 141 |
+
"grad_norm": 56.51182174682617,
|
| 142 |
+
"learning_rate": 1.9786324786324788e-05,
|
| 143 |
+
"loss": 1.5633,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.336551724137931,
|
| 148 |
+
"grad_norm": 36.8007926940918,
|
| 149 |
+
"learning_rate": 1.9633699633699634e-05,
|
| 150 |
+
"loss": 1.4867,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.47448275862069,
|
| 155 |
+
"grad_norm": 43.388999938964844,
|
| 156 |
+
"learning_rate": 1.948107448107448e-05,
|
| 157 |
+
"loss": 1.5222,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.6124137931034483,
|
| 162 |
+
"grad_norm": 45.3205451965332,
|
| 163 |
+
"learning_rate": 1.932844932844933e-05,
|
| 164 |
+
"loss": 1.3935,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.750344827586207,
|
| 169 |
+
"grad_norm": 34.28248596191406,
|
| 170 |
+
"learning_rate": 1.9175824175824178e-05,
|
| 171 |
+
"loss": 1.6564,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.8882758620689657,
|
| 176 |
+
"grad_norm": 48.68022537231445,
|
| 177 |
+
"learning_rate": 1.9023199023199025e-05,
|
| 178 |
+
"loss": 1.7533,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 3.0,
|
| 183 |
+
"eval_accuracy": 0.8149920255183413,
|
| 184 |
+
"eval_loss": 0.5148630142211914,
|
| 185 |
+
"eval_macro_f1": 0.8147536795927627,
|
| 186 |
+
"eval_runtime": 0.4699,
|
| 187 |
+
"eval_samples_per_second": 1334.289,
|
| 188 |
+
"eval_steps_per_second": 85.122,
|
| 189 |
+
"eval_weighted_f1": 0.8157419112240982,
|
| 190 |
+
"step": 546
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.0220689655172412,
|
| 194 |
+
"grad_norm": 49.46675491333008,
|
| 195 |
+
"learning_rate": 1.887057387057387e-05,
|
| 196 |
+
"loss": 1.5452,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.16,
|
| 201 |
+
"grad_norm": 11.307979583740234,
|
| 202 |
+
"learning_rate": 1.8717948717948718e-05,
|
| 203 |
+
"loss": 1.1415,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.2979310344827586,
|
| 208 |
+
"grad_norm": 44.298988342285156,
|
| 209 |
+
"learning_rate": 1.8565323565323568e-05,
|
| 210 |
+
"loss": 1.0489,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.435862068965517,
|
| 215 |
+
"grad_norm": 20.450481414794922,
|
| 216 |
+
"learning_rate": 1.8412698412698415e-05,
|
| 217 |
+
"loss": 1.0247,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.573793103448276,
|
| 222 |
+
"grad_norm": 104.5062484741211,
|
| 223 |
+
"learning_rate": 1.826007326007326e-05,
|
| 224 |
+
"loss": 0.9099,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.7117241379310344,
|
| 229 |
+
"grad_norm": 42.8377571105957,
|
| 230 |
+
"learning_rate": 1.8107448107448108e-05,
|
| 231 |
+
"loss": 0.8576,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.849655172413793,
|
| 236 |
+
"grad_norm": 77.01546478271484,
|
| 237 |
+
"learning_rate": 1.7954822954822958e-05,
|
| 238 |
+
"loss": 1.2518,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.9875862068965517,
|
| 243 |
+
"grad_norm": 33.022186279296875,
|
| 244 |
+
"learning_rate": 1.78021978021978e-05,
|
| 245 |
+
"loss": 1.0455,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 4.0,
|
| 250 |
+
"eval_accuracy": 0.8229665071770335,
|
| 251 |
+
"eval_loss": 0.5552509427070618,
|
| 252 |
+
"eval_macro_f1": 0.8209902209748191,
|
| 253 |
+
"eval_runtime": 0.4353,
|
| 254 |
+
"eval_samples_per_second": 1440.462,
|
| 255 |
+
"eval_steps_per_second": 91.896,
|
| 256 |
+
"eval_weighted_f1": 0.8216399364269156,
|
| 257 |
+
"step": 728
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.121379310344827,
|
| 261 |
+
"grad_norm": 7.092617034912109,
|
| 262 |
+
"learning_rate": 1.764957264957265e-05,
|
| 263 |
+
"loss": 0.679,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.259310344827586,
|
| 268 |
+
"grad_norm": 32.29753112792969,
|
| 269 |
+
"learning_rate": 1.74969474969475e-05,
|
| 270 |
+
"loss": 0.748,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.397241379310345,
|
| 275 |
+
"grad_norm": 40.31268310546875,
|
| 276 |
+
"learning_rate": 1.7344322344322345e-05,
|
| 277 |
+
"loss": 0.7101,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.535172413793103,
|
| 282 |
+
"grad_norm": 43.25869369506836,
|
| 283 |
+
"learning_rate": 1.7191697191697195e-05,
|
| 284 |
+
"loss": 0.7234,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.673103448275862,
|
| 289 |
+
"grad_norm": 41.71521759033203,
|
| 290 |
+
"learning_rate": 1.703907203907204e-05,
|
| 291 |
+
"loss": 0.6363,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.811034482758621,
|
| 296 |
+
"grad_norm": 13.076591491699219,
|
| 297 |
+
"learning_rate": 1.688644688644689e-05,
|
| 298 |
+
"loss": 0.6994,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.94896551724138,
|
| 303 |
+
"grad_norm": 55.197872161865234,
|
| 304 |
+
"learning_rate": 1.6733821733821735e-05,
|
| 305 |
+
"loss": 0.7616,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 5.0,
|
| 310 |
+
"eval_accuracy": 0.8229665071770335,
|
| 311 |
+
"eval_loss": 0.6477085947990417,
|
| 312 |
+
"eval_macro_f1": 0.8216170421915824,
|
| 313 |
+
"eval_runtime": 0.4515,
|
| 314 |
+
"eval_samples_per_second": 1388.572,
|
| 315 |
+
"eval_steps_per_second": 88.585,
|
| 316 |
+
"eval_weighted_f1": 0.8225339518347661,
|
| 317 |
+
"step": 910
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.082758620689655,
|
| 321 |
+
"grad_norm": 105.48500061035156,
|
| 322 |
+
"learning_rate": 1.6581196581196585e-05,
|
| 323 |
+
"loss": 0.6473,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.220689655172414,
|
| 328 |
+
"grad_norm": 24.407155990600586,
|
| 329 |
+
"learning_rate": 1.642857142857143e-05,
|
| 330 |
+
"loss": 0.4367,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.358620689655172,
|
| 335 |
+
"grad_norm": 39.0447883605957,
|
| 336 |
+
"learning_rate": 1.627594627594628e-05,
|
| 337 |
+
"loss": 0.6359,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.496551724137931,
|
| 342 |
+
"grad_norm": 18.864765167236328,
|
| 343 |
+
"learning_rate": 1.6123321123321125e-05,
|
| 344 |
+
"loss": 0.6158,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.63448275862069,
|
| 349 |
+
"grad_norm": 39.232460021972656,
|
| 350 |
+
"learning_rate": 1.5970695970695972e-05,
|
| 351 |
+
"loss": 0.5719,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.772413793103448,
|
| 356 |
+
"grad_norm": 116.5931625366211,
|
| 357 |
+
"learning_rate": 1.581807081807082e-05,
|
| 358 |
+
"loss": 0.6208,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.910344827586207,
|
| 363 |
+
"grad_norm": 15.676041603088379,
|
| 364 |
+
"learning_rate": 1.5665445665445665e-05,
|
| 365 |
+
"loss": 0.715,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 6.0,
|
| 370 |
+
"eval_accuracy": 0.8309409888357256,
|
| 371 |
+
"eval_loss": 0.6961700320243835,
|
| 372 |
+
"eval_macro_f1": 0.8276455899292472,
|
| 373 |
+
"eval_runtime": 0.46,
|
| 374 |
+
"eval_samples_per_second": 1362.959,
|
| 375 |
+
"eval_steps_per_second": 86.951,
|
| 376 |
+
"eval_weighted_f1": 0.8288980583340427,
|
| 377 |
+
"step": 1092
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 6.0441379310344825,
|
| 381 |
+
"grad_norm": 45.342185974121094,
|
| 382 |
+
"learning_rate": 1.5512820512820516e-05,
|
| 383 |
+
"loss": 0.6092,
|
| 384 |
+
"step": 1100
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 6.182068965517241,
|
| 388 |
+
"grad_norm": 18.596826553344727,
|
| 389 |
+
"learning_rate": 1.5360195360195362e-05,
|
| 390 |
+
"loss": 0.4873,
|
| 391 |
+
"step": 1125
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 6.32,
|
| 395 |
+
"grad_norm": 22.515487670898438,
|
| 396 |
+
"learning_rate": 1.5207570207570207e-05,
|
| 397 |
+
"loss": 0.47,
|
| 398 |
+
"step": 1150
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.457931034482758,
|
| 402 |
+
"grad_norm": 50.00104904174805,
|
| 403 |
+
"learning_rate": 1.5054945054945056e-05,
|
| 404 |
+
"loss": 0.249,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.595862068965517,
|
| 409 |
+
"grad_norm": 1.878219723701477,
|
| 410 |
+
"learning_rate": 1.4902319902319904e-05,
|
| 411 |
+
"loss": 0.3166,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.733793103448276,
|
| 416 |
+
"grad_norm": 230.46878051757812,
|
| 417 |
+
"learning_rate": 1.4749694749694752e-05,
|
| 418 |
+
"loss": 0.4405,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.871724137931034,
|
| 423 |
+
"grad_norm": 2.7776849269866943,
|
| 424 |
+
"learning_rate": 1.4597069597069597e-05,
|
| 425 |
+
"loss": 0.598,
|
| 426 |
+
"step": 1250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 7.0,
|
| 430 |
+
"eval_accuracy": 0.8165869218500797,
|
| 431 |
+
"eval_loss": 0.8390244841575623,
|
| 432 |
+
"eval_macro_f1": 0.816228369025199,
|
| 433 |
+
"eval_runtime": 0.4453,
|
| 434 |
+
"eval_samples_per_second": 1408.127,
|
| 435 |
+
"eval_steps_per_second": 89.833,
|
| 436 |
+
"eval_weighted_f1": 0.8164647288261327,
|
| 437 |
+
"step": 1274
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 7.0055172413793105,
|
| 441 |
+
"grad_norm": 97.53742980957031,
|
| 442 |
+
"learning_rate": 1.4444444444444446e-05,
|
| 443 |
+
"loss": 0.309,
|
| 444 |
+
"step": 1275
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 7.1434482758620685,
|
| 448 |
+
"grad_norm": 2.1324219703674316,
|
| 449 |
+
"learning_rate": 1.4291819291819293e-05,
|
| 450 |
+
"loss": 0.2612,
|
| 451 |
+
"step": 1300
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 7.281379310344827,
|
| 455 |
+
"grad_norm": 1.8065520524978638,
|
| 456 |
+
"learning_rate": 1.4139194139194141e-05,
|
| 457 |
+
"loss": 0.3383,
|
| 458 |
+
"step": 1325
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 7.419310344827586,
|
| 462 |
+
"grad_norm": 0.28863993287086487,
|
| 463 |
+
"learning_rate": 1.3986568986568988e-05,
|
| 464 |
+
"loss": 0.4322,
|
| 465 |
+
"step": 1350
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 7.557241379310344,
|
| 469 |
+
"grad_norm": 5.9753851890563965,
|
| 470 |
+
"learning_rate": 1.3833943833943834e-05,
|
| 471 |
+
"loss": 0.4495,
|
| 472 |
+
"step": 1375
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 7.695172413793103,
|
| 476 |
+
"grad_norm": 64.97172546386719,
|
| 477 |
+
"learning_rate": 1.3681318681318683e-05,
|
| 478 |
+
"loss": 0.3148,
|
| 479 |
+
"step": 1400
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 7.833103448275862,
|
| 483 |
+
"grad_norm": 4.758451461791992,
|
| 484 |
+
"learning_rate": 1.3528693528693531e-05,
|
| 485 |
+
"loss": 0.3551,
|
| 486 |
+
"step": 1425
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 7.971034482758621,
|
| 490 |
+
"grad_norm": 0.2615467309951782,
|
| 491 |
+
"learning_rate": 1.3376068376068376e-05,
|
| 492 |
+
"loss": 0.3473,
|
| 493 |
+
"step": 1450
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 8.0,
|
| 497 |
+
"eval_accuracy": 0.8341307814992025,
|
| 498 |
+
"eval_loss": 0.8162709474563599,
|
| 499 |
+
"eval_macro_f1": 0.8325881485992662,
|
| 500 |
+
"eval_runtime": 0.4417,
|
| 501 |
+
"eval_samples_per_second": 1419.381,
|
| 502 |
+
"eval_steps_per_second": 90.551,
|
| 503 |
+
"eval_weighted_f1": 0.8335870769388658,
|
| 504 |
+
"step": 1456
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 8.104827586206897,
|
| 508 |
+
"grad_norm": 27.80624008178711,
|
| 509 |
+
"learning_rate": 1.3223443223443224e-05,
|
| 510 |
+
"loss": 0.2155,
|
| 511 |
+
"step": 1475
|
| 512 |
+
},
|
| 513 |
+
{
|
| 514 |
+
"epoch": 8.242758620689655,
|
| 515 |
+
"grad_norm": 0.5220441222190857,
|
| 516 |
+
"learning_rate": 1.3070818070818073e-05,
|
| 517 |
+
"loss": 0.2857,
|
| 518 |
+
"step": 1500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 8.380689655172414,
|
| 522 |
+
"grad_norm": 0.21475531160831451,
|
| 523 |
+
"learning_rate": 1.291819291819292e-05,
|
| 524 |
+
"loss": 0.0943,
|
| 525 |
+
"step": 1525
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 8.518620689655172,
|
| 529 |
+
"grad_norm": 0.21331074833869934,
|
| 530 |
+
"learning_rate": 1.2765567765567766e-05,
|
| 531 |
+
"loss": 0.3014,
|
| 532 |
+
"step": 1550
|
| 533 |
+
},
|
| 534 |
+
{
|
| 535 |
+
"epoch": 8.65655172413793,
|
| 536 |
+
"grad_norm": 15.7271728515625,
|
| 537 |
+
"learning_rate": 1.2612942612942613e-05,
|
| 538 |
+
"loss": 0.2912,
|
| 539 |
+
"step": 1575
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 8.79448275862069,
|
| 543 |
+
"grad_norm": 0.22531968355178833,
|
| 544 |
+
"learning_rate": 1.2460317460317461e-05,
|
| 545 |
+
"loss": 0.3286,
|
| 546 |
+
"step": 1600
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 8.932413793103448,
|
| 550 |
+
"grad_norm": 0.38372373580932617,
|
| 551 |
+
"learning_rate": 1.230769230769231e-05,
|
| 552 |
+
"loss": 0.283,
|
| 553 |
+
"step": 1625
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 9.0,
|
| 557 |
+
"eval_accuracy": 0.8277511961722488,
|
| 558 |
+
"eval_loss": 0.9389778971672058,
|
| 559 |
+
"eval_macro_f1": 0.825350417598556,
|
| 560 |
+
"eval_runtime": 0.4551,
|
| 561 |
+
"eval_samples_per_second": 1377.836,
|
| 562 |
+
"eval_steps_per_second": 87.9,
|
| 563 |
+
"eval_weighted_f1": 0.8263106414941627,
|
| 564 |
+
"step": 1638
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"epoch": 9.066206896551725,
|
| 568 |
+
"grad_norm": 0.17771022021770477,
|
| 569 |
+
"learning_rate": 1.2155067155067155e-05,
|
| 570 |
+
"loss": 0.2743,
|
| 571 |
+
"step": 1650
|
| 572 |
+
},
|
| 573 |
+
{
|
| 574 |
+
"epoch": 9.204137931034483,
|
| 575 |
+
"grad_norm": 3.238870859146118,
|
| 576 |
+
"learning_rate": 1.2002442002442003e-05,
|
| 577 |
+
"loss": 0.1507,
|
| 578 |
+
"step": 1675
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 9.34206896551724,
|
| 582 |
+
"grad_norm": 0.40190988779067993,
|
| 583 |
+
"learning_rate": 1.1849816849816852e-05,
|
| 584 |
+
"loss": 0.2774,
|
| 585 |
+
"step": 1700
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 9.48,
|
| 589 |
+
"grad_norm": 3.4043807983398438,
|
| 590 |
+
"learning_rate": 1.1697191697191698e-05,
|
| 591 |
+
"loss": 0.2866,
|
| 592 |
+
"step": 1725
|
| 593 |
+
},
|
| 594 |
+
{
|
| 595 |
+
"epoch": 9.617931034482758,
|
| 596 |
+
"grad_norm": 209.0543212890625,
|
| 597 |
+
"learning_rate": 1.1544566544566545e-05,
|
| 598 |
+
"loss": 0.2125,
|
| 599 |
+
"step": 1750
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 9.755862068965516,
|
| 603 |
+
"grad_norm": 312.34130859375,
|
| 604 |
+
"learning_rate": 1.1391941391941392e-05,
|
| 605 |
+
"loss": 0.1535,
|
| 606 |
+
"step": 1775
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"epoch": 9.893793103448276,
|
| 610 |
+
"grad_norm": 0.2288227081298828,
|
| 611 |
+
"learning_rate": 1.123931623931624e-05,
|
| 612 |
+
"loss": 0.2826,
|
| 613 |
+
"step": 1800
|
| 614 |
+
},
|
| 615 |
+
{
|
| 616 |
+
"epoch": 10.0,
|
| 617 |
+
"eval_accuracy": 0.8277511961722488,
|
| 618 |
+
"eval_loss": 0.9429144263267517,
|
| 619 |
+
"eval_macro_f1": 0.8260337718267153,
|
| 620 |
+
"eval_runtime": 0.4858,
|
| 621 |
+
"eval_samples_per_second": 1290.566,
|
| 622 |
+
"eval_steps_per_second": 82.333,
|
| 623 |
+
"eval_weighted_f1": 0.8268623258994476,
|
| 624 |
+
"step": 1820
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"epoch": 10.027586206896551,
|
| 628 |
+
"grad_norm": 5.840743064880371,
|
| 629 |
+
"learning_rate": 1.1086691086691088e-05,
|
| 630 |
+
"loss": 0.105,
|
| 631 |
+
"step": 1825
|
| 632 |
+
},
|
| 633 |
+
{
|
| 634 |
+
"epoch": 10.16551724137931,
|
| 635 |
+
"grad_norm": 0.14806704223155975,
|
| 636 |
+
"learning_rate": 1.0934065934065933e-05,
|
| 637 |
+
"loss": 0.1058,
|
| 638 |
+
"step": 1850
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 10.303448275862069,
|
| 642 |
+
"grad_norm": 0.14528627693653107,
|
| 643 |
+
"learning_rate": 1.0781440781440782e-05,
|
| 644 |
+
"loss": 0.1641,
|
| 645 |
+
"step": 1875
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 10.441379310344828,
|
| 649 |
+
"grad_norm": 0.14897380769252777,
|
| 650 |
+
"learning_rate": 1.062881562881563e-05,
|
| 651 |
+
"loss": 0.13,
|
| 652 |
+
"step": 1900
|
| 653 |
+
},
|
| 654 |
+
{
|
| 655 |
+
"epoch": 10.579310344827586,
|
| 656 |
+
"grad_norm": 0.13757172226905823,
|
| 657 |
+
"learning_rate": 1.0476190476190477e-05,
|
| 658 |
+
"loss": 0.2867,
|
| 659 |
+
"step": 1925
|
| 660 |
+
},
|
| 661 |
+
{
|
| 662 |
+
"epoch": 10.717241379310344,
|
| 663 |
+
"grad_norm": 0.17192251980304718,
|
| 664 |
+
"learning_rate": 1.0323565323565325e-05,
|
| 665 |
+
"loss": 0.1653,
|
| 666 |
+
"step": 1950
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"epoch": 10.855172413793104,
|
| 670 |
+
"grad_norm": 0.1284424513578415,
|
| 671 |
+
"learning_rate": 1.0170940170940172e-05,
|
| 672 |
+
"loss": 0.1984,
|
| 673 |
+
"step": 1975
|
| 674 |
+
},
|
| 675 |
+
{
|
| 676 |
+
"epoch": 10.993103448275862,
|
| 677 |
+
"grad_norm": 0.15068988502025604,
|
| 678 |
+
"learning_rate": 1.0018315018315019e-05,
|
| 679 |
+
"loss": 0.1587,
|
| 680 |
+
"step": 2000
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 11.0,
|
| 684 |
+
"eval_accuracy": 0.8245614035087719,
|
| 685 |
+
"eval_loss": 0.9925813674926758,
|
| 686 |
+
"eval_macro_f1": 0.8241525421588328,
|
| 687 |
+
"eval_runtime": 0.4345,
|
| 688 |
+
"eval_samples_per_second": 1442.987,
|
| 689 |
+
"eval_steps_per_second": 92.057,
|
| 690 |
+
"eval_weighted_f1": 0.8246959240424885,
|
| 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": 3
|
| 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": true
|
| 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-source-source-00-s43-20260826T153605Z-bdd2c308/trainer/checkpoint-2002/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-source-source-01-s43-20260826T153943Z-bdd2c308/trainer/checkpoint-1026/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-source-source-01-s43-20260826T153943Z-bdd2c308/trainer/checkpoint-1026/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-source-source-01-s43-20260826T153943Z-bdd2c308/trainer/checkpoint-1026/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-source-source-01-s43-20260826T153943Z-bdd2c308/trainer/checkpoint-1026/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-source-source-01-s43-20260826T153943Z-bdd2c308/trainer/checkpoint-1026/trainer_state.json
ADDED
|
@@ -0,0 +1,396 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 513,
|
| 3 |
+
"best_metric": 0.8397977857809792,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-source-source-01-s43-20260826T153943Z-bdd2c308/trainer/checkpoint-513",
|
| 5 |
+
"epoch": 6.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1026,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.14684287812041116,
|
| 14 |
+
"grad_norm": 6.580949306488037,
|
| 15 |
+
"learning_rate": 1.4035087719298246e-06,
|
| 16 |
+
"loss": 4.399,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.2936857562408223,
|
| 21 |
+
"grad_norm": 8.75669002532959,
|
| 22 |
+
"learning_rate": 2.865497076023392e-06,
|
| 23 |
+
"loss": 4.4004,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.44052863436123346,
|
| 28 |
+
"grad_norm": 8.719416618347168,
|
| 29 |
+
"learning_rate": 4.3274853801169596e-06,
|
| 30 |
+
"loss": 4.4158,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5873715124816447,
|
| 35 |
+
"grad_norm": 6.355876922607422,
|
| 36 |
+
"learning_rate": 5.789473684210527e-06,
|
| 37 |
+
"loss": 4.3664,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.7342143906020558,
|
| 42 |
+
"grad_norm": 7.792836666107178,
|
| 43 |
+
"learning_rate": 7.251461988304094e-06,
|
| 44 |
+
"loss": 4.3561,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8810572687224669,
|
| 49 |
+
"grad_norm": 8.224111557006836,
|
| 50 |
+
"learning_rate": 8.713450292397661e-06,
|
| 51 |
+
"loss": 4.3037,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 1.0,
|
| 56 |
+
"eval_accuracy": 0.5604982206405694,
|
| 57 |
+
"eval_loss": 0.9842637777328491,
|
| 58 |
+
"eval_macro_f1": 0.5376388922321295,
|
| 59 |
+
"eval_runtime": 0.4105,
|
| 60 |
+
"eval_samples_per_second": 1369.111,
|
| 61 |
+
"eval_steps_per_second": 87.701,
|
| 62 |
+
"eval_weighted_f1": 0.5402353459452872,
|
| 63 |
+
"step": 171
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 1.0234948604992657,
|
| 67 |
+
"grad_norm": 11.988283157348633,
|
| 68 |
+
"learning_rate": 1.017543859649123e-05,
|
| 69 |
+
"loss": 4.0271,
|
| 70 |
+
"step": 175
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.170337738619677,
|
| 74 |
+
"grad_norm": 15.458869934082031,
|
| 75 |
+
"learning_rate": 1.1637426900584797e-05,
|
| 76 |
+
"loss": 3.7639,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.3171806167400881,
|
| 81 |
+
"grad_norm": 38.16561508178711,
|
| 82 |
+
"learning_rate": 1.3099415204678362e-05,
|
| 83 |
+
"loss": 3.3136,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.4640234948604993,
|
| 88 |
+
"grad_norm": 36.84794998168945,
|
| 89 |
+
"learning_rate": 1.4561403508771931e-05,
|
| 90 |
+
"loss": 2.9758,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.6108663729809103,
|
| 95 |
+
"grad_norm": 46.71482849121094,
|
| 96 |
+
"learning_rate": 1.60233918128655e-05,
|
| 97 |
+
"loss": 2.6033,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.7577092511013217,
|
| 102 |
+
"grad_norm": 57.56834030151367,
|
| 103 |
+
"learning_rate": 1.7485380116959064e-05,
|
| 104 |
+
"loss": 2.5563,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.9045521292217327,
|
| 109 |
+
"grad_norm": 78.12703704833984,
|
| 110 |
+
"learning_rate": 1.894736842105263e-05,
|
| 111 |
+
"loss": 2.2356,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 2.0,
|
| 116 |
+
"eval_accuracy": 0.8042704626334519,
|
| 117 |
+
"eval_loss": 0.4996437728404999,
|
| 118 |
+
"eval_macro_f1": 0.8030879007132304,
|
| 119 |
+
"eval_runtime": 0.4227,
|
| 120 |
+
"eval_samples_per_second": 1329.582,
|
| 121 |
+
"eval_steps_per_second": 85.169,
|
| 122 |
+
"eval_weighted_f1": 0.8030392062354804,
|
| 123 |
+
"step": 342
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 2.0469897209985315,
|
| 127 |
+
"grad_norm": 56.90924072265625,
|
| 128 |
+
"learning_rate": 1.99545159194282e-05,
|
| 129 |
+
"loss": 1.9874,
|
| 130 |
+
"step": 350
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.193832599118943,
|
| 134 |
+
"grad_norm": 31.367389678955078,
|
| 135 |
+
"learning_rate": 1.9792072774528916e-05,
|
| 136 |
+
"loss": 1.7393,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.340675477239354,
|
| 141 |
+
"grad_norm": 9.992380142211914,
|
| 142 |
+
"learning_rate": 1.962962962962963e-05,
|
| 143 |
+
"loss": 1.6523,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.4875183553597653,
|
| 148 |
+
"grad_norm": 52.02292251586914,
|
| 149 |
+
"learning_rate": 1.9467186484730346e-05,
|
| 150 |
+
"loss": 1.6325,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.6343612334801763,
|
| 155 |
+
"grad_norm": 27.05382537841797,
|
| 156 |
+
"learning_rate": 1.930474333983106e-05,
|
| 157 |
+
"loss": 1.654,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.7812041116005872,
|
| 162 |
+
"grad_norm": 33.80085754394531,
|
| 163 |
+
"learning_rate": 1.9142300194931776e-05,
|
| 164 |
+
"loss": 1.5247,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.9280469897209986,
|
| 169 |
+
"grad_norm": 81.95863342285156,
|
| 170 |
+
"learning_rate": 1.897985705003249e-05,
|
| 171 |
+
"loss": 1.6259,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 3.0,
|
| 176 |
+
"eval_accuracy": 0.8398576512455516,
|
| 177 |
+
"eval_loss": 0.4565703272819519,
|
| 178 |
+
"eval_macro_f1": 0.8397977857809792,
|
| 179 |
+
"eval_runtime": 0.3942,
|
| 180 |
+
"eval_samples_per_second": 1425.841,
|
| 181 |
+
"eval_steps_per_second": 91.335,
|
| 182 |
+
"eval_weighted_f1": 0.8394908938664434,
|
| 183 |
+
"step": 513
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 3.0704845814977975,
|
| 187 |
+
"grad_norm": 31.45627784729004,
|
| 188 |
+
"learning_rate": 1.8817413905133203e-05,
|
| 189 |
+
"loss": 1.4359,
|
| 190 |
+
"step": 525
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.2173274596182084,
|
| 194 |
+
"grad_norm": 36.0887565612793,
|
| 195 |
+
"learning_rate": 1.8654970760233918e-05,
|
| 196 |
+
"loss": 1.1627,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.36417033773862,
|
| 201 |
+
"grad_norm": 46.31768035888672,
|
| 202 |
+
"learning_rate": 1.8492527615334633e-05,
|
| 203 |
+
"loss": 1.0561,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.511013215859031,
|
| 208 |
+
"grad_norm": 45.30815505981445,
|
| 209 |
+
"learning_rate": 1.833008447043535e-05,
|
| 210 |
+
"loss": 0.9744,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.657856093979442,
|
| 215 |
+
"grad_norm": 19.926179885864258,
|
| 216 |
+
"learning_rate": 1.8167641325536064e-05,
|
| 217 |
+
"loss": 1.0005,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.804698972099853,
|
| 222 |
+
"grad_norm": 42.985836029052734,
|
| 223 |
+
"learning_rate": 1.800519818063678e-05,
|
| 224 |
+
"loss": 0.9474,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.951541850220264,
|
| 229 |
+
"grad_norm": 24.591243743896484,
|
| 230 |
+
"learning_rate": 1.7842755035737494e-05,
|
| 231 |
+
"loss": 0.8966,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 4.0,
|
| 236 |
+
"eval_accuracy": 0.8274021352313167,
|
| 237 |
+
"eval_loss": 0.5256436467170715,
|
| 238 |
+
"eval_macro_f1": 0.8273996161080882,
|
| 239 |
+
"eval_runtime": 0.4073,
|
| 240 |
+
"eval_samples_per_second": 1379.965,
|
| 241 |
+
"eval_steps_per_second": 88.396,
|
| 242 |
+
"eval_weighted_f1": 0.8271863930243845,
|
| 243 |
+
"step": 684
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 4.093979441997063,
|
| 247 |
+
"grad_norm": 5.707021713256836,
|
| 248 |
+
"learning_rate": 1.768031189083821e-05,
|
| 249 |
+
"loss": 0.7773,
|
| 250 |
+
"step": 700
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 4.240822320117474,
|
| 254 |
+
"grad_norm": 63.49526596069336,
|
| 255 |
+
"learning_rate": 1.751786874593892e-05,
|
| 256 |
+
"loss": 0.7973,
|
| 257 |
+
"step": 725
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.387665198237886,
|
| 261 |
+
"grad_norm": 76.21549224853516,
|
| 262 |
+
"learning_rate": 1.7355425601039636e-05,
|
| 263 |
+
"loss": 0.8493,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.534508076358296,
|
| 268 |
+
"grad_norm": 38.05754470825195,
|
| 269 |
+
"learning_rate": 1.719298245614035e-05,
|
| 270 |
+
"loss": 0.8036,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.681350954478708,
|
| 275 |
+
"grad_norm": 30.598918914794922,
|
| 276 |
+
"learning_rate": 1.7030539311241066e-05,
|
| 277 |
+
"loss": 0.7502,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.828193832599119,
|
| 282 |
+
"grad_norm": 42.295555114746094,
|
| 283 |
+
"learning_rate": 1.686809616634178e-05,
|
| 284 |
+
"loss": 0.6921,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.975036710719531,
|
| 289 |
+
"grad_norm": 88.1772232055664,
|
| 290 |
+
"learning_rate": 1.6705653021442497e-05,
|
| 291 |
+
"loss": 0.7119,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 5.0,
|
| 296 |
+
"eval_accuracy": 0.8202846975088968,
|
| 297 |
+
"eval_loss": 0.6434597969055176,
|
| 298 |
+
"eval_macro_f1": 0.8189568294159987,
|
| 299 |
+
"eval_runtime": 0.4025,
|
| 300 |
+
"eval_samples_per_second": 1396.322,
|
| 301 |
+
"eval_steps_per_second": 89.444,
|
| 302 |
+
"eval_weighted_f1": 0.8191643978980605,
|
| 303 |
+
"step": 855
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 5.117474302496329,
|
| 307 |
+
"grad_norm": 85.53871154785156,
|
| 308 |
+
"learning_rate": 1.654320987654321e-05,
|
| 309 |
+
"loss": 0.5731,
|
| 310 |
+
"step": 875
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 5.26431718061674,
|
| 314 |
+
"grad_norm": 38.318050384521484,
|
| 315 |
+
"learning_rate": 1.6380766731643927e-05,
|
| 316 |
+
"loss": 0.6352,
|
| 317 |
+
"step": 900
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.411160058737151,
|
| 321 |
+
"grad_norm": 43.86847686767578,
|
| 322 |
+
"learning_rate": 1.621832358674464e-05,
|
| 323 |
+
"loss": 0.4375,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.558002936857562,
|
| 328 |
+
"grad_norm": 94.6635513305664,
|
| 329 |
+
"learning_rate": 1.6055880441845354e-05,
|
| 330 |
+
"loss": 0.5495,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.704845814977974,
|
| 335 |
+
"grad_norm": 9.676894187927246,
|
| 336 |
+
"learning_rate": 1.589343729694607e-05,
|
| 337 |
+
"loss": 0.495,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.851688693098385,
|
| 342 |
+
"grad_norm": 52.328853607177734,
|
| 343 |
+
"learning_rate": 1.5730994152046787e-05,
|
| 344 |
+
"loss": 0.6385,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.998531571218796,
|
| 349 |
+
"grad_norm": 113.49581909179688,
|
| 350 |
+
"learning_rate": 1.55685510071475e-05,
|
| 351 |
+
"loss": 0.5591,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 6.0,
|
| 356 |
+
"eval_accuracy": 0.8149466192170819,
|
| 357 |
+
"eval_loss": 0.7918155193328857,
|
| 358 |
+
"eval_macro_f1": 0.8140392911384309,
|
| 359 |
+
"eval_runtime": 0.4216,
|
| 360 |
+
"eval_samples_per_second": 1332.997,
|
| 361 |
+
"eval_steps_per_second": 85.388,
|
| 362 |
+
"eval_weighted_f1": 0.814167633470207,
|
| 363 |
+
"step": 1026
|
| 364 |
+
}
|
| 365 |
+
],
|
| 366 |
+
"logging_steps": 25,
|
| 367 |
+
"max_steps": 3420,
|
| 368 |
+
"num_input_tokens_seen": 0,
|
| 369 |
+
"num_train_epochs": 20,
|
| 370 |
+
"save_steps": 500,
|
| 371 |
+
"stateful_callbacks": {
|
| 372 |
+
"EarlyStoppingCallback": {
|
| 373 |
+
"args": {
|
| 374 |
+
"early_stopping_patience": 3,
|
| 375 |
+
"early_stopping_threshold": 0.0
|
| 376 |
+
},
|
| 377 |
+
"attributes": {
|
| 378 |
+
"early_stopping_patience_counter": 3
|
| 379 |
+
}
|
| 380 |
+
},
|
| 381 |
+
"TrainerControl": {
|
| 382 |
+
"args": {
|
| 383 |
+
"should_epoch_stop": false,
|
| 384 |
+
"should_evaluate": false,
|
| 385 |
+
"should_log": false,
|
| 386 |
+
"should_save": true,
|
| 387 |
+
"should_training_stop": true
|
| 388 |
+
},
|
| 389 |
+
"attributes": {}
|
| 390 |
+
}
|
| 391 |
+
},
|
| 392 |
+
"total_flos": 0.0,
|
| 393 |
+
"train_batch_size": 8,
|
| 394 |
+
"trial_name": null,
|
| 395 |
+
"trial_params": null
|
| 396 |
+
}
|
ood-vifactcheck-phobert-source-source-01-s43-20260826T153943Z-bdd2c308/trainer/checkpoint-1026/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-source-source-01-s43-20260826T153943Z-bdd2c308/trainer/checkpoint-513/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-source-source-01-s43-20260826T153943Z-bdd2c308/trainer/checkpoint-513/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-source-source-01-s43-20260826T153943Z-bdd2c308/trainer/checkpoint-513/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-source-source-01-s43-20260826T153943Z-bdd2c308/trainer/checkpoint-513/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-source-source-01-s43-20260826T153943Z-bdd2c308/trainer/checkpoint-513/trainer_state.json
ADDED
|
@@ -0,0 +1,216 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 513,
|
| 3 |
+
"best_metric": 0.8397977857809792,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-source-source-01-s43-20260826T153943Z-bdd2c308/trainer/checkpoint-513",
|
| 5 |
+
"epoch": 3.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 513,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.14684287812041116,
|
| 14 |
+
"grad_norm": 6.580949306488037,
|
| 15 |
+
"learning_rate": 1.4035087719298246e-06,
|
| 16 |
+
"loss": 4.399,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.2936857562408223,
|
| 21 |
+
"grad_norm": 8.75669002532959,
|
| 22 |
+
"learning_rate": 2.865497076023392e-06,
|
| 23 |
+
"loss": 4.4004,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.44052863436123346,
|
| 28 |
+
"grad_norm": 8.719416618347168,
|
| 29 |
+
"learning_rate": 4.3274853801169596e-06,
|
| 30 |
+
"loss": 4.4158,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5873715124816447,
|
| 35 |
+
"grad_norm": 6.355876922607422,
|
| 36 |
+
"learning_rate": 5.789473684210527e-06,
|
| 37 |
+
"loss": 4.3664,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.7342143906020558,
|
| 42 |
+
"grad_norm": 7.792836666107178,
|
| 43 |
+
"learning_rate": 7.251461988304094e-06,
|
| 44 |
+
"loss": 4.3561,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8810572687224669,
|
| 49 |
+
"grad_norm": 8.224111557006836,
|
| 50 |
+
"learning_rate": 8.713450292397661e-06,
|
| 51 |
+
"loss": 4.3037,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 1.0,
|
| 56 |
+
"eval_accuracy": 0.5604982206405694,
|
| 57 |
+
"eval_loss": 0.9842637777328491,
|
| 58 |
+
"eval_macro_f1": 0.5376388922321295,
|
| 59 |
+
"eval_runtime": 0.4105,
|
| 60 |
+
"eval_samples_per_second": 1369.111,
|
| 61 |
+
"eval_steps_per_second": 87.701,
|
| 62 |
+
"eval_weighted_f1": 0.5402353459452872,
|
| 63 |
+
"step": 171
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 1.0234948604992657,
|
| 67 |
+
"grad_norm": 11.988283157348633,
|
| 68 |
+
"learning_rate": 1.017543859649123e-05,
|
| 69 |
+
"loss": 4.0271,
|
| 70 |
+
"step": 175
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.170337738619677,
|
| 74 |
+
"grad_norm": 15.458869934082031,
|
| 75 |
+
"learning_rate": 1.1637426900584797e-05,
|
| 76 |
+
"loss": 3.7639,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.3171806167400881,
|
| 81 |
+
"grad_norm": 38.16561508178711,
|
| 82 |
+
"learning_rate": 1.3099415204678362e-05,
|
| 83 |
+
"loss": 3.3136,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.4640234948604993,
|
| 88 |
+
"grad_norm": 36.84794998168945,
|
| 89 |
+
"learning_rate": 1.4561403508771931e-05,
|
| 90 |
+
"loss": 2.9758,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.6108663729809103,
|
| 95 |
+
"grad_norm": 46.71482849121094,
|
| 96 |
+
"learning_rate": 1.60233918128655e-05,
|
| 97 |
+
"loss": 2.6033,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.7577092511013217,
|
| 102 |
+
"grad_norm": 57.56834030151367,
|
| 103 |
+
"learning_rate": 1.7485380116959064e-05,
|
| 104 |
+
"loss": 2.5563,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.9045521292217327,
|
| 109 |
+
"grad_norm": 78.12703704833984,
|
| 110 |
+
"learning_rate": 1.894736842105263e-05,
|
| 111 |
+
"loss": 2.2356,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 2.0,
|
| 116 |
+
"eval_accuracy": 0.8042704626334519,
|
| 117 |
+
"eval_loss": 0.4996437728404999,
|
| 118 |
+
"eval_macro_f1": 0.8030879007132304,
|
| 119 |
+
"eval_runtime": 0.4227,
|
| 120 |
+
"eval_samples_per_second": 1329.582,
|
| 121 |
+
"eval_steps_per_second": 85.169,
|
| 122 |
+
"eval_weighted_f1": 0.8030392062354804,
|
| 123 |
+
"step": 342
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 2.0469897209985315,
|
| 127 |
+
"grad_norm": 56.90924072265625,
|
| 128 |
+
"learning_rate": 1.99545159194282e-05,
|
| 129 |
+
"loss": 1.9874,
|
| 130 |
+
"step": 350
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.193832599118943,
|
| 134 |
+
"grad_norm": 31.367389678955078,
|
| 135 |
+
"learning_rate": 1.9792072774528916e-05,
|
| 136 |
+
"loss": 1.7393,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.340675477239354,
|
| 141 |
+
"grad_norm": 9.992380142211914,
|
| 142 |
+
"learning_rate": 1.962962962962963e-05,
|
| 143 |
+
"loss": 1.6523,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.4875183553597653,
|
| 148 |
+
"grad_norm": 52.02292251586914,
|
| 149 |
+
"learning_rate": 1.9467186484730346e-05,
|
| 150 |
+
"loss": 1.6325,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.6343612334801763,
|
| 155 |
+
"grad_norm": 27.05382537841797,
|
| 156 |
+
"learning_rate": 1.930474333983106e-05,
|
| 157 |
+
"loss": 1.654,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.7812041116005872,
|
| 162 |
+
"grad_norm": 33.80085754394531,
|
| 163 |
+
"learning_rate": 1.9142300194931776e-05,
|
| 164 |
+
"loss": 1.5247,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.9280469897209986,
|
| 169 |
+
"grad_norm": 81.95863342285156,
|
| 170 |
+
"learning_rate": 1.897985705003249e-05,
|
| 171 |
+
"loss": 1.6259,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 3.0,
|
| 176 |
+
"eval_accuracy": 0.8398576512455516,
|
| 177 |
+
"eval_loss": 0.4565703272819519,
|
| 178 |
+
"eval_macro_f1": 0.8397977857809792,
|
| 179 |
+
"eval_runtime": 0.3942,
|
| 180 |
+
"eval_samples_per_second": 1425.841,
|
| 181 |
+
"eval_steps_per_second": 91.335,
|
| 182 |
+
"eval_weighted_f1": 0.8394908938664434,
|
| 183 |
+
"step": 513
|
| 184 |
+
}
|
| 185 |
+
],
|
| 186 |
+
"logging_steps": 25,
|
| 187 |
+
"max_steps": 3420,
|
| 188 |
+
"num_input_tokens_seen": 0,
|
| 189 |
+
"num_train_epochs": 20,
|
| 190 |
+
"save_steps": 500,
|
| 191 |
+
"stateful_callbacks": {
|
| 192 |
+
"EarlyStoppingCallback": {
|
| 193 |
+
"args": {
|
| 194 |
+
"early_stopping_patience": 3,
|
| 195 |
+
"early_stopping_threshold": 0.0
|
| 196 |
+
},
|
| 197 |
+
"attributes": {
|
| 198 |
+
"early_stopping_patience_counter": 0
|
| 199 |
+
}
|
| 200 |
+
},
|
| 201 |
+
"TrainerControl": {
|
| 202 |
+
"args": {
|
| 203 |
+
"should_epoch_stop": false,
|
| 204 |
+
"should_evaluate": false,
|
| 205 |
+
"should_log": false,
|
| 206 |
+
"should_save": true,
|
| 207 |
+
"should_training_stop": false
|
| 208 |
+
},
|
| 209 |
+
"attributes": {}
|
| 210 |
+
}
|
| 211 |
+
},
|
| 212 |
+
"total_flos": 0.0,
|
| 213 |
+
"train_batch_size": 8,
|
| 214 |
+
"trial_name": null,
|
| 215 |
+
"trial_params": null
|
| 216 |
+
}
|
ood-vifactcheck-phobert-source-source-01-s43-20260826T153943Z-bdd2c308/trainer/checkpoint-513/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-source-source-02-s43-20260826T154150Z-bdd2c308/trainer/checkpoint-1448/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-source-source-02-s43-20260826T154150Z-bdd2c308/trainer/checkpoint-1448/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-source-source-02-s43-20260826T154150Z-bdd2c308/trainer/checkpoint-1448/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-source-source-02-s43-20260826T154150Z-bdd2c308/trainer/checkpoint-1448/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-source-source-02-s43-20260826T154150Z-bdd2c308/trainer/checkpoint-1448/trainer_state.json
ADDED
|
@@ -0,0 +1,530 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 905,
|
| 3 |
+
"best_metric": 0.8337704818761008,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-source-source-02-s43-20260826T154150Z-bdd2c308/trainer/checkpoint-905",
|
| 5 |
+
"epoch": 8.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1448,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.13812154696132597,
|
| 14 |
+
"grad_norm": 6.335784912109375,
|
| 15 |
+
"learning_rate": 1.3259668508287293e-06,
|
| 16 |
+
"loss": 4.4522,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.27624309392265195,
|
| 21 |
+
"grad_norm": 6.684322834014893,
|
| 22 |
+
"learning_rate": 2.707182320441989e-06,
|
| 23 |
+
"loss": 4.4467,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.4143646408839779,
|
| 28 |
+
"grad_norm": 7.152876377105713,
|
| 29 |
+
"learning_rate": 4.088397790055249e-06,
|
| 30 |
+
"loss": 4.4027,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5524861878453039,
|
| 35 |
+
"grad_norm": 8.315168380737305,
|
| 36 |
+
"learning_rate": 5.469613259668509e-06,
|
| 37 |
+
"loss": 4.3901,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6906077348066298,
|
| 42 |
+
"grad_norm": 7.618349552154541,
|
| 43 |
+
"learning_rate": 6.850828729281769e-06,
|
| 44 |
+
"loss": 4.3232,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8287292817679558,
|
| 49 |
+
"grad_norm": 9.473018646240234,
|
| 50 |
+
"learning_rate": 8.232044198895029e-06,
|
| 51 |
+
"loss": 4.2777,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9668508287292817,
|
| 56 |
+
"grad_norm": 13.824647903442383,
|
| 57 |
+
"learning_rate": 9.613259668508288e-06,
|
| 58 |
+
"loss": 4.2007,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.4381443298969072,
|
| 64 |
+
"eval_loss": 1.0333575010299683,
|
| 65 |
+
"eval_macro_f1": 0.3744094254510921,
|
| 66 |
+
"eval_runtime": 0.42,
|
| 67 |
+
"eval_samples_per_second": 1385.648,
|
| 68 |
+
"eval_steps_per_second": 88.091,
|
| 69 |
+
"eval_weighted_f1": 0.3812797230054532,
|
| 70 |
+
"step": 181
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.1049723756906078,
|
| 74 |
+
"grad_norm": 14.590435028076172,
|
| 75 |
+
"learning_rate": 1.0994475138121547e-05,
|
| 76 |
+
"loss": 3.9303,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.2430939226519337,
|
| 81 |
+
"grad_norm": 19.318132400512695,
|
| 82 |
+
"learning_rate": 1.2375690607734806e-05,
|
| 83 |
+
"loss": 3.5442,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.3812154696132597,
|
| 88 |
+
"grad_norm": 25.510618209838867,
|
| 89 |
+
"learning_rate": 1.3756906077348067e-05,
|
| 90 |
+
"loss": 3.2065,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.5193370165745856,
|
| 95 |
+
"grad_norm": 32.776275634765625,
|
| 96 |
+
"learning_rate": 1.5138121546961326e-05,
|
| 97 |
+
"loss": 2.7508,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.6574585635359116,
|
| 102 |
+
"grad_norm": 35.182769775390625,
|
| 103 |
+
"learning_rate": 1.6519337016574586e-05,
|
| 104 |
+
"loss": 2.4866,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7955801104972375,
|
| 109 |
+
"grad_norm": 29.76224708557129,
|
| 110 |
+
"learning_rate": 1.7900552486187846e-05,
|
| 111 |
+
"loss": 2.4372,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.9337016574585635,
|
| 116 |
+
"grad_norm": 65.1946792602539,
|
| 117 |
+
"learning_rate": 1.9281767955801107e-05,
|
| 118 |
+
"loss": 2.2584,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.8195876288659794,
|
| 124 |
+
"eval_loss": 0.5012013912200928,
|
| 125 |
+
"eval_macro_f1": 0.8172247387275364,
|
| 126 |
+
"eval_runtime": 0.4178,
|
| 127 |
+
"eval_samples_per_second": 1393.177,
|
| 128 |
+
"eval_steps_per_second": 88.57,
|
| 129 |
+
"eval_weighted_f1": 0.8180222984651606,
|
| 130 |
+
"step": 362
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.0718232044198897,
|
| 134 |
+
"grad_norm": 59.23931884765625,
|
| 135 |
+
"learning_rate": 1.9926335174953962e-05,
|
| 136 |
+
"loss": 1.8542,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.2099447513812156,
|
| 141 |
+
"grad_norm": 35.68587875366211,
|
| 142 |
+
"learning_rate": 1.9772866789441377e-05,
|
| 143 |
+
"loss": 1.6,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.3480662983425415,
|
| 148 |
+
"grad_norm": 65.08133697509766,
|
| 149 |
+
"learning_rate": 1.9619398403928794e-05,
|
| 150 |
+
"loss": 1.5752,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.4861878453038675,
|
| 155 |
+
"grad_norm": 61.21116256713867,
|
| 156 |
+
"learning_rate": 1.946593001841621e-05,
|
| 157 |
+
"loss": 1.5893,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.6243093922651934,
|
| 162 |
+
"grad_norm": 53.618019104003906,
|
| 163 |
+
"learning_rate": 1.9312461632903623e-05,
|
| 164 |
+
"loss": 1.492,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.7624309392265194,
|
| 169 |
+
"grad_norm": 38.58893585205078,
|
| 170 |
+
"learning_rate": 1.915899324739104e-05,
|
| 171 |
+
"loss": 1.6762,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.9005524861878453,
|
| 176 |
+
"grad_norm": 33.73582458496094,
|
| 177 |
+
"learning_rate": 1.9005524861878455e-05,
|
| 178 |
+
"loss": 1.682,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 3.0,
|
| 183 |
+
"eval_accuracy": 0.8333333333333334,
|
| 184 |
+
"eval_loss": 0.4494641423225403,
|
| 185 |
+
"eval_macro_f1": 0.8326897917378258,
|
| 186 |
+
"eval_runtime": 0.4043,
|
| 187 |
+
"eval_samples_per_second": 1439.513,
|
| 188 |
+
"eval_steps_per_second": 91.515,
|
| 189 |
+
"eval_weighted_f1": 0.8334230690325918,
|
| 190 |
+
"step": 543
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.0386740331491713,
|
| 194 |
+
"grad_norm": 58.56112289428711,
|
| 195 |
+
"learning_rate": 1.885205647636587e-05,
|
| 196 |
+
"loss": 1.3738,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.1767955801104972,
|
| 201 |
+
"grad_norm": 61.42835235595703,
|
| 202 |
+
"learning_rate": 1.8698588090853287e-05,
|
| 203 |
+
"loss": 1.0249,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.314917127071823,
|
| 208 |
+
"grad_norm": 34.996360778808594,
|
| 209 |
+
"learning_rate": 1.85451197053407e-05,
|
| 210 |
+
"loss": 1.0205,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.453038674033149,
|
| 215 |
+
"grad_norm": 94.53072357177734,
|
| 216 |
+
"learning_rate": 1.8391651319828116e-05,
|
| 217 |
+
"loss": 1.1667,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.591160220994475,
|
| 222 |
+
"grad_norm": 53.063297271728516,
|
| 223 |
+
"learning_rate": 1.823818293431553e-05,
|
| 224 |
+
"loss": 1.1452,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.729281767955801,
|
| 229 |
+
"grad_norm": 38.83993911743164,
|
| 230 |
+
"learning_rate": 1.8084714548802948e-05,
|
| 231 |
+
"loss": 1.0453,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.867403314917127,
|
| 236 |
+
"grad_norm": 100.19972229003906,
|
| 237 |
+
"learning_rate": 1.7931246163290362e-05,
|
| 238 |
+
"loss": 1.1064,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 4.0,
|
| 243 |
+
"eval_accuracy": 0.8264604810996563,
|
| 244 |
+
"eval_loss": 0.5307199954986572,
|
| 245 |
+
"eval_macro_f1": 0.8254789152183964,
|
| 246 |
+
"eval_runtime": 0.4258,
|
| 247 |
+
"eval_samples_per_second": 1366.706,
|
| 248 |
+
"eval_steps_per_second": 86.887,
|
| 249 |
+
"eval_weighted_f1": 0.826051444447233,
|
| 250 |
+
"step": 724
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 4.005524861878453,
|
| 254 |
+
"grad_norm": 39.464176177978516,
|
| 255 |
+
"learning_rate": 1.7777777777777777e-05,
|
| 256 |
+
"loss": 1.13,
|
| 257 |
+
"step": 725
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.143646408839779,
|
| 261 |
+
"grad_norm": 38.51054763793945,
|
| 262 |
+
"learning_rate": 1.7624309392265194e-05,
|
| 263 |
+
"loss": 0.8192,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.281767955801105,
|
| 268 |
+
"grad_norm": 17.538677215576172,
|
| 269 |
+
"learning_rate": 1.747084100675261e-05,
|
| 270 |
+
"loss": 0.6238,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.419889502762431,
|
| 275 |
+
"grad_norm": 76.97430419921875,
|
| 276 |
+
"learning_rate": 1.7317372621240026e-05,
|
| 277 |
+
"loss": 0.7208,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.558011049723757,
|
| 282 |
+
"grad_norm": 28.344594955444336,
|
| 283 |
+
"learning_rate": 1.716390423572744e-05,
|
| 284 |
+
"loss": 0.6024,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.696132596685083,
|
| 289 |
+
"grad_norm": 2.690808057785034,
|
| 290 |
+
"learning_rate": 1.701043585021486e-05,
|
| 291 |
+
"loss": 0.6563,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.834254143646409,
|
| 296 |
+
"grad_norm": 11.658122062683105,
|
| 297 |
+
"learning_rate": 1.6856967464702273e-05,
|
| 298 |
+
"loss": 0.7523,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.972375690607735,
|
| 303 |
+
"grad_norm": 19.206592559814453,
|
| 304 |
+
"learning_rate": 1.670349907918969e-05,
|
| 305 |
+
"loss": 0.7208,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 5.0,
|
| 310 |
+
"eval_accuracy": 0.8350515463917526,
|
| 311 |
+
"eval_loss": 0.6184484958648682,
|
| 312 |
+
"eval_macro_f1": 0.8337704818761008,
|
| 313 |
+
"eval_runtime": 0.4141,
|
| 314 |
+
"eval_samples_per_second": 1405.5,
|
| 315 |
+
"eval_steps_per_second": 89.353,
|
| 316 |
+
"eval_weighted_f1": 0.8344647651968053,
|
| 317 |
+
"step": 905
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.110497237569061,
|
| 321 |
+
"grad_norm": 10.873888969421387,
|
| 322 |
+
"learning_rate": 1.6550030693677105e-05,
|
| 323 |
+
"loss": 0.5289,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.248618784530387,
|
| 328 |
+
"grad_norm": 13.146443367004395,
|
| 329 |
+
"learning_rate": 1.639656230816452e-05,
|
| 330 |
+
"loss": 0.4617,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.386740331491713,
|
| 335 |
+
"grad_norm": 42.1235237121582,
|
| 336 |
+
"learning_rate": 1.6243093922651937e-05,
|
| 337 |
+
"loss": 0.6367,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.524861878453039,
|
| 342 |
+
"grad_norm": 68.03278350830078,
|
| 343 |
+
"learning_rate": 1.608962553713935e-05,
|
| 344 |
+
"loss": 0.6179,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.662983425414365,
|
| 349 |
+
"grad_norm": 2.3678650856018066,
|
| 350 |
+
"learning_rate": 1.5936157151626766e-05,
|
| 351 |
+
"loss": 0.456,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.801104972375691,
|
| 356 |
+
"grad_norm": 74.92092895507812,
|
| 357 |
+
"learning_rate": 1.5782688766114183e-05,
|
| 358 |
+
"loss": 0.61,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.939226519337017,
|
| 363 |
+
"grad_norm": 26.87939453125,
|
| 364 |
+
"learning_rate": 1.5629220380601598e-05,
|
| 365 |
+
"loss": 0.4892,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 6.0,
|
| 370 |
+
"eval_accuracy": 0.8333333333333334,
|
| 371 |
+
"eval_loss": 0.7233725190162659,
|
| 372 |
+
"eval_macro_f1": 0.8302068474313072,
|
| 373 |
+
"eval_runtime": 0.4122,
|
| 374 |
+
"eval_samples_per_second": 1411.821,
|
| 375 |
+
"eval_steps_per_second": 89.755,
|
| 376 |
+
"eval_weighted_f1": 0.8309938455164565,
|
| 377 |
+
"step": 1086
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 6.077348066298343,
|
| 381 |
+
"grad_norm": 7.154663562774658,
|
| 382 |
+
"learning_rate": 1.5475751995089012e-05,
|
| 383 |
+
"loss": 0.4874,
|
| 384 |
+
"step": 1100
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 6.2154696132596685,
|
| 388 |
+
"grad_norm": 3.5838303565979004,
|
| 389 |
+
"learning_rate": 1.532228360957643e-05,
|
| 390 |
+
"loss": 0.6082,
|
| 391 |
+
"step": 1125
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 6.3535911602209945,
|
| 395 |
+
"grad_norm": 6.32533597946167,
|
| 396 |
+
"learning_rate": 1.5168815224063844e-05,
|
| 397 |
+
"loss": 0.4121,
|
| 398 |
+
"step": 1150
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.49171270718232,
|
| 402 |
+
"grad_norm": 9.479781150817871,
|
| 403 |
+
"learning_rate": 1.5015346838551258e-05,
|
| 404 |
+
"loss": 0.3843,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.629834254143646,
|
| 409 |
+
"grad_norm": 11.685283660888672,
|
| 410 |
+
"learning_rate": 1.4861878453038676e-05,
|
| 411 |
+
"loss": 0.3472,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.767955801104972,
|
| 416 |
+
"grad_norm": 18.950719833374023,
|
| 417 |
+
"learning_rate": 1.470841006752609e-05,
|
| 418 |
+
"loss": 0.4416,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.906077348066298,
|
| 423 |
+
"grad_norm": 108.25150299072266,
|
| 424 |
+
"learning_rate": 1.4554941682013506e-05,
|
| 425 |
+
"loss": 0.7049,
|
| 426 |
+
"step": 1250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 7.0,
|
| 430 |
+
"eval_accuracy": 0.8298969072164949,
|
| 431 |
+
"eval_loss": 0.7750498056411743,
|
| 432 |
+
"eval_macro_f1": 0.8290962019215967,
|
| 433 |
+
"eval_runtime": 0.4163,
|
| 434 |
+
"eval_samples_per_second": 1397.908,
|
| 435 |
+
"eval_steps_per_second": 88.87,
|
| 436 |
+
"eval_weighted_f1": 0.8298012186054777,
|
| 437 |
+
"step": 1267
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 7.044198895027624,
|
| 441 |
+
"grad_norm": 74.38285064697266,
|
| 442 |
+
"learning_rate": 1.4401473296500922e-05,
|
| 443 |
+
"loss": 0.2615,
|
| 444 |
+
"step": 1275
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 7.18232044198895,
|
| 448 |
+
"grad_norm": 0.2990298867225647,
|
| 449 |
+
"learning_rate": 1.4248004910988338e-05,
|
| 450 |
+
"loss": 0.154,
|
| 451 |
+
"step": 1300
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 7.320441988950276,
|
| 455 |
+
"grad_norm": 0.29604440927505493,
|
| 456 |
+
"learning_rate": 1.4094536525475753e-05,
|
| 457 |
+
"loss": 0.3229,
|
| 458 |
+
"step": 1325
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 7.458563535911602,
|
| 462 |
+
"grad_norm": 0.8504543900489807,
|
| 463 |
+
"learning_rate": 1.394106813996317e-05,
|
| 464 |
+
"loss": 0.2784,
|
| 465 |
+
"step": 1350
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 7.596685082872928,
|
| 469 |
+
"grad_norm": 63.97197341918945,
|
| 470 |
+
"learning_rate": 1.3787599754450585e-05,
|
| 471 |
+
"loss": 0.2941,
|
| 472 |
+
"step": 1375
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 7.734806629834254,
|
| 476 |
+
"grad_norm": 76.94757843017578,
|
| 477 |
+
"learning_rate": 1.3634131368937999e-05,
|
| 478 |
+
"loss": 0.2114,
|
| 479 |
+
"step": 1400
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 7.87292817679558,
|
| 483 |
+
"grad_norm": 0.32140347361564636,
|
| 484 |
+
"learning_rate": 1.3480662983425417e-05,
|
| 485 |
+
"loss": 0.3671,
|
| 486 |
+
"step": 1425
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 8.0,
|
| 490 |
+
"eval_accuracy": 0.8298969072164949,
|
| 491 |
+
"eval_loss": 0.8607999086380005,
|
| 492 |
+
"eval_macro_f1": 0.827746023684425,
|
| 493 |
+
"eval_runtime": 0.3919,
|
| 494 |
+
"eval_samples_per_second": 1484.951,
|
| 495 |
+
"eval_steps_per_second": 94.404,
|
| 496 |
+
"eval_weighted_f1": 0.8287579905371756,
|
| 497 |
+
"step": 1448
|
| 498 |
+
}
|
| 499 |
+
],
|
| 500 |
+
"logging_steps": 25,
|
| 501 |
+
"max_steps": 3620,
|
| 502 |
+
"num_input_tokens_seen": 0,
|
| 503 |
+
"num_train_epochs": 20,
|
| 504 |
+
"save_steps": 500,
|
| 505 |
+
"stateful_callbacks": {
|
| 506 |
+
"EarlyStoppingCallback": {
|
| 507 |
+
"args": {
|
| 508 |
+
"early_stopping_patience": 3,
|
| 509 |
+
"early_stopping_threshold": 0.0
|
| 510 |
+
},
|
| 511 |
+
"attributes": {
|
| 512 |
+
"early_stopping_patience_counter": 3
|
| 513 |
+
}
|
| 514 |
+
},
|
| 515 |
+
"TrainerControl": {
|
| 516 |
+
"args": {
|
| 517 |
+
"should_epoch_stop": false,
|
| 518 |
+
"should_evaluate": false,
|
| 519 |
+
"should_log": false,
|
| 520 |
+
"should_save": true,
|
| 521 |
+
"should_training_stop": true
|
| 522 |
+
},
|
| 523 |
+
"attributes": {}
|
| 524 |
+
}
|
| 525 |
+
},
|
| 526 |
+
"total_flos": 0.0,
|
| 527 |
+
"train_batch_size": 8,
|
| 528 |
+
"trial_name": null,
|
| 529 |
+
"trial_params": null
|
| 530 |
+
}
|
ood-vifactcheck-phobert-source-source-02-s43-20260826T154150Z-bdd2c308/trainer/checkpoint-1448/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-source-source-02-s43-20260826T154150Z-bdd2c308/trainer/checkpoint-905/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-source-source-02-s43-20260826T154150Z-bdd2c308/trainer/checkpoint-905/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-source-source-02-s43-20260826T154150Z-bdd2c308/trainer/checkpoint-905/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-source-source-02-s43-20260826T154150Z-bdd2c308/trainer/checkpoint-905/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-source-source-02-s43-20260826T154150Z-bdd2c308/trainer/checkpoint-905/trainer_state.json
ADDED
|
@@ -0,0 +1,350 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 905,
|
| 3 |
+
"best_metric": 0.8337704818761008,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-source-source-02-s43-20260826T154150Z-bdd2c308/trainer/checkpoint-905",
|
| 5 |
+
"epoch": 5.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 905,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.13812154696132597,
|
| 14 |
+
"grad_norm": 6.335784912109375,
|
| 15 |
+
"learning_rate": 1.3259668508287293e-06,
|
| 16 |
+
"loss": 4.4522,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.27624309392265195,
|
| 21 |
+
"grad_norm": 6.684322834014893,
|
| 22 |
+
"learning_rate": 2.707182320441989e-06,
|
| 23 |
+
"loss": 4.4467,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.4143646408839779,
|
| 28 |
+
"grad_norm": 7.152876377105713,
|
| 29 |
+
"learning_rate": 4.088397790055249e-06,
|
| 30 |
+
"loss": 4.4027,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5524861878453039,
|
| 35 |
+
"grad_norm": 8.315168380737305,
|
| 36 |
+
"learning_rate": 5.469613259668509e-06,
|
| 37 |
+
"loss": 4.3901,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6906077348066298,
|
| 42 |
+
"grad_norm": 7.618349552154541,
|
| 43 |
+
"learning_rate": 6.850828729281769e-06,
|
| 44 |
+
"loss": 4.3232,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8287292817679558,
|
| 49 |
+
"grad_norm": 9.473018646240234,
|
| 50 |
+
"learning_rate": 8.232044198895029e-06,
|
| 51 |
+
"loss": 4.2777,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9668508287292817,
|
| 56 |
+
"grad_norm": 13.824647903442383,
|
| 57 |
+
"learning_rate": 9.613259668508288e-06,
|
| 58 |
+
"loss": 4.2007,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.4381443298969072,
|
| 64 |
+
"eval_loss": 1.0333575010299683,
|
| 65 |
+
"eval_macro_f1": 0.3744094254510921,
|
| 66 |
+
"eval_runtime": 0.42,
|
| 67 |
+
"eval_samples_per_second": 1385.648,
|
| 68 |
+
"eval_steps_per_second": 88.091,
|
| 69 |
+
"eval_weighted_f1": 0.3812797230054532,
|
| 70 |
+
"step": 181
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.1049723756906078,
|
| 74 |
+
"grad_norm": 14.590435028076172,
|
| 75 |
+
"learning_rate": 1.0994475138121547e-05,
|
| 76 |
+
"loss": 3.9303,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.2430939226519337,
|
| 81 |
+
"grad_norm": 19.318132400512695,
|
| 82 |
+
"learning_rate": 1.2375690607734806e-05,
|
| 83 |
+
"loss": 3.5442,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.3812154696132597,
|
| 88 |
+
"grad_norm": 25.510618209838867,
|
| 89 |
+
"learning_rate": 1.3756906077348067e-05,
|
| 90 |
+
"loss": 3.2065,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.5193370165745856,
|
| 95 |
+
"grad_norm": 32.776275634765625,
|
| 96 |
+
"learning_rate": 1.5138121546961326e-05,
|
| 97 |
+
"loss": 2.7508,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.6574585635359116,
|
| 102 |
+
"grad_norm": 35.182769775390625,
|
| 103 |
+
"learning_rate": 1.6519337016574586e-05,
|
| 104 |
+
"loss": 2.4866,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7955801104972375,
|
| 109 |
+
"grad_norm": 29.76224708557129,
|
| 110 |
+
"learning_rate": 1.7900552486187846e-05,
|
| 111 |
+
"loss": 2.4372,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.9337016574585635,
|
| 116 |
+
"grad_norm": 65.1946792602539,
|
| 117 |
+
"learning_rate": 1.9281767955801107e-05,
|
| 118 |
+
"loss": 2.2584,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.8195876288659794,
|
| 124 |
+
"eval_loss": 0.5012013912200928,
|
| 125 |
+
"eval_macro_f1": 0.8172247387275364,
|
| 126 |
+
"eval_runtime": 0.4178,
|
| 127 |
+
"eval_samples_per_second": 1393.177,
|
| 128 |
+
"eval_steps_per_second": 88.57,
|
| 129 |
+
"eval_weighted_f1": 0.8180222984651606,
|
| 130 |
+
"step": 362
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.0718232044198897,
|
| 134 |
+
"grad_norm": 59.23931884765625,
|
| 135 |
+
"learning_rate": 1.9926335174953962e-05,
|
| 136 |
+
"loss": 1.8542,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.2099447513812156,
|
| 141 |
+
"grad_norm": 35.68587875366211,
|
| 142 |
+
"learning_rate": 1.9772866789441377e-05,
|
| 143 |
+
"loss": 1.6,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.3480662983425415,
|
| 148 |
+
"grad_norm": 65.08133697509766,
|
| 149 |
+
"learning_rate": 1.9619398403928794e-05,
|
| 150 |
+
"loss": 1.5752,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.4861878453038675,
|
| 155 |
+
"grad_norm": 61.21116256713867,
|
| 156 |
+
"learning_rate": 1.946593001841621e-05,
|
| 157 |
+
"loss": 1.5893,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.6243093922651934,
|
| 162 |
+
"grad_norm": 53.618019104003906,
|
| 163 |
+
"learning_rate": 1.9312461632903623e-05,
|
| 164 |
+
"loss": 1.492,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.7624309392265194,
|
| 169 |
+
"grad_norm": 38.58893585205078,
|
| 170 |
+
"learning_rate": 1.915899324739104e-05,
|
| 171 |
+
"loss": 1.6762,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.9005524861878453,
|
| 176 |
+
"grad_norm": 33.73582458496094,
|
| 177 |
+
"learning_rate": 1.9005524861878455e-05,
|
| 178 |
+
"loss": 1.682,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 3.0,
|
| 183 |
+
"eval_accuracy": 0.8333333333333334,
|
| 184 |
+
"eval_loss": 0.4494641423225403,
|
| 185 |
+
"eval_macro_f1": 0.8326897917378258,
|
| 186 |
+
"eval_runtime": 0.4043,
|
| 187 |
+
"eval_samples_per_second": 1439.513,
|
| 188 |
+
"eval_steps_per_second": 91.515,
|
| 189 |
+
"eval_weighted_f1": 0.8334230690325918,
|
| 190 |
+
"step": 543
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.0386740331491713,
|
| 194 |
+
"grad_norm": 58.56112289428711,
|
| 195 |
+
"learning_rate": 1.885205647636587e-05,
|
| 196 |
+
"loss": 1.3738,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.1767955801104972,
|
| 201 |
+
"grad_norm": 61.42835235595703,
|
| 202 |
+
"learning_rate": 1.8698588090853287e-05,
|
| 203 |
+
"loss": 1.0249,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.314917127071823,
|
| 208 |
+
"grad_norm": 34.996360778808594,
|
| 209 |
+
"learning_rate": 1.85451197053407e-05,
|
| 210 |
+
"loss": 1.0205,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.453038674033149,
|
| 215 |
+
"grad_norm": 94.53072357177734,
|
| 216 |
+
"learning_rate": 1.8391651319828116e-05,
|
| 217 |
+
"loss": 1.1667,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.591160220994475,
|
| 222 |
+
"grad_norm": 53.063297271728516,
|
| 223 |
+
"learning_rate": 1.823818293431553e-05,
|
| 224 |
+
"loss": 1.1452,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.729281767955801,
|
| 229 |
+
"grad_norm": 38.83993911743164,
|
| 230 |
+
"learning_rate": 1.8084714548802948e-05,
|
| 231 |
+
"loss": 1.0453,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.867403314917127,
|
| 236 |
+
"grad_norm": 100.19972229003906,
|
| 237 |
+
"learning_rate": 1.7931246163290362e-05,
|
| 238 |
+
"loss": 1.1064,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 4.0,
|
| 243 |
+
"eval_accuracy": 0.8264604810996563,
|
| 244 |
+
"eval_loss": 0.5307199954986572,
|
| 245 |
+
"eval_macro_f1": 0.8254789152183964,
|
| 246 |
+
"eval_runtime": 0.4258,
|
| 247 |
+
"eval_samples_per_second": 1366.706,
|
| 248 |
+
"eval_steps_per_second": 86.887,
|
| 249 |
+
"eval_weighted_f1": 0.826051444447233,
|
| 250 |
+
"step": 724
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 4.005524861878453,
|
| 254 |
+
"grad_norm": 39.464176177978516,
|
| 255 |
+
"learning_rate": 1.7777777777777777e-05,
|
| 256 |
+
"loss": 1.13,
|
| 257 |
+
"step": 725
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.143646408839779,
|
| 261 |
+
"grad_norm": 38.51054763793945,
|
| 262 |
+
"learning_rate": 1.7624309392265194e-05,
|
| 263 |
+
"loss": 0.8192,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.281767955801105,
|
| 268 |
+
"grad_norm": 17.538677215576172,
|
| 269 |
+
"learning_rate": 1.747084100675261e-05,
|
| 270 |
+
"loss": 0.6238,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.419889502762431,
|
| 275 |
+
"grad_norm": 76.97430419921875,
|
| 276 |
+
"learning_rate": 1.7317372621240026e-05,
|
| 277 |
+
"loss": 0.7208,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.558011049723757,
|
| 282 |
+
"grad_norm": 28.344594955444336,
|
| 283 |
+
"learning_rate": 1.716390423572744e-05,
|
| 284 |
+
"loss": 0.6024,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.696132596685083,
|
| 289 |
+
"grad_norm": 2.690808057785034,
|
| 290 |
+
"learning_rate": 1.701043585021486e-05,
|
| 291 |
+
"loss": 0.6563,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.834254143646409,
|
| 296 |
+
"grad_norm": 11.658122062683105,
|
| 297 |
+
"learning_rate": 1.6856967464702273e-05,
|
| 298 |
+
"loss": 0.7523,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.972375690607735,
|
| 303 |
+
"grad_norm": 19.206592559814453,
|
| 304 |
+
"learning_rate": 1.670349907918969e-05,
|
| 305 |
+
"loss": 0.7208,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 5.0,
|
| 310 |
+
"eval_accuracy": 0.8350515463917526,
|
| 311 |
+
"eval_loss": 0.6184484958648682,
|
| 312 |
+
"eval_macro_f1": 0.8337704818761008,
|
| 313 |
+
"eval_runtime": 0.4141,
|
| 314 |
+
"eval_samples_per_second": 1405.5,
|
| 315 |
+
"eval_steps_per_second": 89.353,
|
| 316 |
+
"eval_weighted_f1": 0.8344647651968053,
|
| 317 |
+
"step": 905
|
| 318 |
+
}
|
| 319 |
+
],
|
| 320 |
+
"logging_steps": 25,
|
| 321 |
+
"max_steps": 3620,
|
| 322 |
+
"num_input_tokens_seen": 0,
|
| 323 |
+
"num_train_epochs": 20,
|
| 324 |
+
"save_steps": 500,
|
| 325 |
+
"stateful_callbacks": {
|
| 326 |
+
"EarlyStoppingCallback": {
|
| 327 |
+
"args": {
|
| 328 |
+
"early_stopping_patience": 3,
|
| 329 |
+
"early_stopping_threshold": 0.0
|
| 330 |
+
},
|
| 331 |
+
"attributes": {
|
| 332 |
+
"early_stopping_patience_counter": 0
|
| 333 |
+
}
|
| 334 |
+
},
|
| 335 |
+
"TrainerControl": {
|
| 336 |
+
"args": {
|
| 337 |
+
"should_epoch_stop": false,
|
| 338 |
+
"should_evaluate": false,
|
| 339 |
+
"should_log": false,
|
| 340 |
+
"should_save": true,
|
| 341 |
+
"should_training_stop": false
|
| 342 |
+
},
|
| 343 |
+
"attributes": {}
|
| 344 |
+
}
|
| 345 |
+
},
|
| 346 |
+
"total_flos": 0.0,
|
| 347 |
+
"train_batch_size": 8,
|
| 348 |
+
"trial_name": null,
|
| 349 |
+
"trial_params": null
|
| 350 |
+
}
|
ood-vifactcheck-phobert-source-source-02-s43-20260826T154150Z-bdd2c308/trainer/checkpoint-905/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-source-source-03-s43-20260826T154435Z-bdd2c308/trainer/checkpoint-1464/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-source-source-03-s43-20260826T154435Z-bdd2c308/trainer/checkpoint-1464/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-source-source-03-s43-20260826T154435Z-bdd2c308/trainer/checkpoint-1464/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-source-source-03-s43-20260826T154435Z-bdd2c308/trainer/checkpoint-1464/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-source-source-03-s43-20260826T154435Z-bdd2c308/trainer/checkpoint-1464/trainer_state.json
ADDED
|
@@ -0,0 +1,537 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 915,
|
| 3 |
+
"best_metric": 0.853797945488543,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-source-source-03-s43-20260826T154435Z-bdd2c308/trainer/checkpoint-915",
|
| 5 |
+
"epoch": 8.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1464,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.13717421124828533,
|
| 14 |
+
"grad_norm": 6.4076714515686035,
|
| 15 |
+
"learning_rate": 1.3114754098360657e-06,
|
| 16 |
+
"loss": 4.4479,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.27434842249657065,
|
| 21 |
+
"grad_norm": 6.469189643859863,
|
| 22 |
+
"learning_rate": 2.677595628415301e-06,
|
| 23 |
+
"loss": 4.4062,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.411522633744856,
|
| 28 |
+
"grad_norm": 9.004118919372559,
|
| 29 |
+
"learning_rate": 4.043715846994536e-06,
|
| 30 |
+
"loss": 4.4233,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5486968449931413,
|
| 35 |
+
"grad_norm": 10.774386405944824,
|
| 36 |
+
"learning_rate": 5.409836065573772e-06,
|
| 37 |
+
"loss": 4.3641,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6858710562414266,
|
| 42 |
+
"grad_norm": 7.2467169761657715,
|
| 43 |
+
"learning_rate": 6.775956284153006e-06,
|
| 44 |
+
"loss": 4.3555,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.823045267489712,
|
| 49 |
+
"grad_norm": 8.021151542663574,
|
| 50 |
+
"learning_rate": 8.14207650273224e-06,
|
| 51 |
+
"loss": 4.3166,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9602194787379973,
|
| 56 |
+
"grad_norm": 8.692425727844238,
|
| 57 |
+
"learning_rate": 9.508196721311476e-06,
|
| 58 |
+
"loss": 4.256,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.5440931780366056,
|
| 64 |
+
"eval_loss": 0.9838680028915405,
|
| 65 |
+
"eval_macro_f1": 0.5369983375409565,
|
| 66 |
+
"eval_runtime": 0.4171,
|
| 67 |
+
"eval_samples_per_second": 1440.887,
|
| 68 |
+
"eval_steps_per_second": 91.104,
|
| 69 |
+
"eval_weighted_f1": 0.5389635726736528,
|
| 70 |
+
"step": 183
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.093278463648834,
|
| 74 |
+
"grad_norm": 12.057026863098145,
|
| 75 |
+
"learning_rate": 1.0874316939890712e-05,
|
| 76 |
+
"loss": 3.8209,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.2304526748971194,
|
| 81 |
+
"grad_norm": 23.84331512451172,
|
| 82 |
+
"learning_rate": 1.2240437158469946e-05,
|
| 83 |
+
"loss": 3.552,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.3676268861454046,
|
| 88 |
+
"grad_norm": 34.658504486083984,
|
| 89 |
+
"learning_rate": 1.3606557377049181e-05,
|
| 90 |
+
"loss": 3.2955,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.50480109739369,
|
| 95 |
+
"grad_norm": 35.16084289550781,
|
| 96 |
+
"learning_rate": 1.4972677595628417e-05,
|
| 97 |
+
"loss": 2.9104,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.6419753086419753,
|
| 102 |
+
"grad_norm": 41.14034652709961,
|
| 103 |
+
"learning_rate": 1.633879781420765e-05,
|
| 104 |
+
"loss": 2.6576,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7791495198902605,
|
| 109 |
+
"grad_norm": 48.54668045043945,
|
| 110 |
+
"learning_rate": 1.7704918032786887e-05,
|
| 111 |
+
"loss": 2.6274,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.916323731138546,
|
| 116 |
+
"grad_norm": 51.720664978027344,
|
| 117 |
+
"learning_rate": 1.9071038251366122e-05,
|
| 118 |
+
"loss": 2.349,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.8419301164725458,
|
| 124 |
+
"eval_loss": 0.4657325744628906,
|
| 125 |
+
"eval_macro_f1": 0.8420572992124283,
|
| 126 |
+
"eval_runtime": 0.3881,
|
| 127 |
+
"eval_samples_per_second": 1548.683,
|
| 128 |
+
"eval_steps_per_second": 97.92,
|
| 129 |
+
"eval_weighted_f1": 0.841929537922293,
|
| 130 |
+
"step": 366
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.049382716049383,
|
| 134 |
+
"grad_norm": 33.88608932495117,
|
| 135 |
+
"learning_rate": 1.995142683667274e-05,
|
| 136 |
+
"loss": 1.9099,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.186556927297668,
|
| 141 |
+
"grad_norm": 38.441776275634766,
|
| 142 |
+
"learning_rate": 1.9799635701275045e-05,
|
| 143 |
+
"loss": 1.8178,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.3237311385459534,
|
| 148 |
+
"grad_norm": 23.70911407470703,
|
| 149 |
+
"learning_rate": 1.9647844565877354e-05,
|
| 150 |
+
"loss": 1.5067,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.460905349794239,
|
| 155 |
+
"grad_norm": 24.64794921875,
|
| 156 |
+
"learning_rate": 1.9496053430479663e-05,
|
| 157 |
+
"loss": 1.4754,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.598079561042524,
|
| 162 |
+
"grad_norm": 53.598976135253906,
|
| 163 |
+
"learning_rate": 1.934426229508197e-05,
|
| 164 |
+
"loss": 1.6642,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.7352537722908092,
|
| 169 |
+
"grad_norm": 28.306968688964844,
|
| 170 |
+
"learning_rate": 1.9192471159684274e-05,
|
| 171 |
+
"loss": 1.7823,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.8724279835390947,
|
| 176 |
+
"grad_norm": 58.7725715637207,
|
| 177 |
+
"learning_rate": 1.9040680024286583e-05,
|
| 178 |
+
"loss": 1.7782,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 3.0,
|
| 183 |
+
"eval_accuracy": 0.8535773710482529,
|
| 184 |
+
"eval_loss": 0.43532466888427734,
|
| 185 |
+
"eval_macro_f1": 0.8534562192837053,
|
| 186 |
+
"eval_runtime": 0.5288,
|
| 187 |
+
"eval_samples_per_second": 1136.623,
|
| 188 |
+
"eval_steps_per_second": 71.866,
|
| 189 |
+
"eval_weighted_f1": 0.8528918706451937,
|
| 190 |
+
"step": 549
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.0054869684499312,
|
| 194 |
+
"grad_norm": 98.39491271972656,
|
| 195 |
+
"learning_rate": 1.888888888888889e-05,
|
| 196 |
+
"loss": 1.5376,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.1426611796982167,
|
| 201 |
+
"grad_norm": 56.169395446777344,
|
| 202 |
+
"learning_rate": 1.8737097753491198e-05,
|
| 203 |
+
"loss": 1.1509,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.279835390946502,
|
| 208 |
+
"grad_norm": 100.4572525024414,
|
| 209 |
+
"learning_rate": 1.8585306618093504e-05,
|
| 210 |
+
"loss": 1.1175,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.4170096021947876,
|
| 215 |
+
"grad_norm": 32.61418533325195,
|
| 216 |
+
"learning_rate": 1.8433515482695813e-05,
|
| 217 |
+
"loss": 1.1605,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.5541838134430725,
|
| 222 |
+
"grad_norm": 50.624610900878906,
|
| 223 |
+
"learning_rate": 1.828172434729812e-05,
|
| 224 |
+
"loss": 1.1586,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.691358024691358,
|
| 229 |
+
"grad_norm": 23.299829483032227,
|
| 230 |
+
"learning_rate": 1.8129933211900424e-05,
|
| 231 |
+
"loss": 1.1789,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.8285322359396434,
|
| 236 |
+
"grad_norm": 111.1823959350586,
|
| 237 |
+
"learning_rate": 1.7978142076502733e-05,
|
| 238 |
+
"loss": 0.9352,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.9657064471879284,
|
| 243 |
+
"grad_norm": 33.47234344482422,
|
| 244 |
+
"learning_rate": 1.7826350941105042e-05,
|
| 245 |
+
"loss": 1.0522,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 4.0,
|
| 250 |
+
"eval_accuracy": 0.848585690515807,
|
| 251 |
+
"eval_loss": 0.43581661581993103,
|
| 252 |
+
"eval_macro_f1": 0.848643240658458,
|
| 253 |
+
"eval_runtime": 0.4208,
|
| 254 |
+
"eval_samples_per_second": 1428.299,
|
| 255 |
+
"eval_steps_per_second": 90.308,
|
| 256 |
+
"eval_weighted_f1": 0.8482517042566979,
|
| 257 |
+
"step": 732
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.098765432098766,
|
| 261 |
+
"grad_norm": 26.348663330078125,
|
| 262 |
+
"learning_rate": 1.7674559805707348e-05,
|
| 263 |
+
"loss": 0.8097,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.23593964334705,
|
| 268 |
+
"grad_norm": 15.588164329528809,
|
| 269 |
+
"learning_rate": 1.7522768670309657e-05,
|
| 270 |
+
"loss": 0.7674,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.373113854595336,
|
| 275 |
+
"grad_norm": 9.365654945373535,
|
| 276 |
+
"learning_rate": 1.7370977534911962e-05,
|
| 277 |
+
"loss": 0.8267,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.510288065843621,
|
| 282 |
+
"grad_norm": 26.54583168029785,
|
| 283 |
+
"learning_rate": 1.7219186399514268e-05,
|
| 284 |
+
"loss": 0.8376,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.647462277091907,
|
| 289 |
+
"grad_norm": 44.178321838378906,
|
| 290 |
+
"learning_rate": 1.7067395264116577e-05,
|
| 291 |
+
"loss": 0.7836,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.784636488340192,
|
| 296 |
+
"grad_norm": 45.240657806396484,
|
| 297 |
+
"learning_rate": 1.6915604128718883e-05,
|
| 298 |
+
"loss": 0.7959,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.921810699588478,
|
| 303 |
+
"grad_norm": 17.089048385620117,
|
| 304 |
+
"learning_rate": 1.6763812993321192e-05,
|
| 305 |
+
"loss": 0.7625,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 5.0,
|
| 310 |
+
"eval_accuracy": 0.8535773710482529,
|
| 311 |
+
"eval_loss": 0.507513701915741,
|
| 312 |
+
"eval_macro_f1": 0.853797945488543,
|
| 313 |
+
"eval_runtime": 0.4211,
|
| 314 |
+
"eval_samples_per_second": 1427.139,
|
| 315 |
+
"eval_steps_per_second": 90.235,
|
| 316 |
+
"eval_weighted_f1": 0.8538355369232116,
|
| 317 |
+
"step": 915
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.054869684499314,
|
| 321 |
+
"grad_norm": 99.6161117553711,
|
| 322 |
+
"learning_rate": 1.66120218579235e-05,
|
| 323 |
+
"loss": 0.6111,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.1920438957476,
|
| 328 |
+
"grad_norm": 29.69748306274414,
|
| 329 |
+
"learning_rate": 1.6460230722525807e-05,
|
| 330 |
+
"loss": 0.6004,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.329218106995885,
|
| 335 |
+
"grad_norm": 10.094265937805176,
|
| 336 |
+
"learning_rate": 1.6308439587128112e-05,
|
| 337 |
+
"loss": 0.5867,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.4663923182441705,
|
| 342 |
+
"grad_norm": 56.10636520385742,
|
| 343 |
+
"learning_rate": 1.615664845173042e-05,
|
| 344 |
+
"loss": 0.6944,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.603566529492455,
|
| 349 |
+
"grad_norm": 60.17007827758789,
|
| 350 |
+
"learning_rate": 1.6004857316332727e-05,
|
| 351 |
+
"loss": 0.4378,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.7407407407407405,
|
| 356 |
+
"grad_norm": 248.58416748046875,
|
| 357 |
+
"learning_rate": 1.5853066180935036e-05,
|
| 358 |
+
"loss": 0.4382,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.877914951989026,
|
| 363 |
+
"grad_norm": 109.17561340332031,
|
| 364 |
+
"learning_rate": 1.570127504553734e-05,
|
| 365 |
+
"loss": 0.5545,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 6.0,
|
| 370 |
+
"eval_accuracy": 0.8352745424292846,
|
| 371 |
+
"eval_loss": 0.6536827683448792,
|
| 372 |
+
"eval_macro_f1": 0.8349438329901168,
|
| 373 |
+
"eval_runtime": 0.4294,
|
| 374 |
+
"eval_samples_per_second": 1399.773,
|
| 375 |
+
"eval_steps_per_second": 88.505,
|
| 376 |
+
"eval_weighted_f1": 0.8345025828805359,
|
| 377 |
+
"step": 1098
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 6.0109739368998625,
|
| 381 |
+
"grad_norm": 21.607284545898438,
|
| 382 |
+
"learning_rate": 1.5549483910139647e-05,
|
| 383 |
+
"loss": 0.6047,
|
| 384 |
+
"step": 1100
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 6.148148148148148,
|
| 388 |
+
"grad_norm": 8.471403121948242,
|
| 389 |
+
"learning_rate": 1.5397692774741956e-05,
|
| 390 |
+
"loss": 0.226,
|
| 391 |
+
"step": 1125
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 6.285322359396433,
|
| 395 |
+
"grad_norm": 122.19269561767578,
|
| 396 |
+
"learning_rate": 1.5245901639344264e-05,
|
| 397 |
+
"loss": 0.5928,
|
| 398 |
+
"step": 1150
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.422496570644719,
|
| 402 |
+
"grad_norm": 2.4058661460876465,
|
| 403 |
+
"learning_rate": 1.5094110503946571e-05,
|
| 404 |
+
"loss": 0.5368,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.559670781893004,
|
| 409 |
+
"grad_norm": 96.64785766601562,
|
| 410 |
+
"learning_rate": 1.4942319368548878e-05,
|
| 411 |
+
"loss": 0.5003,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.69684499314129,
|
| 416 |
+
"grad_norm": 66.92341613769531,
|
| 417 |
+
"learning_rate": 1.4790528233151184e-05,
|
| 418 |
+
"loss": 0.4638,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.834019204389575,
|
| 423 |
+
"grad_norm": 4.732495307922363,
|
| 424 |
+
"learning_rate": 1.4638737097753493e-05,
|
| 425 |
+
"loss": 0.4124,
|
| 426 |
+
"step": 1250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 6.97119341563786,
|
| 430 |
+
"grad_norm": 0.9478699564933777,
|
| 431 |
+
"learning_rate": 1.44869459623558e-05,
|
| 432 |
+
"loss": 0.3049,
|
| 433 |
+
"step": 1275
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 7.0,
|
| 437 |
+
"eval_accuracy": 0.848585690515807,
|
| 438 |
+
"eval_loss": 0.6374513506889343,
|
| 439 |
+
"eval_macro_f1": 0.8491974829662329,
|
| 440 |
+
"eval_runtime": 0.429,
|
| 441 |
+
"eval_samples_per_second": 1400.936,
|
| 442 |
+
"eval_steps_per_second": 88.578,
|
| 443 |
+
"eval_weighted_f1": 0.8486492313929089,
|
| 444 |
+
"step": 1281
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 7.104252400548697,
|
| 448 |
+
"grad_norm": 2.8059935569763184,
|
| 449 |
+
"learning_rate": 1.4335154826958106e-05,
|
| 450 |
+
"loss": 0.3915,
|
| 451 |
+
"step": 1300
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 7.2414266117969825,
|
| 455 |
+
"grad_norm": 31.428237915039062,
|
| 456 |
+
"learning_rate": 1.4183363691560415e-05,
|
| 457 |
+
"loss": 0.3007,
|
| 458 |
+
"step": 1325
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 7.378600823045267,
|
| 462 |
+
"grad_norm": 0.7204816341400146,
|
| 463 |
+
"learning_rate": 1.403157255616272e-05,
|
| 464 |
+
"loss": 0.2583,
|
| 465 |
+
"step": 1350
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 7.5157750342935525,
|
| 469 |
+
"grad_norm": 0.43895164132118225,
|
| 470 |
+
"learning_rate": 1.3879781420765028e-05,
|
| 471 |
+
"loss": 0.3497,
|
| 472 |
+
"step": 1375
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 7.652949245541838,
|
| 476 |
+
"grad_norm": 164.42591857910156,
|
| 477 |
+
"learning_rate": 1.3727990285367337e-05,
|
| 478 |
+
"loss": 0.414,
|
| 479 |
+
"step": 1400
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 7.790123456790123,
|
| 483 |
+
"grad_norm": 17.342838287353516,
|
| 484 |
+
"learning_rate": 1.3576199149969643e-05,
|
| 485 |
+
"loss": 0.3485,
|
| 486 |
+
"step": 1425
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 7.927297668038409,
|
| 490 |
+
"grad_norm": 0.37304502725601196,
|
| 491 |
+
"learning_rate": 1.342440801457195e-05,
|
| 492 |
+
"loss": 0.2061,
|
| 493 |
+
"step": 1450
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 8.0,
|
| 497 |
+
"eval_accuracy": 0.8386023294509152,
|
| 498 |
+
"eval_loss": 0.7873337268829346,
|
| 499 |
+
"eval_macro_f1": 0.8389631959724527,
|
| 500 |
+
"eval_runtime": 0.4185,
|
| 501 |
+
"eval_samples_per_second": 1436.064,
|
| 502 |
+
"eval_steps_per_second": 90.799,
|
| 503 |
+
"eval_weighted_f1": 0.8384356508017139,
|
| 504 |
+
"step": 1464
|
| 505 |
+
}
|
| 506 |
+
],
|
| 507 |
+
"logging_steps": 25,
|
| 508 |
+
"max_steps": 3660,
|
| 509 |
+
"num_input_tokens_seen": 0,
|
| 510 |
+
"num_train_epochs": 20,
|
| 511 |
+
"save_steps": 500,
|
| 512 |
+
"stateful_callbacks": {
|
| 513 |
+
"EarlyStoppingCallback": {
|
| 514 |
+
"args": {
|
| 515 |
+
"early_stopping_patience": 3,
|
| 516 |
+
"early_stopping_threshold": 0.0
|
| 517 |
+
},
|
| 518 |
+
"attributes": {
|
| 519 |
+
"early_stopping_patience_counter": 3
|
| 520 |
+
}
|
| 521 |
+
},
|
| 522 |
+
"TrainerControl": {
|
| 523 |
+
"args": {
|
| 524 |
+
"should_epoch_stop": false,
|
| 525 |
+
"should_evaluate": false,
|
| 526 |
+
"should_log": false,
|
| 527 |
+
"should_save": true,
|
| 528 |
+
"should_training_stop": true
|
| 529 |
+
},
|
| 530 |
+
"attributes": {}
|
| 531 |
+
}
|
| 532 |
+
},
|
| 533 |
+
"total_flos": 0.0,
|
| 534 |
+
"train_batch_size": 8,
|
| 535 |
+
"trial_name": null,
|
| 536 |
+
"trial_params": null
|
| 537 |
+
}
|
ood-vifactcheck-phobert-source-source-03-s43-20260826T154435Z-bdd2c308/trainer/checkpoint-1464/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-source-source-03-s43-20260826T154435Z-bdd2c308/trainer/checkpoint-915/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-source-source-03-s43-20260826T154435Z-bdd2c308/trainer/checkpoint-915/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-source-source-03-s43-20260826T154435Z-bdd2c308/trainer/checkpoint-915/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-source-source-03-s43-20260826T154435Z-bdd2c308/trainer/checkpoint-915/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-source-source-03-s43-20260826T154435Z-bdd2c308/trainer/checkpoint-915/trainer_state.json
ADDED
|
@@ -0,0 +1,350 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 915,
|
| 3 |
+
"best_metric": 0.853797945488543,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-source-source-03-s43-20260826T154435Z-bdd2c308/trainer/checkpoint-915",
|
| 5 |
+
"epoch": 5.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 915,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.13717421124828533,
|
| 14 |
+
"grad_norm": 6.4076714515686035,
|
| 15 |
+
"learning_rate": 1.3114754098360657e-06,
|
| 16 |
+
"loss": 4.4479,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.27434842249657065,
|
| 21 |
+
"grad_norm": 6.469189643859863,
|
| 22 |
+
"learning_rate": 2.677595628415301e-06,
|
| 23 |
+
"loss": 4.4062,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.411522633744856,
|
| 28 |
+
"grad_norm": 9.004118919372559,
|
| 29 |
+
"learning_rate": 4.043715846994536e-06,
|
| 30 |
+
"loss": 4.4233,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5486968449931413,
|
| 35 |
+
"grad_norm": 10.774386405944824,
|
| 36 |
+
"learning_rate": 5.409836065573772e-06,
|
| 37 |
+
"loss": 4.3641,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6858710562414266,
|
| 42 |
+
"grad_norm": 7.2467169761657715,
|
| 43 |
+
"learning_rate": 6.775956284153006e-06,
|
| 44 |
+
"loss": 4.3555,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.823045267489712,
|
| 49 |
+
"grad_norm": 8.021151542663574,
|
| 50 |
+
"learning_rate": 8.14207650273224e-06,
|
| 51 |
+
"loss": 4.3166,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9602194787379973,
|
| 56 |
+
"grad_norm": 8.692425727844238,
|
| 57 |
+
"learning_rate": 9.508196721311476e-06,
|
| 58 |
+
"loss": 4.256,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.5440931780366056,
|
| 64 |
+
"eval_loss": 0.9838680028915405,
|
| 65 |
+
"eval_macro_f1": 0.5369983375409565,
|
| 66 |
+
"eval_runtime": 0.4171,
|
| 67 |
+
"eval_samples_per_second": 1440.887,
|
| 68 |
+
"eval_steps_per_second": 91.104,
|
| 69 |
+
"eval_weighted_f1": 0.5389635726736528,
|
| 70 |
+
"step": 183
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.093278463648834,
|
| 74 |
+
"grad_norm": 12.057026863098145,
|
| 75 |
+
"learning_rate": 1.0874316939890712e-05,
|
| 76 |
+
"loss": 3.8209,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.2304526748971194,
|
| 81 |
+
"grad_norm": 23.84331512451172,
|
| 82 |
+
"learning_rate": 1.2240437158469946e-05,
|
| 83 |
+
"loss": 3.552,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.3676268861454046,
|
| 88 |
+
"grad_norm": 34.658504486083984,
|
| 89 |
+
"learning_rate": 1.3606557377049181e-05,
|
| 90 |
+
"loss": 3.2955,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.50480109739369,
|
| 95 |
+
"grad_norm": 35.16084289550781,
|
| 96 |
+
"learning_rate": 1.4972677595628417e-05,
|
| 97 |
+
"loss": 2.9104,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.6419753086419753,
|
| 102 |
+
"grad_norm": 41.14034652709961,
|
| 103 |
+
"learning_rate": 1.633879781420765e-05,
|
| 104 |
+
"loss": 2.6576,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7791495198902605,
|
| 109 |
+
"grad_norm": 48.54668045043945,
|
| 110 |
+
"learning_rate": 1.7704918032786887e-05,
|
| 111 |
+
"loss": 2.6274,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.916323731138546,
|
| 116 |
+
"grad_norm": 51.720664978027344,
|
| 117 |
+
"learning_rate": 1.9071038251366122e-05,
|
| 118 |
+
"loss": 2.349,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.8419301164725458,
|
| 124 |
+
"eval_loss": 0.4657325744628906,
|
| 125 |
+
"eval_macro_f1": 0.8420572992124283,
|
| 126 |
+
"eval_runtime": 0.3881,
|
| 127 |
+
"eval_samples_per_second": 1548.683,
|
| 128 |
+
"eval_steps_per_second": 97.92,
|
| 129 |
+
"eval_weighted_f1": 0.841929537922293,
|
| 130 |
+
"step": 366
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.049382716049383,
|
| 134 |
+
"grad_norm": 33.88608932495117,
|
| 135 |
+
"learning_rate": 1.995142683667274e-05,
|
| 136 |
+
"loss": 1.9099,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.186556927297668,
|
| 141 |
+
"grad_norm": 38.441776275634766,
|
| 142 |
+
"learning_rate": 1.9799635701275045e-05,
|
| 143 |
+
"loss": 1.8178,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.3237311385459534,
|
| 148 |
+
"grad_norm": 23.70911407470703,
|
| 149 |
+
"learning_rate": 1.9647844565877354e-05,
|
| 150 |
+
"loss": 1.5067,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.460905349794239,
|
| 155 |
+
"grad_norm": 24.64794921875,
|
| 156 |
+
"learning_rate": 1.9496053430479663e-05,
|
| 157 |
+
"loss": 1.4754,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.598079561042524,
|
| 162 |
+
"grad_norm": 53.598976135253906,
|
| 163 |
+
"learning_rate": 1.934426229508197e-05,
|
| 164 |
+
"loss": 1.6642,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.7352537722908092,
|
| 169 |
+
"grad_norm": 28.306968688964844,
|
| 170 |
+
"learning_rate": 1.9192471159684274e-05,
|
| 171 |
+
"loss": 1.7823,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.8724279835390947,
|
| 176 |
+
"grad_norm": 58.7725715637207,
|
| 177 |
+
"learning_rate": 1.9040680024286583e-05,
|
| 178 |
+
"loss": 1.7782,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 3.0,
|
| 183 |
+
"eval_accuracy": 0.8535773710482529,
|
| 184 |
+
"eval_loss": 0.43532466888427734,
|
| 185 |
+
"eval_macro_f1": 0.8534562192837053,
|
| 186 |
+
"eval_runtime": 0.5288,
|
| 187 |
+
"eval_samples_per_second": 1136.623,
|
| 188 |
+
"eval_steps_per_second": 71.866,
|
| 189 |
+
"eval_weighted_f1": 0.8528918706451937,
|
| 190 |
+
"step": 549
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.0054869684499312,
|
| 194 |
+
"grad_norm": 98.39491271972656,
|
| 195 |
+
"learning_rate": 1.888888888888889e-05,
|
| 196 |
+
"loss": 1.5376,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.1426611796982167,
|
| 201 |
+
"grad_norm": 56.169395446777344,
|
| 202 |
+
"learning_rate": 1.8737097753491198e-05,
|
| 203 |
+
"loss": 1.1509,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.279835390946502,
|
| 208 |
+
"grad_norm": 100.4572525024414,
|
| 209 |
+
"learning_rate": 1.8585306618093504e-05,
|
| 210 |
+
"loss": 1.1175,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.4170096021947876,
|
| 215 |
+
"grad_norm": 32.61418533325195,
|
| 216 |
+
"learning_rate": 1.8433515482695813e-05,
|
| 217 |
+
"loss": 1.1605,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.5541838134430725,
|
| 222 |
+
"grad_norm": 50.624610900878906,
|
| 223 |
+
"learning_rate": 1.828172434729812e-05,
|
| 224 |
+
"loss": 1.1586,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.691358024691358,
|
| 229 |
+
"grad_norm": 23.299829483032227,
|
| 230 |
+
"learning_rate": 1.8129933211900424e-05,
|
| 231 |
+
"loss": 1.1789,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.8285322359396434,
|
| 236 |
+
"grad_norm": 111.1823959350586,
|
| 237 |
+
"learning_rate": 1.7978142076502733e-05,
|
| 238 |
+
"loss": 0.9352,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.9657064471879284,
|
| 243 |
+
"grad_norm": 33.47234344482422,
|
| 244 |
+
"learning_rate": 1.7826350941105042e-05,
|
| 245 |
+
"loss": 1.0522,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 4.0,
|
| 250 |
+
"eval_accuracy": 0.848585690515807,
|
| 251 |
+
"eval_loss": 0.43581661581993103,
|
| 252 |
+
"eval_macro_f1": 0.848643240658458,
|
| 253 |
+
"eval_runtime": 0.4208,
|
| 254 |
+
"eval_samples_per_second": 1428.299,
|
| 255 |
+
"eval_steps_per_second": 90.308,
|
| 256 |
+
"eval_weighted_f1": 0.8482517042566979,
|
| 257 |
+
"step": 732
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.098765432098766,
|
| 261 |
+
"grad_norm": 26.348663330078125,
|
| 262 |
+
"learning_rate": 1.7674559805707348e-05,
|
| 263 |
+
"loss": 0.8097,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.23593964334705,
|
| 268 |
+
"grad_norm": 15.588164329528809,
|
| 269 |
+
"learning_rate": 1.7522768670309657e-05,
|
| 270 |
+
"loss": 0.7674,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.373113854595336,
|
| 275 |
+
"grad_norm": 9.365654945373535,
|
| 276 |
+
"learning_rate": 1.7370977534911962e-05,
|
| 277 |
+
"loss": 0.8267,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.510288065843621,
|
| 282 |
+
"grad_norm": 26.54583168029785,
|
| 283 |
+
"learning_rate": 1.7219186399514268e-05,
|
| 284 |
+
"loss": 0.8376,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.647462277091907,
|
| 289 |
+
"grad_norm": 44.178321838378906,
|
| 290 |
+
"learning_rate": 1.7067395264116577e-05,
|
| 291 |
+
"loss": 0.7836,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.784636488340192,
|
| 296 |
+
"grad_norm": 45.240657806396484,
|
| 297 |
+
"learning_rate": 1.6915604128718883e-05,
|
| 298 |
+
"loss": 0.7959,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.921810699588478,
|
| 303 |
+
"grad_norm": 17.089048385620117,
|
| 304 |
+
"learning_rate": 1.6763812993321192e-05,
|
| 305 |
+
"loss": 0.7625,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 5.0,
|
| 310 |
+
"eval_accuracy": 0.8535773710482529,
|
| 311 |
+
"eval_loss": 0.507513701915741,
|
| 312 |
+
"eval_macro_f1": 0.853797945488543,
|
| 313 |
+
"eval_runtime": 0.4211,
|
| 314 |
+
"eval_samples_per_second": 1427.139,
|
| 315 |
+
"eval_steps_per_second": 90.235,
|
| 316 |
+
"eval_weighted_f1": 0.8538355369232116,
|
| 317 |
+
"step": 915
|
| 318 |
+
}
|
| 319 |
+
],
|
| 320 |
+
"logging_steps": 25,
|
| 321 |
+
"max_steps": 3660,
|
| 322 |
+
"num_input_tokens_seen": 0,
|
| 323 |
+
"num_train_epochs": 20,
|
| 324 |
+
"save_steps": 500,
|
| 325 |
+
"stateful_callbacks": {
|
| 326 |
+
"EarlyStoppingCallback": {
|
| 327 |
+
"args": {
|
| 328 |
+
"early_stopping_patience": 3,
|
| 329 |
+
"early_stopping_threshold": 0.0
|
| 330 |
+
},
|
| 331 |
+
"attributes": {
|
| 332 |
+
"early_stopping_patience_counter": 0
|
| 333 |
+
}
|
| 334 |
+
},
|
| 335 |
+
"TrainerControl": {
|
| 336 |
+
"args": {
|
| 337 |
+
"should_epoch_stop": false,
|
| 338 |
+
"should_evaluate": false,
|
| 339 |
+
"should_log": false,
|
| 340 |
+
"should_save": true,
|
| 341 |
+
"should_training_stop": false
|
| 342 |
+
},
|
| 343 |
+
"attributes": {}
|
| 344 |
+
}
|
| 345 |
+
},
|
| 346 |
+
"total_flos": 0.0,
|
| 347 |
+
"train_batch_size": 8,
|
| 348 |
+
"trial_name": null,
|
| 349 |
+
"trial_params": null
|
| 350 |
+
}
|
ood-vifactcheck-phobert-source-source-03-s43-20260826T154435Z-bdd2c308/trainer/checkpoint-915/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-source-source-04-s43-20260826T154723Z-bdd2c308/trainer/checkpoint-1408/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-source-source-04-s43-20260826T154723Z-bdd2c308/trainer/checkpoint-1408/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|