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-s42-20260826T150739Z-00ca7096/trainer/checkpoint-1092/added_tokens.json +3 -0
- ood-vifactcheck-phobert-source-source-00-s42-20260826T150739Z-00ca7096/trainer/checkpoint-1092/bpe.codes +0 -0
- ood-vifactcheck-phobert-source-source-00-s42-20260826T150739Z-00ca7096/trainer/checkpoint-1092/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-source-source-00-s42-20260826T150739Z-00ca7096/trainer/checkpoint-1092/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-source-source-00-s42-20260826T150739Z-00ca7096/trainer/checkpoint-1092/trainer_state.json +410 -0
- ood-vifactcheck-phobert-source-source-00-s42-20260826T150739Z-00ca7096/trainer/checkpoint-1092/vocab.txt +0 -0
- ood-vifactcheck-phobert-source-source-00-s42-20260826T150739Z-00ca7096/trainer/checkpoint-1638/added_tokens.json +3 -0
- ood-vifactcheck-phobert-source-source-00-s42-20260826T150739Z-00ca7096/trainer/checkpoint-1638/bpe.codes +0 -0
- ood-vifactcheck-phobert-source-source-00-s42-20260826T150739Z-00ca7096/trainer/checkpoint-1638/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-source-source-00-s42-20260826T150739Z-00ca7096/trainer/checkpoint-1638/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-source-source-00-s42-20260826T150739Z-00ca7096/trainer/checkpoint-1638/trainer_state.json +597 -0
- ood-vifactcheck-phobert-source-source-00-s42-20260826T150739Z-00ca7096/trainer/checkpoint-1638/vocab.txt +0 -0
- ood-vifactcheck-phobert-source-source-01-s42-20260826T151112Z-00ca7096/trainer/checkpoint-1026/added_tokens.json +3 -0
- ood-vifactcheck-phobert-source-source-01-s42-20260826T151112Z-00ca7096/trainer/checkpoint-1026/bpe.codes +0 -0
- ood-vifactcheck-phobert-source-source-01-s42-20260826T151112Z-00ca7096/trainer/checkpoint-1026/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-source-source-01-s42-20260826T151112Z-00ca7096/trainer/checkpoint-1026/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-source-source-01-s42-20260826T151112Z-00ca7096/trainer/checkpoint-1026/trainer_state.json +396 -0
- ood-vifactcheck-phobert-source-source-01-s42-20260826T151112Z-00ca7096/trainer/checkpoint-1026/vocab.txt +0 -0
- ood-vifactcheck-phobert-source-source-01-s42-20260826T151112Z-00ca7096/trainer/checkpoint-513/added_tokens.json +3 -0
- ood-vifactcheck-phobert-source-source-01-s42-20260826T151112Z-00ca7096/trainer/checkpoint-513/bpe.codes +0 -0
- ood-vifactcheck-phobert-source-source-01-s42-20260826T151112Z-00ca7096/trainer/checkpoint-513/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-source-source-01-s42-20260826T151112Z-00ca7096/trainer/checkpoint-513/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-source-source-01-s42-20260826T151112Z-00ca7096/trainer/checkpoint-513/trainer_state.json +216 -0
- ood-vifactcheck-phobert-source-source-01-s42-20260826T151112Z-00ca7096/trainer/checkpoint-513/vocab.txt +0 -0
- ood-vifactcheck-phobert-source-source-02-s42-20260826T151325Z-00ca7096/trainer/checkpoint-1086/added_tokens.json +3 -0
- ood-vifactcheck-phobert-source-source-02-s42-20260826T151325Z-00ca7096/trainer/checkpoint-1086/bpe.codes +0 -0
- ood-vifactcheck-phobert-source-source-02-s42-20260826T151325Z-00ca7096/trainer/checkpoint-1086/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-source-source-02-s42-20260826T151325Z-00ca7096/trainer/checkpoint-1086/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-source-source-02-s42-20260826T151325Z-00ca7096/trainer/checkpoint-1086/trainer_state.json +410 -0
- ood-vifactcheck-phobert-source-source-02-s42-20260826T151325Z-00ca7096/trainer/checkpoint-1086/vocab.txt +0 -0
- ood-vifactcheck-phobert-source-source-02-s42-20260826T151325Z-00ca7096/trainer/checkpoint-1629/added_tokens.json +3 -0
- ood-vifactcheck-phobert-source-source-02-s42-20260826T151325Z-00ca7096/trainer/checkpoint-1629/bpe.codes +0 -0
- ood-vifactcheck-phobert-source-source-02-s42-20260826T151325Z-00ca7096/trainer/checkpoint-1629/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-source-source-02-s42-20260826T151325Z-00ca7096/trainer/checkpoint-1629/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-source-source-02-s42-20260826T151325Z-00ca7096/trainer/checkpoint-1629/trainer_state.json +597 -0
- ood-vifactcheck-phobert-source-source-02-s42-20260826T151325Z-00ca7096/trainer/checkpoint-1629/vocab.txt +0 -0
- ood-vifactcheck-phobert-source-source-03-s42-20260826T151649Z-00ca7096/trainer/checkpoint-1464/added_tokens.json +3 -0
- ood-vifactcheck-phobert-source-source-03-s42-20260826T151649Z-00ca7096/trainer/checkpoint-1464/bpe.codes +0 -0
- ood-vifactcheck-phobert-source-source-03-s42-20260826T151649Z-00ca7096/trainer/checkpoint-1464/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-source-source-03-s42-20260826T151649Z-00ca7096/trainer/checkpoint-1464/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-source-source-03-s42-20260826T151649Z-00ca7096/trainer/checkpoint-1464/trainer_state.json +537 -0
- ood-vifactcheck-phobert-source-source-03-s42-20260826T151649Z-00ca7096/trainer/checkpoint-1464/vocab.txt +0 -0
- ood-vifactcheck-phobert-source-source-03-s42-20260826T151649Z-00ca7096/trainer/checkpoint-2013/added_tokens.json +3 -0
- ood-vifactcheck-phobert-source-source-03-s42-20260826T151649Z-00ca7096/trainer/checkpoint-2013/bpe.codes +0 -0
- ood-vifactcheck-phobert-source-source-03-s42-20260826T151649Z-00ca7096/trainer/checkpoint-2013/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-source-source-03-s42-20260826T151649Z-00ca7096/trainer/checkpoint-2013/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-source-source-03-s42-20260826T151649Z-00ca7096/trainer/checkpoint-2013/trainer_state.json +724 -0
- ood-vifactcheck-phobert-source-source-03-s42-20260826T151649Z-00ca7096/trainer/checkpoint-2013/vocab.txt +0 -0
- ood-vifactcheck-phobert-source-source-04-s42-20260826T152033Z-00ca7096/trainer/checkpoint-1408/added_tokens.json +3 -0
- ood-vifactcheck-phobert-source-source-04-s42-20260826T152033Z-00ca7096/trainer/checkpoint-1408/bpe.codes +0 -0
ood-vifactcheck-phobert-source-source-00-s42-20260826T150739Z-00ca7096/trainer/checkpoint-1092/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-source-source-00-s42-20260826T150739Z-00ca7096/trainer/checkpoint-1092/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-source-source-00-s42-20260826T150739Z-00ca7096/trainer/checkpoint-1092/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-s42-20260826T150739Z-00ca7096/trainer/checkpoint-1092/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-s42-20260826T150739Z-00ca7096/trainer/checkpoint-1092/trainer_state.json
ADDED
|
@@ -0,0 +1,410 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1092,
|
| 3 |
+
"best_metric": 0.8372604723707665,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-source-source-00-s42-20260826T150739Z-00ca7096/trainer/checkpoint-1092",
|
| 5 |
+
"epoch": 6.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1092,
|
| 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": 8.781030654907227,
|
| 15 |
+
"learning_rate": 1.3186813186813187e-06,
|
| 16 |
+
"loss": 4.4288,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.27586206896551724,
|
| 21 |
+
"grad_norm": 9.2155122756958,
|
| 22 |
+
"learning_rate": 2.6923076923076923e-06,
|
| 23 |
+
"loss": 4.396,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.41379310344827586,
|
| 28 |
+
"grad_norm": 10.585765838623047,
|
| 29 |
+
"learning_rate": 4.065934065934066e-06,
|
| 30 |
+
"loss": 4.3774,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5517241379310345,
|
| 35 |
+
"grad_norm": 7.612149715423584,
|
| 36 |
+
"learning_rate": 5.43956043956044e-06,
|
| 37 |
+
"loss": 4.3767,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6896551724137931,
|
| 42 |
+
"grad_norm": 8.606959342956543,
|
| 43 |
+
"learning_rate": 6.813186813186814e-06,
|
| 44 |
+
"loss": 4.3419,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8275862068965517,
|
| 49 |
+
"grad_norm": 7.943285942077637,
|
| 50 |
+
"learning_rate": 8.186813186813188e-06,
|
| 51 |
+
"loss": 4.2616,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9655172413793104,
|
| 56 |
+
"grad_norm": 13.161110877990723,
|
| 57 |
+
"learning_rate": 9.560439560439562e-06,
|
| 58 |
+
"loss": 4.1452,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.5055821371610846,
|
| 64 |
+
"eval_loss": 0.9872875809669495,
|
| 65 |
+
"eval_macro_f1": 0.4581447856224181,
|
| 66 |
+
"eval_runtime": 0.5325,
|
| 67 |
+
"eval_samples_per_second": 1177.553,
|
| 68 |
+
"eval_steps_per_second": 75.123,
|
| 69 |
+
"eval_weighted_f1": 0.46422434349631536,
|
| 70 |
+
"step": 182
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0993103448275863,
|
| 74 |
+
"grad_norm": 51.72880554199219,
|
| 75 |
+
"learning_rate": 1.0934065934065933e-05,
|
| 76 |
+
"loss": 3.7636,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.237241379310345,
|
| 81 |
+
"grad_norm": 43.20492935180664,
|
| 82 |
+
"learning_rate": 1.230769230769231e-05,
|
| 83 |
+
"loss": 3.3896,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.3751724137931034,
|
| 88 |
+
"grad_norm": 40.41683578491211,
|
| 89 |
+
"learning_rate": 1.3681318681318683e-05,
|
| 90 |
+
"loss": 3.1445,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.513103448275862,
|
| 95 |
+
"grad_norm": 46.37562561035156,
|
| 96 |
+
"learning_rate": 1.5054945054945056e-05,
|
| 97 |
+
"loss": 2.807,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.6510344827586207,
|
| 102 |
+
"grad_norm": 38.70273208618164,
|
| 103 |
+
"learning_rate": 1.642857142857143e-05,
|
| 104 |
+
"loss": 2.4614,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7889655172413792,
|
| 109 |
+
"grad_norm": 44.936439514160156,
|
| 110 |
+
"learning_rate": 1.78021978021978e-05,
|
| 111 |
+
"loss": 2.3582,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.926896551724138,
|
| 116 |
+
"grad_norm": 36.20042037963867,
|
| 117 |
+
"learning_rate": 1.9175824175824178e-05,
|
| 118 |
+
"loss": 2.0656,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.7878787878787878,
|
| 124 |
+
"eval_loss": 0.5488452315330505,
|
| 125 |
+
"eval_macro_f1": 0.790145583937237,
|
| 126 |
+
"eval_runtime": 0.4441,
|
| 127 |
+
"eval_samples_per_second": 1411.831,
|
| 128 |
+
"eval_steps_per_second": 90.069,
|
| 129 |
+
"eval_weighted_f1": 0.7898133410168393,
|
| 130 |
+
"step": 364
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.0606896551724136,
|
| 134 |
+
"grad_norm": 34.499168395996094,
|
| 135 |
+
"learning_rate": 1.993894993894994e-05,
|
| 136 |
+
"loss": 1.8135,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.1986206896551725,
|
| 141 |
+
"grad_norm": 47.76192092895508,
|
| 142 |
+
"learning_rate": 1.9786324786324788e-05,
|
| 143 |
+
"loss": 1.6279,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.336551724137931,
|
| 148 |
+
"grad_norm": 47.168907165527344,
|
| 149 |
+
"learning_rate": 1.9633699633699634e-05,
|
| 150 |
+
"loss": 1.6226,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.47448275862069,
|
| 155 |
+
"grad_norm": 26.09088706970215,
|
| 156 |
+
"learning_rate": 1.948107448107448e-05,
|
| 157 |
+
"loss": 1.5349,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.6124137931034483,
|
| 162 |
+
"grad_norm": 44.584800720214844,
|
| 163 |
+
"learning_rate": 1.932844932844933e-05,
|
| 164 |
+
"loss": 1.535,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.750344827586207,
|
| 169 |
+
"grad_norm": 55.80675506591797,
|
| 170 |
+
"learning_rate": 1.9175824175824178e-05,
|
| 171 |
+
"loss": 1.4094,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.8882758620689657,
|
| 176 |
+
"grad_norm": 25.840045928955078,
|
| 177 |
+
"learning_rate": 1.9023199023199025e-05,
|
| 178 |
+
"loss": 1.7045,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 3.0,
|
| 183 |
+
"eval_accuracy": 0.8245614035087719,
|
| 184 |
+
"eval_loss": 0.5314666628837585,
|
| 185 |
+
"eval_macro_f1": 0.8234695861563542,
|
| 186 |
+
"eval_runtime": 0.5736,
|
| 187 |
+
"eval_samples_per_second": 1093.11,
|
| 188 |
+
"eval_steps_per_second": 69.736,
|
| 189 |
+
"eval_weighted_f1": 0.8238643206727031,
|
| 190 |
+
"step": 546
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.0220689655172412,
|
| 194 |
+
"grad_norm": 26.39698600769043,
|
| 195 |
+
"learning_rate": 1.887057387057387e-05,
|
| 196 |
+
"loss": 1.3128,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.16,
|
| 201 |
+
"grad_norm": 77.0368881225586,
|
| 202 |
+
"learning_rate": 1.8717948717948718e-05,
|
| 203 |
+
"loss": 0.8658,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.2979310344827586,
|
| 208 |
+
"grad_norm": 15.0227632522583,
|
| 209 |
+
"learning_rate": 1.8565323565323568e-05,
|
| 210 |
+
"loss": 0.9315,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.435862068965517,
|
| 215 |
+
"grad_norm": 34.49467086791992,
|
| 216 |
+
"learning_rate": 1.8412698412698415e-05,
|
| 217 |
+
"loss": 1.0376,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.573793103448276,
|
| 222 |
+
"grad_norm": 42.42506790161133,
|
| 223 |
+
"learning_rate": 1.826007326007326e-05,
|
| 224 |
+
"loss": 0.8994,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.7117241379310344,
|
| 229 |
+
"grad_norm": 117.4176025390625,
|
| 230 |
+
"learning_rate": 1.8107448107448108e-05,
|
| 231 |
+
"loss": 0.9095,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.849655172413793,
|
| 236 |
+
"grad_norm": 25.557952880859375,
|
| 237 |
+
"learning_rate": 1.7954822954822958e-05,
|
| 238 |
+
"loss": 1.2512,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.9875862068965517,
|
| 243 |
+
"grad_norm": 67.30958557128906,
|
| 244 |
+
"learning_rate": 1.78021978021978e-05,
|
| 245 |
+
"loss": 1.0467,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 4.0,
|
| 250 |
+
"eval_accuracy": 0.8245614035087719,
|
| 251 |
+
"eval_loss": 0.564399778842926,
|
| 252 |
+
"eval_macro_f1": 0.8240187480081348,
|
| 253 |
+
"eval_runtime": 0.4811,
|
| 254 |
+
"eval_samples_per_second": 1303.139,
|
| 255 |
+
"eval_steps_per_second": 83.135,
|
| 256 |
+
"eval_weighted_f1": 0.8244343276068646,
|
| 257 |
+
"step": 728
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.121379310344827,
|
| 261 |
+
"grad_norm": 27.550617218017578,
|
| 262 |
+
"learning_rate": 1.764957264957265e-05,
|
| 263 |
+
"loss": 0.7177,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.259310344827586,
|
| 268 |
+
"grad_norm": 82.70658111572266,
|
| 269 |
+
"learning_rate": 1.74969474969475e-05,
|
| 270 |
+
"loss": 0.671,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.397241379310345,
|
| 275 |
+
"grad_norm": 41.79597854614258,
|
| 276 |
+
"learning_rate": 1.7344322344322345e-05,
|
| 277 |
+
"loss": 0.7191,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.535172413793103,
|
| 282 |
+
"grad_norm": 44.62472915649414,
|
| 283 |
+
"learning_rate": 1.7191697191697195e-05,
|
| 284 |
+
"loss": 0.6617,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.673103448275862,
|
| 289 |
+
"grad_norm": 62.12500762939453,
|
| 290 |
+
"learning_rate": 1.703907203907204e-05,
|
| 291 |
+
"loss": 0.5849,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.811034482758621,
|
| 296 |
+
"grad_norm": 85.72279357910156,
|
| 297 |
+
"learning_rate": 1.688644688644689e-05,
|
| 298 |
+
"loss": 0.8882,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.94896551724138,
|
| 303 |
+
"grad_norm": 107.17594146728516,
|
| 304 |
+
"learning_rate": 1.6733821733821735e-05,
|
| 305 |
+
"loss": 0.5734,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 5.0,
|
| 310 |
+
"eval_accuracy": 0.8245614035087719,
|
| 311 |
+
"eval_loss": 0.6386942863464355,
|
| 312 |
+
"eval_macro_f1": 0.8242676694017517,
|
| 313 |
+
"eval_runtime": 0.5887,
|
| 314 |
+
"eval_samples_per_second": 1065.106,
|
| 315 |
+
"eval_steps_per_second": 67.949,
|
| 316 |
+
"eval_weighted_f1": 0.8248482095684856,
|
| 317 |
+
"step": 910
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.082758620689655,
|
| 321 |
+
"grad_norm": 98.38258361816406,
|
| 322 |
+
"learning_rate": 1.6581196581196585e-05,
|
| 323 |
+
"loss": 0.5686,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.220689655172414,
|
| 328 |
+
"grad_norm": 52.097164154052734,
|
| 329 |
+
"learning_rate": 1.642857142857143e-05,
|
| 330 |
+
"loss": 0.5493,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.358620689655172,
|
| 335 |
+
"grad_norm": 28.596345901489258,
|
| 336 |
+
"learning_rate": 1.627594627594628e-05,
|
| 337 |
+
"loss": 0.5761,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.496551724137931,
|
| 342 |
+
"grad_norm": 26.687185287475586,
|
| 343 |
+
"learning_rate": 1.6123321123321125e-05,
|
| 344 |
+
"loss": 0.5517,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.63448275862069,
|
| 349 |
+
"grad_norm": 73.08152770996094,
|
| 350 |
+
"learning_rate": 1.5970695970695972e-05,
|
| 351 |
+
"loss": 0.5936,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.772413793103448,
|
| 356 |
+
"grad_norm": 0.5139356851577759,
|
| 357 |
+
"learning_rate": 1.581807081807082e-05,
|
| 358 |
+
"loss": 0.4198,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.910344827586207,
|
| 363 |
+
"grad_norm": 18.04680633544922,
|
| 364 |
+
"learning_rate": 1.5665445665445665e-05,
|
| 365 |
+
"loss": 0.6345,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 6.0,
|
| 370 |
+
"eval_accuracy": 0.8373205741626795,
|
| 371 |
+
"eval_loss": 0.6907906532287598,
|
| 372 |
+
"eval_macro_f1": 0.8372604723707665,
|
| 373 |
+
"eval_runtime": 0.4883,
|
| 374 |
+
"eval_samples_per_second": 1283.949,
|
| 375 |
+
"eval_steps_per_second": 81.911,
|
| 376 |
+
"eval_weighted_f1": 0.8374247861389009,
|
| 377 |
+
"step": 1092
|
| 378 |
+
}
|
| 379 |
+
],
|
| 380 |
+
"logging_steps": 25,
|
| 381 |
+
"max_steps": 3640,
|
| 382 |
+
"num_input_tokens_seen": 0,
|
| 383 |
+
"num_train_epochs": 20,
|
| 384 |
+
"save_steps": 500,
|
| 385 |
+
"stateful_callbacks": {
|
| 386 |
+
"EarlyStoppingCallback": {
|
| 387 |
+
"args": {
|
| 388 |
+
"early_stopping_patience": 3,
|
| 389 |
+
"early_stopping_threshold": 0.0
|
| 390 |
+
},
|
| 391 |
+
"attributes": {
|
| 392 |
+
"early_stopping_patience_counter": 0
|
| 393 |
+
}
|
| 394 |
+
},
|
| 395 |
+
"TrainerControl": {
|
| 396 |
+
"args": {
|
| 397 |
+
"should_epoch_stop": false,
|
| 398 |
+
"should_evaluate": false,
|
| 399 |
+
"should_log": false,
|
| 400 |
+
"should_save": true,
|
| 401 |
+
"should_training_stop": false
|
| 402 |
+
},
|
| 403 |
+
"attributes": {}
|
| 404 |
+
}
|
| 405 |
+
},
|
| 406 |
+
"total_flos": 0.0,
|
| 407 |
+
"train_batch_size": 8,
|
| 408 |
+
"trial_name": null,
|
| 409 |
+
"trial_params": null
|
| 410 |
+
}
|
ood-vifactcheck-phobert-source-source-00-s42-20260826T150739Z-00ca7096/trainer/checkpoint-1092/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-source-source-00-s42-20260826T150739Z-00ca7096/trainer/checkpoint-1638/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-source-source-00-s42-20260826T150739Z-00ca7096/trainer/checkpoint-1638/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-source-source-00-s42-20260826T150739Z-00ca7096/trainer/checkpoint-1638/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-s42-20260826T150739Z-00ca7096/trainer/checkpoint-1638/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-s42-20260826T150739Z-00ca7096/trainer/checkpoint-1638/trainer_state.json
ADDED
|
@@ -0,0 +1,597 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1092,
|
| 3 |
+
"best_metric": 0.8372604723707665,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-source-source-00-s42-20260826T150739Z-00ca7096/trainer/checkpoint-1092",
|
| 5 |
+
"epoch": 9.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1638,
|
| 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": 8.781030654907227,
|
| 15 |
+
"learning_rate": 1.3186813186813187e-06,
|
| 16 |
+
"loss": 4.4288,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.27586206896551724,
|
| 21 |
+
"grad_norm": 9.2155122756958,
|
| 22 |
+
"learning_rate": 2.6923076923076923e-06,
|
| 23 |
+
"loss": 4.396,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.41379310344827586,
|
| 28 |
+
"grad_norm": 10.585765838623047,
|
| 29 |
+
"learning_rate": 4.065934065934066e-06,
|
| 30 |
+
"loss": 4.3774,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5517241379310345,
|
| 35 |
+
"grad_norm": 7.612149715423584,
|
| 36 |
+
"learning_rate": 5.43956043956044e-06,
|
| 37 |
+
"loss": 4.3767,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6896551724137931,
|
| 42 |
+
"grad_norm": 8.606959342956543,
|
| 43 |
+
"learning_rate": 6.813186813186814e-06,
|
| 44 |
+
"loss": 4.3419,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8275862068965517,
|
| 49 |
+
"grad_norm": 7.943285942077637,
|
| 50 |
+
"learning_rate": 8.186813186813188e-06,
|
| 51 |
+
"loss": 4.2616,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9655172413793104,
|
| 56 |
+
"grad_norm": 13.161110877990723,
|
| 57 |
+
"learning_rate": 9.560439560439562e-06,
|
| 58 |
+
"loss": 4.1452,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.5055821371610846,
|
| 64 |
+
"eval_loss": 0.9872875809669495,
|
| 65 |
+
"eval_macro_f1": 0.4581447856224181,
|
| 66 |
+
"eval_runtime": 0.5325,
|
| 67 |
+
"eval_samples_per_second": 1177.553,
|
| 68 |
+
"eval_steps_per_second": 75.123,
|
| 69 |
+
"eval_weighted_f1": 0.46422434349631536,
|
| 70 |
+
"step": 182
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0993103448275863,
|
| 74 |
+
"grad_norm": 51.72880554199219,
|
| 75 |
+
"learning_rate": 1.0934065934065933e-05,
|
| 76 |
+
"loss": 3.7636,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.237241379310345,
|
| 81 |
+
"grad_norm": 43.20492935180664,
|
| 82 |
+
"learning_rate": 1.230769230769231e-05,
|
| 83 |
+
"loss": 3.3896,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.3751724137931034,
|
| 88 |
+
"grad_norm": 40.41683578491211,
|
| 89 |
+
"learning_rate": 1.3681318681318683e-05,
|
| 90 |
+
"loss": 3.1445,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.513103448275862,
|
| 95 |
+
"grad_norm": 46.37562561035156,
|
| 96 |
+
"learning_rate": 1.5054945054945056e-05,
|
| 97 |
+
"loss": 2.807,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.6510344827586207,
|
| 102 |
+
"grad_norm": 38.70273208618164,
|
| 103 |
+
"learning_rate": 1.642857142857143e-05,
|
| 104 |
+
"loss": 2.4614,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7889655172413792,
|
| 109 |
+
"grad_norm": 44.936439514160156,
|
| 110 |
+
"learning_rate": 1.78021978021978e-05,
|
| 111 |
+
"loss": 2.3582,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.926896551724138,
|
| 116 |
+
"grad_norm": 36.20042037963867,
|
| 117 |
+
"learning_rate": 1.9175824175824178e-05,
|
| 118 |
+
"loss": 2.0656,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.7878787878787878,
|
| 124 |
+
"eval_loss": 0.5488452315330505,
|
| 125 |
+
"eval_macro_f1": 0.790145583937237,
|
| 126 |
+
"eval_runtime": 0.4441,
|
| 127 |
+
"eval_samples_per_second": 1411.831,
|
| 128 |
+
"eval_steps_per_second": 90.069,
|
| 129 |
+
"eval_weighted_f1": 0.7898133410168393,
|
| 130 |
+
"step": 364
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.0606896551724136,
|
| 134 |
+
"grad_norm": 34.499168395996094,
|
| 135 |
+
"learning_rate": 1.993894993894994e-05,
|
| 136 |
+
"loss": 1.8135,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.1986206896551725,
|
| 141 |
+
"grad_norm": 47.76192092895508,
|
| 142 |
+
"learning_rate": 1.9786324786324788e-05,
|
| 143 |
+
"loss": 1.6279,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.336551724137931,
|
| 148 |
+
"grad_norm": 47.168907165527344,
|
| 149 |
+
"learning_rate": 1.9633699633699634e-05,
|
| 150 |
+
"loss": 1.6226,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.47448275862069,
|
| 155 |
+
"grad_norm": 26.09088706970215,
|
| 156 |
+
"learning_rate": 1.948107448107448e-05,
|
| 157 |
+
"loss": 1.5349,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.6124137931034483,
|
| 162 |
+
"grad_norm": 44.584800720214844,
|
| 163 |
+
"learning_rate": 1.932844932844933e-05,
|
| 164 |
+
"loss": 1.535,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.750344827586207,
|
| 169 |
+
"grad_norm": 55.80675506591797,
|
| 170 |
+
"learning_rate": 1.9175824175824178e-05,
|
| 171 |
+
"loss": 1.4094,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.8882758620689657,
|
| 176 |
+
"grad_norm": 25.840045928955078,
|
| 177 |
+
"learning_rate": 1.9023199023199025e-05,
|
| 178 |
+
"loss": 1.7045,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 3.0,
|
| 183 |
+
"eval_accuracy": 0.8245614035087719,
|
| 184 |
+
"eval_loss": 0.5314666628837585,
|
| 185 |
+
"eval_macro_f1": 0.8234695861563542,
|
| 186 |
+
"eval_runtime": 0.5736,
|
| 187 |
+
"eval_samples_per_second": 1093.11,
|
| 188 |
+
"eval_steps_per_second": 69.736,
|
| 189 |
+
"eval_weighted_f1": 0.8238643206727031,
|
| 190 |
+
"step": 546
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.0220689655172412,
|
| 194 |
+
"grad_norm": 26.39698600769043,
|
| 195 |
+
"learning_rate": 1.887057387057387e-05,
|
| 196 |
+
"loss": 1.3128,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.16,
|
| 201 |
+
"grad_norm": 77.0368881225586,
|
| 202 |
+
"learning_rate": 1.8717948717948718e-05,
|
| 203 |
+
"loss": 0.8658,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.2979310344827586,
|
| 208 |
+
"grad_norm": 15.0227632522583,
|
| 209 |
+
"learning_rate": 1.8565323565323568e-05,
|
| 210 |
+
"loss": 0.9315,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.435862068965517,
|
| 215 |
+
"grad_norm": 34.49467086791992,
|
| 216 |
+
"learning_rate": 1.8412698412698415e-05,
|
| 217 |
+
"loss": 1.0376,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.573793103448276,
|
| 222 |
+
"grad_norm": 42.42506790161133,
|
| 223 |
+
"learning_rate": 1.826007326007326e-05,
|
| 224 |
+
"loss": 0.8994,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.7117241379310344,
|
| 229 |
+
"grad_norm": 117.4176025390625,
|
| 230 |
+
"learning_rate": 1.8107448107448108e-05,
|
| 231 |
+
"loss": 0.9095,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.849655172413793,
|
| 236 |
+
"grad_norm": 25.557952880859375,
|
| 237 |
+
"learning_rate": 1.7954822954822958e-05,
|
| 238 |
+
"loss": 1.2512,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.9875862068965517,
|
| 243 |
+
"grad_norm": 67.30958557128906,
|
| 244 |
+
"learning_rate": 1.78021978021978e-05,
|
| 245 |
+
"loss": 1.0467,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 4.0,
|
| 250 |
+
"eval_accuracy": 0.8245614035087719,
|
| 251 |
+
"eval_loss": 0.564399778842926,
|
| 252 |
+
"eval_macro_f1": 0.8240187480081348,
|
| 253 |
+
"eval_runtime": 0.4811,
|
| 254 |
+
"eval_samples_per_second": 1303.139,
|
| 255 |
+
"eval_steps_per_second": 83.135,
|
| 256 |
+
"eval_weighted_f1": 0.8244343276068646,
|
| 257 |
+
"step": 728
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.121379310344827,
|
| 261 |
+
"grad_norm": 27.550617218017578,
|
| 262 |
+
"learning_rate": 1.764957264957265e-05,
|
| 263 |
+
"loss": 0.7177,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.259310344827586,
|
| 268 |
+
"grad_norm": 82.70658111572266,
|
| 269 |
+
"learning_rate": 1.74969474969475e-05,
|
| 270 |
+
"loss": 0.671,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.397241379310345,
|
| 275 |
+
"grad_norm": 41.79597854614258,
|
| 276 |
+
"learning_rate": 1.7344322344322345e-05,
|
| 277 |
+
"loss": 0.7191,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.535172413793103,
|
| 282 |
+
"grad_norm": 44.62472915649414,
|
| 283 |
+
"learning_rate": 1.7191697191697195e-05,
|
| 284 |
+
"loss": 0.6617,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.673103448275862,
|
| 289 |
+
"grad_norm": 62.12500762939453,
|
| 290 |
+
"learning_rate": 1.703907203907204e-05,
|
| 291 |
+
"loss": 0.5849,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.811034482758621,
|
| 296 |
+
"grad_norm": 85.72279357910156,
|
| 297 |
+
"learning_rate": 1.688644688644689e-05,
|
| 298 |
+
"loss": 0.8882,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.94896551724138,
|
| 303 |
+
"grad_norm": 107.17594146728516,
|
| 304 |
+
"learning_rate": 1.6733821733821735e-05,
|
| 305 |
+
"loss": 0.5734,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 5.0,
|
| 310 |
+
"eval_accuracy": 0.8245614035087719,
|
| 311 |
+
"eval_loss": 0.6386942863464355,
|
| 312 |
+
"eval_macro_f1": 0.8242676694017517,
|
| 313 |
+
"eval_runtime": 0.5887,
|
| 314 |
+
"eval_samples_per_second": 1065.106,
|
| 315 |
+
"eval_steps_per_second": 67.949,
|
| 316 |
+
"eval_weighted_f1": 0.8248482095684856,
|
| 317 |
+
"step": 910
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.082758620689655,
|
| 321 |
+
"grad_norm": 98.38258361816406,
|
| 322 |
+
"learning_rate": 1.6581196581196585e-05,
|
| 323 |
+
"loss": 0.5686,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.220689655172414,
|
| 328 |
+
"grad_norm": 52.097164154052734,
|
| 329 |
+
"learning_rate": 1.642857142857143e-05,
|
| 330 |
+
"loss": 0.5493,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.358620689655172,
|
| 335 |
+
"grad_norm": 28.596345901489258,
|
| 336 |
+
"learning_rate": 1.627594627594628e-05,
|
| 337 |
+
"loss": 0.5761,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.496551724137931,
|
| 342 |
+
"grad_norm": 26.687185287475586,
|
| 343 |
+
"learning_rate": 1.6123321123321125e-05,
|
| 344 |
+
"loss": 0.5517,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.63448275862069,
|
| 349 |
+
"grad_norm": 73.08152770996094,
|
| 350 |
+
"learning_rate": 1.5970695970695972e-05,
|
| 351 |
+
"loss": 0.5936,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.772413793103448,
|
| 356 |
+
"grad_norm": 0.5139356851577759,
|
| 357 |
+
"learning_rate": 1.581807081807082e-05,
|
| 358 |
+
"loss": 0.4198,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.910344827586207,
|
| 363 |
+
"grad_norm": 18.04680633544922,
|
| 364 |
+
"learning_rate": 1.5665445665445665e-05,
|
| 365 |
+
"loss": 0.6345,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 6.0,
|
| 370 |
+
"eval_accuracy": 0.8373205741626795,
|
| 371 |
+
"eval_loss": 0.6907906532287598,
|
| 372 |
+
"eval_macro_f1": 0.8372604723707665,
|
| 373 |
+
"eval_runtime": 0.4883,
|
| 374 |
+
"eval_samples_per_second": 1283.949,
|
| 375 |
+
"eval_steps_per_second": 81.911,
|
| 376 |
+
"eval_weighted_f1": 0.8374247861389009,
|
| 377 |
+
"step": 1092
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 6.0441379310344825,
|
| 381 |
+
"grad_norm": 2.420173168182373,
|
| 382 |
+
"learning_rate": 1.5512820512820516e-05,
|
| 383 |
+
"loss": 0.4033,
|
| 384 |
+
"step": 1100
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 6.182068965517241,
|
| 388 |
+
"grad_norm": 5.75001335144043,
|
| 389 |
+
"learning_rate": 1.5360195360195362e-05,
|
| 390 |
+
"loss": 0.3476,
|
| 391 |
+
"step": 1125
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 6.32,
|
| 395 |
+
"grad_norm": 1.0390547513961792,
|
| 396 |
+
"learning_rate": 1.5207570207570207e-05,
|
| 397 |
+
"loss": 0.4482,
|
| 398 |
+
"step": 1150
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.457931034482758,
|
| 402 |
+
"grad_norm": 3.1745517253875732,
|
| 403 |
+
"learning_rate": 1.5054945054945056e-05,
|
| 404 |
+
"loss": 0.4839,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.595862068965517,
|
| 409 |
+
"grad_norm": 4.390712738037109,
|
| 410 |
+
"learning_rate": 1.4902319902319904e-05,
|
| 411 |
+
"loss": 0.3,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.733793103448276,
|
| 416 |
+
"grad_norm": 58.13725662231445,
|
| 417 |
+
"learning_rate": 1.4749694749694752e-05,
|
| 418 |
+
"loss": 0.4689,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.871724137931034,
|
| 423 |
+
"grad_norm": 0.8821493983268738,
|
| 424 |
+
"learning_rate": 1.4597069597069597e-05,
|
| 425 |
+
"loss": 0.4705,
|
| 426 |
+
"step": 1250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 7.0,
|
| 430 |
+
"eval_accuracy": 0.8325358851674641,
|
| 431 |
+
"eval_loss": 0.7787495255470276,
|
| 432 |
+
"eval_macro_f1": 0.8330605762139252,
|
| 433 |
+
"eval_runtime": 0.5135,
|
| 434 |
+
"eval_samples_per_second": 1221.007,
|
| 435 |
+
"eval_steps_per_second": 77.895,
|
| 436 |
+
"eval_weighted_f1": 0.8330675654494112,
|
| 437 |
+
"step": 1274
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 7.0055172413793105,
|
| 441 |
+
"grad_norm": 110.0478744506836,
|
| 442 |
+
"learning_rate": 1.4444444444444446e-05,
|
| 443 |
+
"loss": 0.4595,
|
| 444 |
+
"step": 1275
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 7.1434482758620685,
|
| 448 |
+
"grad_norm": 0.38922569155693054,
|
| 449 |
+
"learning_rate": 1.4291819291819293e-05,
|
| 450 |
+
"loss": 0.3367,
|
| 451 |
+
"step": 1300
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 7.281379310344827,
|
| 455 |
+
"grad_norm": 6.160050392150879,
|
| 456 |
+
"learning_rate": 1.4139194139194141e-05,
|
| 457 |
+
"loss": 0.3945,
|
| 458 |
+
"step": 1325
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 7.419310344827586,
|
| 462 |
+
"grad_norm": 56.95576477050781,
|
| 463 |
+
"learning_rate": 1.3986568986568988e-05,
|
| 464 |
+
"loss": 0.3382,
|
| 465 |
+
"step": 1350
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 7.557241379310344,
|
| 469 |
+
"grad_norm": 0.6677722334861755,
|
| 470 |
+
"learning_rate": 1.3833943833943834e-05,
|
| 471 |
+
"loss": 0.28,
|
| 472 |
+
"step": 1375
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 7.695172413793103,
|
| 476 |
+
"grad_norm": 93.78797912597656,
|
| 477 |
+
"learning_rate": 1.3681318681318683e-05,
|
| 478 |
+
"loss": 0.3953,
|
| 479 |
+
"step": 1400
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 7.833103448275862,
|
| 483 |
+
"grad_norm": 7.6941447257995605,
|
| 484 |
+
"learning_rate": 1.3528693528693531e-05,
|
| 485 |
+
"loss": 0.5206,
|
| 486 |
+
"step": 1425
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 7.971034482758621,
|
| 490 |
+
"grad_norm": 1.6897352933883667,
|
| 491 |
+
"learning_rate": 1.3376068376068376e-05,
|
| 492 |
+
"loss": 0.3847,
|
| 493 |
+
"step": 1450
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 8.0,
|
| 497 |
+
"eval_accuracy": 0.8309409888357256,
|
| 498 |
+
"eval_loss": 0.7827369570732117,
|
| 499 |
+
"eval_macro_f1": 0.8316951709306393,
|
| 500 |
+
"eval_runtime": 0.4758,
|
| 501 |
+
"eval_samples_per_second": 1317.705,
|
| 502 |
+
"eval_steps_per_second": 84.064,
|
| 503 |
+
"eval_weighted_f1": 0.8317954891558011,
|
| 504 |
+
"step": 1456
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 8.104827586206897,
|
| 508 |
+
"grad_norm": 0.2674339711666107,
|
| 509 |
+
"learning_rate": 1.3223443223443224e-05,
|
| 510 |
+
"loss": 0.2504,
|
| 511 |
+
"step": 1475
|
| 512 |
+
},
|
| 513 |
+
{
|
| 514 |
+
"epoch": 8.242758620689655,
|
| 515 |
+
"grad_norm": 116.59619903564453,
|
| 516 |
+
"learning_rate": 1.3070818070818073e-05,
|
| 517 |
+
"loss": 0.2873,
|
| 518 |
+
"step": 1500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 8.380689655172414,
|
| 522 |
+
"grad_norm": 1.4999762773513794,
|
| 523 |
+
"learning_rate": 1.291819291819292e-05,
|
| 524 |
+
"loss": 0.3916,
|
| 525 |
+
"step": 1525
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 8.518620689655172,
|
| 529 |
+
"grad_norm": 89.30589294433594,
|
| 530 |
+
"learning_rate": 1.2765567765567766e-05,
|
| 531 |
+
"loss": 0.2112,
|
| 532 |
+
"step": 1550
|
| 533 |
+
},
|
| 534 |
+
{
|
| 535 |
+
"epoch": 8.65655172413793,
|
| 536 |
+
"grad_norm": 0.2756333351135254,
|
| 537 |
+
"learning_rate": 1.2612942612942613e-05,
|
| 538 |
+
"loss": 0.2541,
|
| 539 |
+
"step": 1575
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 8.79448275862069,
|
| 543 |
+
"grad_norm": 0.23524503409862518,
|
| 544 |
+
"learning_rate": 1.2460317460317461e-05,
|
| 545 |
+
"loss": 0.2271,
|
| 546 |
+
"step": 1600
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 8.932413793103448,
|
| 550 |
+
"grad_norm": 479.52410888671875,
|
| 551 |
+
"learning_rate": 1.230769230769231e-05,
|
| 552 |
+
"loss": 0.2214,
|
| 553 |
+
"step": 1625
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 9.0,
|
| 557 |
+
"eval_accuracy": 0.8293460925039873,
|
| 558 |
+
"eval_loss": 0.9199625849723816,
|
| 559 |
+
"eval_macro_f1": 0.8280296092796093,
|
| 560 |
+
"eval_runtime": 0.4975,
|
| 561 |
+
"eval_samples_per_second": 1260.197,
|
| 562 |
+
"eval_steps_per_second": 80.395,
|
| 563 |
+
"eval_weighted_f1": 0.828773748181643,
|
| 564 |
+
"step": 1638
|
| 565 |
+
}
|
| 566 |
+
],
|
| 567 |
+
"logging_steps": 25,
|
| 568 |
+
"max_steps": 3640,
|
| 569 |
+
"num_input_tokens_seen": 0,
|
| 570 |
+
"num_train_epochs": 20,
|
| 571 |
+
"save_steps": 500,
|
| 572 |
+
"stateful_callbacks": {
|
| 573 |
+
"EarlyStoppingCallback": {
|
| 574 |
+
"args": {
|
| 575 |
+
"early_stopping_patience": 3,
|
| 576 |
+
"early_stopping_threshold": 0.0
|
| 577 |
+
},
|
| 578 |
+
"attributes": {
|
| 579 |
+
"early_stopping_patience_counter": 3
|
| 580 |
+
}
|
| 581 |
+
},
|
| 582 |
+
"TrainerControl": {
|
| 583 |
+
"args": {
|
| 584 |
+
"should_epoch_stop": false,
|
| 585 |
+
"should_evaluate": false,
|
| 586 |
+
"should_log": false,
|
| 587 |
+
"should_save": true,
|
| 588 |
+
"should_training_stop": true
|
| 589 |
+
},
|
| 590 |
+
"attributes": {}
|
| 591 |
+
}
|
| 592 |
+
},
|
| 593 |
+
"total_flos": 0.0,
|
| 594 |
+
"train_batch_size": 8,
|
| 595 |
+
"trial_name": null,
|
| 596 |
+
"trial_params": null
|
| 597 |
+
}
|
ood-vifactcheck-phobert-source-source-00-s42-20260826T150739Z-00ca7096/trainer/checkpoint-1638/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-source-source-01-s42-20260826T151112Z-00ca7096/trainer/checkpoint-1026/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-source-source-01-s42-20260826T151112Z-00ca7096/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-s42-20260826T151112Z-00ca7096/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-s42-20260826T151112Z-00ca7096/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-s42-20260826T151112Z-00ca7096/trainer/checkpoint-1026/trainer_state.json
ADDED
|
@@ -0,0 +1,396 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 513,
|
| 3 |
+
"best_metric": 0.8376747116852657,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-source-source-01-s42-20260826T151112Z-00ca7096/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": 8.040509223937988,
|
| 15 |
+
"learning_rate": 1.4035087719298246e-06,
|
| 16 |
+
"loss": 4.3976,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.2936857562408223,
|
| 21 |
+
"grad_norm": 8.21379280090332,
|
| 22 |
+
"learning_rate": 2.865497076023392e-06,
|
| 23 |
+
"loss": 4.4125,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.44052863436123346,
|
| 28 |
+
"grad_norm": 7.445639133453369,
|
| 29 |
+
"learning_rate": 4.3274853801169596e-06,
|
| 30 |
+
"loss": 4.4092,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5873715124816447,
|
| 35 |
+
"grad_norm": 6.265374183654785,
|
| 36 |
+
"learning_rate": 5.789473684210527e-06,
|
| 37 |
+
"loss": 4.3707,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.7342143906020558,
|
| 42 |
+
"grad_norm": 8.169437408447266,
|
| 43 |
+
"learning_rate": 7.251461988304094e-06,
|
| 44 |
+
"loss": 4.3017,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8810572687224669,
|
| 49 |
+
"grad_norm": 10.967144966125488,
|
| 50 |
+
"learning_rate": 8.713450292397661e-06,
|
| 51 |
+
"loss": 4.2093,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 1.0,
|
| 56 |
+
"eval_accuracy": 0.5587188612099644,
|
| 57 |
+
"eval_loss": 0.9680792689323425,
|
| 58 |
+
"eval_macro_f1": 0.5427663304919622,
|
| 59 |
+
"eval_runtime": 0.4105,
|
| 60 |
+
"eval_samples_per_second": 1368.977,
|
| 61 |
+
"eval_steps_per_second": 87.692,
|
| 62 |
+
"eval_weighted_f1": 0.5453519685928437,
|
| 63 |
+
"step": 171
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 1.0234948604992657,
|
| 67 |
+
"grad_norm": 12.033961296081543,
|
| 68 |
+
"learning_rate": 1.017543859649123e-05,
|
| 69 |
+
"loss": 3.8961,
|
| 70 |
+
"step": 175
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.170337738619677,
|
| 74 |
+
"grad_norm": 22.953153610229492,
|
| 75 |
+
"learning_rate": 1.1637426900584797e-05,
|
| 76 |
+
"loss": 3.6748,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.3171806167400881,
|
| 81 |
+
"grad_norm": 29.5552978515625,
|
| 82 |
+
"learning_rate": 1.3099415204678362e-05,
|
| 83 |
+
"loss": 3.4525,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.4640234948604993,
|
| 88 |
+
"grad_norm": 55.18943405151367,
|
| 89 |
+
"learning_rate": 1.4561403508771931e-05,
|
| 90 |
+
"loss": 3.2086,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.6108663729809103,
|
| 95 |
+
"grad_norm": 23.347143173217773,
|
| 96 |
+
"learning_rate": 1.60233918128655e-05,
|
| 97 |
+
"loss": 2.8791,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.7577092511013217,
|
| 102 |
+
"grad_norm": 31.090740203857422,
|
| 103 |
+
"learning_rate": 1.7485380116959064e-05,
|
| 104 |
+
"loss": 2.391,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.9045521292217327,
|
| 109 |
+
"grad_norm": 36.05583953857422,
|
| 110 |
+
"learning_rate": 1.894736842105263e-05,
|
| 111 |
+
"loss": 2.3446,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 2.0,
|
| 116 |
+
"eval_accuracy": 0.7900355871886121,
|
| 117 |
+
"eval_loss": 0.5226938128471375,
|
| 118 |
+
"eval_macro_f1": 0.7877885278223689,
|
| 119 |
+
"eval_runtime": 0.4419,
|
| 120 |
+
"eval_samples_per_second": 1271.888,
|
| 121 |
+
"eval_steps_per_second": 81.473,
|
| 122 |
+
"eval_weighted_f1": 0.7887050657511485,
|
| 123 |
+
"step": 342
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 2.0469897209985315,
|
| 127 |
+
"grad_norm": 35.47526168823242,
|
| 128 |
+
"learning_rate": 1.99545159194282e-05,
|
| 129 |
+
"loss": 1.977,
|
| 130 |
+
"step": 350
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.193832599118943,
|
| 134 |
+
"grad_norm": 28.79005241394043,
|
| 135 |
+
"learning_rate": 1.9792072774528916e-05,
|
| 136 |
+
"loss": 1.8711,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.340675477239354,
|
| 141 |
+
"grad_norm": 34.664554595947266,
|
| 142 |
+
"learning_rate": 1.962962962962963e-05,
|
| 143 |
+
"loss": 1.8345,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.4875183553597653,
|
| 148 |
+
"grad_norm": 41.97056579589844,
|
| 149 |
+
"learning_rate": 1.9467186484730346e-05,
|
| 150 |
+
"loss": 1.8327,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.6343612334801763,
|
| 155 |
+
"grad_norm": 48.226043701171875,
|
| 156 |
+
"learning_rate": 1.930474333983106e-05,
|
| 157 |
+
"loss": 1.5421,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.7812041116005872,
|
| 162 |
+
"grad_norm": 43.560482025146484,
|
| 163 |
+
"learning_rate": 1.9142300194931776e-05,
|
| 164 |
+
"loss": 1.8928,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.9280469897209986,
|
| 169 |
+
"grad_norm": 24.29937171936035,
|
| 170 |
+
"learning_rate": 1.897985705003249e-05,
|
| 171 |
+
"loss": 1.4658,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 3.0,
|
| 176 |
+
"eval_accuracy": 0.8380782918149466,
|
| 177 |
+
"eval_loss": 0.45560288429260254,
|
| 178 |
+
"eval_macro_f1": 0.8376747116852657,
|
| 179 |
+
"eval_runtime": 0.4143,
|
| 180 |
+
"eval_samples_per_second": 1356.658,
|
| 181 |
+
"eval_steps_per_second": 86.903,
|
| 182 |
+
"eval_weighted_f1": 0.8378847133861265,
|
| 183 |
+
"step": 513
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 3.0704845814977975,
|
| 187 |
+
"grad_norm": 61.14263916015625,
|
| 188 |
+
"learning_rate": 1.8817413905133203e-05,
|
| 189 |
+
"loss": 1.3769,
|
| 190 |
+
"step": 525
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.2173274596182084,
|
| 194 |
+
"grad_norm": 47.20503234863281,
|
| 195 |
+
"learning_rate": 1.8654970760233918e-05,
|
| 196 |
+
"loss": 1.0785,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.36417033773862,
|
| 201 |
+
"grad_norm": 107.21454620361328,
|
| 202 |
+
"learning_rate": 1.8492527615334633e-05,
|
| 203 |
+
"loss": 0.9517,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.511013215859031,
|
| 208 |
+
"grad_norm": 40.34237289428711,
|
| 209 |
+
"learning_rate": 1.833008447043535e-05,
|
| 210 |
+
"loss": 1.3178,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.657856093979442,
|
| 215 |
+
"grad_norm": 63.67678451538086,
|
| 216 |
+
"learning_rate": 1.8167641325536064e-05,
|
| 217 |
+
"loss": 1.0918,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.804698972099853,
|
| 222 |
+
"grad_norm": 26.53922462463379,
|
| 223 |
+
"learning_rate": 1.800519818063678e-05,
|
| 224 |
+
"loss": 0.9443,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.951541850220264,
|
| 229 |
+
"grad_norm": 73.66229248046875,
|
| 230 |
+
"learning_rate": 1.7842755035737494e-05,
|
| 231 |
+
"loss": 1.1101,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 4.0,
|
| 236 |
+
"eval_accuracy": 0.8327402135231317,
|
| 237 |
+
"eval_loss": 0.5146844983100891,
|
| 238 |
+
"eval_macro_f1": 0.8333633418090097,
|
| 239 |
+
"eval_runtime": 0.3927,
|
| 240 |
+
"eval_samples_per_second": 1431.002,
|
| 241 |
+
"eval_steps_per_second": 91.666,
|
| 242 |
+
"eval_weighted_f1": 0.833285542087449,
|
| 243 |
+
"step": 684
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 4.093979441997063,
|
| 247 |
+
"grad_norm": 24.502891540527344,
|
| 248 |
+
"learning_rate": 1.768031189083821e-05,
|
| 249 |
+
"loss": 1.0206,
|
| 250 |
+
"step": 700
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 4.240822320117474,
|
| 254 |
+
"grad_norm": 60.3140754699707,
|
| 255 |
+
"learning_rate": 1.751786874593892e-05,
|
| 256 |
+
"loss": 0.7447,
|
| 257 |
+
"step": 725
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.387665198237886,
|
| 261 |
+
"grad_norm": 12.493896484375,
|
| 262 |
+
"learning_rate": 1.7355425601039636e-05,
|
| 263 |
+
"loss": 0.7799,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.534508076358296,
|
| 268 |
+
"grad_norm": 58.10353088378906,
|
| 269 |
+
"learning_rate": 1.719298245614035e-05,
|
| 270 |
+
"loss": 0.6645,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.681350954478708,
|
| 275 |
+
"grad_norm": 6.0228800773620605,
|
| 276 |
+
"learning_rate": 1.7030539311241066e-05,
|
| 277 |
+
"loss": 0.9136,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.828193832599119,
|
| 282 |
+
"grad_norm": 21.521942138671875,
|
| 283 |
+
"learning_rate": 1.686809616634178e-05,
|
| 284 |
+
"loss": 0.8031,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.975036710719531,
|
| 289 |
+
"grad_norm": 44.193153381347656,
|
| 290 |
+
"learning_rate": 1.6705653021442497e-05,
|
| 291 |
+
"loss": 0.6747,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 5.0,
|
| 296 |
+
"eval_accuracy": 0.8274021352313167,
|
| 297 |
+
"eval_loss": 0.6087470650672913,
|
| 298 |
+
"eval_macro_f1": 0.8264231031607672,
|
| 299 |
+
"eval_runtime": 0.4786,
|
| 300 |
+
"eval_samples_per_second": 1174.249,
|
| 301 |
+
"eval_steps_per_second": 75.219,
|
| 302 |
+
"eval_weighted_f1": 0.8265146269468361,
|
| 303 |
+
"step": 855
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 5.117474302496329,
|
| 307 |
+
"grad_norm": 65.67449951171875,
|
| 308 |
+
"learning_rate": 1.654320987654321e-05,
|
| 309 |
+
"loss": 0.5932,
|
| 310 |
+
"step": 875
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 5.26431718061674,
|
| 314 |
+
"grad_norm": 53.35767364501953,
|
| 315 |
+
"learning_rate": 1.6380766731643927e-05,
|
| 316 |
+
"loss": 0.5874,
|
| 317 |
+
"step": 900
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.411160058737151,
|
| 321 |
+
"grad_norm": 20.13292694091797,
|
| 322 |
+
"learning_rate": 1.621832358674464e-05,
|
| 323 |
+
"loss": 0.4872,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.558002936857562,
|
| 328 |
+
"grad_norm": 16.55441665649414,
|
| 329 |
+
"learning_rate": 1.6055880441845354e-05,
|
| 330 |
+
"loss": 0.4874,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.704845814977974,
|
| 335 |
+
"grad_norm": 0.5804534554481506,
|
| 336 |
+
"learning_rate": 1.589343729694607e-05,
|
| 337 |
+
"loss": 0.5545,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.851688693098385,
|
| 342 |
+
"grad_norm": 56.35222625732422,
|
| 343 |
+
"learning_rate": 1.5730994152046787e-05,
|
| 344 |
+
"loss": 0.6182,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.998531571218796,
|
| 349 |
+
"grad_norm": 58.47371292114258,
|
| 350 |
+
"learning_rate": 1.55685510071475e-05,
|
| 351 |
+
"loss": 0.5293,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 6.0,
|
| 356 |
+
"eval_accuracy": 0.8238434163701067,
|
| 357 |
+
"eval_loss": 0.6401006579399109,
|
| 358 |
+
"eval_macro_f1": 0.823553548860387,
|
| 359 |
+
"eval_runtime": 0.4497,
|
| 360 |
+
"eval_samples_per_second": 1249.775,
|
| 361 |
+
"eval_steps_per_second": 80.057,
|
| 362 |
+
"eval_weighted_f1": 0.8236468309857787,
|
| 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-s42-20260826T151112Z-00ca7096/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-s42-20260826T151112Z-00ca7096/trainer/checkpoint-513/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-source-source-01-s42-20260826T151112Z-00ca7096/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-s42-20260826T151112Z-00ca7096/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-s42-20260826T151112Z-00ca7096/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-s42-20260826T151112Z-00ca7096/trainer/checkpoint-513/trainer_state.json
ADDED
|
@@ -0,0 +1,216 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 513,
|
| 3 |
+
"best_metric": 0.8376747116852657,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-source-source-01-s42-20260826T151112Z-00ca7096/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": 8.040509223937988,
|
| 15 |
+
"learning_rate": 1.4035087719298246e-06,
|
| 16 |
+
"loss": 4.3976,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.2936857562408223,
|
| 21 |
+
"grad_norm": 8.21379280090332,
|
| 22 |
+
"learning_rate": 2.865497076023392e-06,
|
| 23 |
+
"loss": 4.4125,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.44052863436123346,
|
| 28 |
+
"grad_norm": 7.445639133453369,
|
| 29 |
+
"learning_rate": 4.3274853801169596e-06,
|
| 30 |
+
"loss": 4.4092,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5873715124816447,
|
| 35 |
+
"grad_norm": 6.265374183654785,
|
| 36 |
+
"learning_rate": 5.789473684210527e-06,
|
| 37 |
+
"loss": 4.3707,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.7342143906020558,
|
| 42 |
+
"grad_norm": 8.169437408447266,
|
| 43 |
+
"learning_rate": 7.251461988304094e-06,
|
| 44 |
+
"loss": 4.3017,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8810572687224669,
|
| 49 |
+
"grad_norm": 10.967144966125488,
|
| 50 |
+
"learning_rate": 8.713450292397661e-06,
|
| 51 |
+
"loss": 4.2093,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 1.0,
|
| 56 |
+
"eval_accuracy": 0.5587188612099644,
|
| 57 |
+
"eval_loss": 0.9680792689323425,
|
| 58 |
+
"eval_macro_f1": 0.5427663304919622,
|
| 59 |
+
"eval_runtime": 0.4105,
|
| 60 |
+
"eval_samples_per_second": 1368.977,
|
| 61 |
+
"eval_steps_per_second": 87.692,
|
| 62 |
+
"eval_weighted_f1": 0.5453519685928437,
|
| 63 |
+
"step": 171
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 1.0234948604992657,
|
| 67 |
+
"grad_norm": 12.033961296081543,
|
| 68 |
+
"learning_rate": 1.017543859649123e-05,
|
| 69 |
+
"loss": 3.8961,
|
| 70 |
+
"step": 175
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.170337738619677,
|
| 74 |
+
"grad_norm": 22.953153610229492,
|
| 75 |
+
"learning_rate": 1.1637426900584797e-05,
|
| 76 |
+
"loss": 3.6748,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.3171806167400881,
|
| 81 |
+
"grad_norm": 29.5552978515625,
|
| 82 |
+
"learning_rate": 1.3099415204678362e-05,
|
| 83 |
+
"loss": 3.4525,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.4640234948604993,
|
| 88 |
+
"grad_norm": 55.18943405151367,
|
| 89 |
+
"learning_rate": 1.4561403508771931e-05,
|
| 90 |
+
"loss": 3.2086,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.6108663729809103,
|
| 95 |
+
"grad_norm": 23.347143173217773,
|
| 96 |
+
"learning_rate": 1.60233918128655e-05,
|
| 97 |
+
"loss": 2.8791,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.7577092511013217,
|
| 102 |
+
"grad_norm": 31.090740203857422,
|
| 103 |
+
"learning_rate": 1.7485380116959064e-05,
|
| 104 |
+
"loss": 2.391,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.9045521292217327,
|
| 109 |
+
"grad_norm": 36.05583953857422,
|
| 110 |
+
"learning_rate": 1.894736842105263e-05,
|
| 111 |
+
"loss": 2.3446,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 2.0,
|
| 116 |
+
"eval_accuracy": 0.7900355871886121,
|
| 117 |
+
"eval_loss": 0.5226938128471375,
|
| 118 |
+
"eval_macro_f1": 0.7877885278223689,
|
| 119 |
+
"eval_runtime": 0.4419,
|
| 120 |
+
"eval_samples_per_second": 1271.888,
|
| 121 |
+
"eval_steps_per_second": 81.473,
|
| 122 |
+
"eval_weighted_f1": 0.7887050657511485,
|
| 123 |
+
"step": 342
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 2.0469897209985315,
|
| 127 |
+
"grad_norm": 35.47526168823242,
|
| 128 |
+
"learning_rate": 1.99545159194282e-05,
|
| 129 |
+
"loss": 1.977,
|
| 130 |
+
"step": 350
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.193832599118943,
|
| 134 |
+
"grad_norm": 28.79005241394043,
|
| 135 |
+
"learning_rate": 1.9792072774528916e-05,
|
| 136 |
+
"loss": 1.8711,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.340675477239354,
|
| 141 |
+
"grad_norm": 34.664554595947266,
|
| 142 |
+
"learning_rate": 1.962962962962963e-05,
|
| 143 |
+
"loss": 1.8345,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.4875183553597653,
|
| 148 |
+
"grad_norm": 41.97056579589844,
|
| 149 |
+
"learning_rate": 1.9467186484730346e-05,
|
| 150 |
+
"loss": 1.8327,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.6343612334801763,
|
| 155 |
+
"grad_norm": 48.226043701171875,
|
| 156 |
+
"learning_rate": 1.930474333983106e-05,
|
| 157 |
+
"loss": 1.5421,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.7812041116005872,
|
| 162 |
+
"grad_norm": 43.560482025146484,
|
| 163 |
+
"learning_rate": 1.9142300194931776e-05,
|
| 164 |
+
"loss": 1.8928,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.9280469897209986,
|
| 169 |
+
"grad_norm": 24.29937171936035,
|
| 170 |
+
"learning_rate": 1.897985705003249e-05,
|
| 171 |
+
"loss": 1.4658,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 3.0,
|
| 176 |
+
"eval_accuracy": 0.8380782918149466,
|
| 177 |
+
"eval_loss": 0.45560288429260254,
|
| 178 |
+
"eval_macro_f1": 0.8376747116852657,
|
| 179 |
+
"eval_runtime": 0.4143,
|
| 180 |
+
"eval_samples_per_second": 1356.658,
|
| 181 |
+
"eval_steps_per_second": 86.903,
|
| 182 |
+
"eval_weighted_f1": 0.8378847133861265,
|
| 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-s42-20260826T151112Z-00ca7096/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-s42-20260826T151325Z-00ca7096/trainer/checkpoint-1086/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-source-source-02-s42-20260826T151325Z-00ca7096/trainer/checkpoint-1086/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-source-source-02-s42-20260826T151325Z-00ca7096/trainer/checkpoint-1086/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-s42-20260826T151325Z-00ca7096/trainer/checkpoint-1086/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-s42-20260826T151325Z-00ca7096/trainer/checkpoint-1086/trainer_state.json
ADDED
|
@@ -0,0 +1,410 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1086,
|
| 3 |
+
"best_metric": 0.8688248539452187,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-source-source-02-s42-20260826T151325Z-00ca7096/trainer/checkpoint-1086",
|
| 5 |
+
"epoch": 6.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1086,
|
| 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": 8.235882759094238,
|
| 15 |
+
"learning_rate": 1.3259668508287293e-06,
|
| 16 |
+
"loss": 4.4025,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.27624309392265195,
|
| 21 |
+
"grad_norm": 7.7691545486450195,
|
| 22 |
+
"learning_rate": 2.707182320441989e-06,
|
| 23 |
+
"loss": 4.4426,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.4143646408839779,
|
| 28 |
+
"grad_norm": 8.282538414001465,
|
| 29 |
+
"learning_rate": 4.088397790055249e-06,
|
| 30 |
+
"loss": 4.3814,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5524861878453039,
|
| 35 |
+
"grad_norm": 6.672089099884033,
|
| 36 |
+
"learning_rate": 5.469613259668509e-06,
|
| 37 |
+
"loss": 4.3904,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6906077348066298,
|
| 42 |
+
"grad_norm": 6.964391708374023,
|
| 43 |
+
"learning_rate": 6.850828729281769e-06,
|
| 44 |
+
"loss": 4.3592,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8287292817679558,
|
| 49 |
+
"grad_norm": 9.143580436706543,
|
| 50 |
+
"learning_rate": 8.232044198895029e-06,
|
| 51 |
+
"loss": 4.2893,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9668508287292817,
|
| 56 |
+
"grad_norm": 20.1214599609375,
|
| 57 |
+
"learning_rate": 9.613259668508288e-06,
|
| 58 |
+
"loss": 4.1041,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.570446735395189,
|
| 64 |
+
"eval_loss": 0.9425908923149109,
|
| 65 |
+
"eval_macro_f1": 0.5624859767487972,
|
| 66 |
+
"eval_runtime": 0.4425,
|
| 67 |
+
"eval_samples_per_second": 1315.313,
|
| 68 |
+
"eval_steps_per_second": 83.62,
|
| 69 |
+
"eval_weighted_f1": 0.5655470497527066,
|
| 70 |
+
"step": 181
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.1049723756906078,
|
| 74 |
+
"grad_norm": 15.14535903930664,
|
| 75 |
+
"learning_rate": 1.0994475138121547e-05,
|
| 76 |
+
"loss": 3.7434,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.2430939226519337,
|
| 81 |
+
"grad_norm": 29.24941635131836,
|
| 82 |
+
"learning_rate": 1.2375690607734806e-05,
|
| 83 |
+
"loss": 3.3674,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.3812154696132597,
|
| 88 |
+
"grad_norm": 29.76244354248047,
|
| 89 |
+
"learning_rate": 1.3756906077348067e-05,
|
| 90 |
+
"loss": 3.2089,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.5193370165745856,
|
| 95 |
+
"grad_norm": 36.082275390625,
|
| 96 |
+
"learning_rate": 1.5138121546961326e-05,
|
| 97 |
+
"loss": 2.7138,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.6574585635359116,
|
| 102 |
+
"grad_norm": 31.891477584838867,
|
| 103 |
+
"learning_rate": 1.6519337016574586e-05,
|
| 104 |
+
"loss": 2.7224,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7955801104972375,
|
| 109 |
+
"grad_norm": 31.563203811645508,
|
| 110 |
+
"learning_rate": 1.7900552486187846e-05,
|
| 111 |
+
"loss": 2.2393,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.9337016574585635,
|
| 116 |
+
"grad_norm": 53.246212005615234,
|
| 117 |
+
"learning_rate": 1.9281767955801107e-05,
|
| 118 |
+
"loss": 2.2948,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.8402061855670103,
|
| 124 |
+
"eval_loss": 0.4632522165775299,
|
| 125 |
+
"eval_macro_f1": 0.8372743995427464,
|
| 126 |
+
"eval_runtime": 0.4452,
|
| 127 |
+
"eval_samples_per_second": 1307.306,
|
| 128 |
+
"eval_steps_per_second": 83.111,
|
| 129 |
+
"eval_weighted_f1": 0.8378945019163432,
|
| 130 |
+
"step": 362
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.0718232044198897,
|
| 134 |
+
"grad_norm": 50.67064666748047,
|
| 135 |
+
"learning_rate": 1.9926335174953962e-05,
|
| 136 |
+
"loss": 1.8427,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.2099447513812156,
|
| 141 |
+
"grad_norm": 59.71549606323242,
|
| 142 |
+
"learning_rate": 1.9772866789441377e-05,
|
| 143 |
+
"loss": 1.6347,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.3480662983425415,
|
| 148 |
+
"grad_norm": 53.55520248413086,
|
| 149 |
+
"learning_rate": 1.9619398403928794e-05,
|
| 150 |
+
"loss": 1.7446,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.4861878453038675,
|
| 155 |
+
"grad_norm": 26.377918243408203,
|
| 156 |
+
"learning_rate": 1.946593001841621e-05,
|
| 157 |
+
"loss": 1.6044,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.6243093922651934,
|
| 162 |
+
"grad_norm": 74.2736587524414,
|
| 163 |
+
"learning_rate": 1.9312461632903623e-05,
|
| 164 |
+
"loss": 1.5534,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.7624309392265194,
|
| 169 |
+
"grad_norm": 61.817203521728516,
|
| 170 |
+
"learning_rate": 1.915899324739104e-05,
|
| 171 |
+
"loss": 1.6098,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.9005524861878453,
|
| 176 |
+
"grad_norm": 39.13019943237305,
|
| 177 |
+
"learning_rate": 1.9005524861878455e-05,
|
| 178 |
+
"loss": 1.654,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 3.0,
|
| 183 |
+
"eval_accuracy": 0.845360824742268,
|
| 184 |
+
"eval_loss": 0.45580676198005676,
|
| 185 |
+
"eval_macro_f1": 0.8427808567735955,
|
| 186 |
+
"eval_runtime": 0.4167,
|
| 187 |
+
"eval_samples_per_second": 1396.818,
|
| 188 |
+
"eval_steps_per_second": 88.801,
|
| 189 |
+
"eval_weighted_f1": 0.8435995679422058,
|
| 190 |
+
"step": 543
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.0386740331491713,
|
| 194 |
+
"grad_norm": 49.5412712097168,
|
| 195 |
+
"learning_rate": 1.885205647636587e-05,
|
| 196 |
+
"loss": 1.268,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.1767955801104972,
|
| 201 |
+
"grad_norm": 27.812705993652344,
|
| 202 |
+
"learning_rate": 1.8698588090853287e-05,
|
| 203 |
+
"loss": 1.1552,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.314917127071823,
|
| 208 |
+
"grad_norm": 131.9677276611328,
|
| 209 |
+
"learning_rate": 1.85451197053407e-05,
|
| 210 |
+
"loss": 1.1495,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.453038674033149,
|
| 215 |
+
"grad_norm": 38.567710876464844,
|
| 216 |
+
"learning_rate": 1.8391651319828116e-05,
|
| 217 |
+
"loss": 1.1873,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.591160220994475,
|
| 222 |
+
"grad_norm": 31.839149475097656,
|
| 223 |
+
"learning_rate": 1.823818293431553e-05,
|
| 224 |
+
"loss": 1.0613,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.729281767955801,
|
| 229 |
+
"grad_norm": 53.34172821044922,
|
| 230 |
+
"learning_rate": 1.8084714548802948e-05,
|
| 231 |
+
"loss": 0.9166,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.867403314917127,
|
| 236 |
+
"grad_norm": 8.729403495788574,
|
| 237 |
+
"learning_rate": 1.7931246163290362e-05,
|
| 238 |
+
"loss": 1.1275,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 4.0,
|
| 243 |
+
"eval_accuracy": 0.8384879725085911,
|
| 244 |
+
"eval_loss": 0.5181735754013062,
|
| 245 |
+
"eval_macro_f1": 0.8361727325756103,
|
| 246 |
+
"eval_runtime": 0.4146,
|
| 247 |
+
"eval_samples_per_second": 1403.765,
|
| 248 |
+
"eval_steps_per_second": 89.243,
|
| 249 |
+
"eval_weighted_f1": 0.8371518889652836,
|
| 250 |
+
"step": 724
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 4.005524861878453,
|
| 254 |
+
"grad_norm": 20.16485023498535,
|
| 255 |
+
"learning_rate": 1.7777777777777777e-05,
|
| 256 |
+
"loss": 1.0104,
|
| 257 |
+
"step": 725
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.143646408839779,
|
| 261 |
+
"grad_norm": 36.119789123535156,
|
| 262 |
+
"learning_rate": 1.7624309392265194e-05,
|
| 263 |
+
"loss": 0.6214,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.281767955801105,
|
| 268 |
+
"grad_norm": 21.186372756958008,
|
| 269 |
+
"learning_rate": 1.747084100675261e-05,
|
| 270 |
+
"loss": 0.7556,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.419889502762431,
|
| 275 |
+
"grad_norm": 80.20263671875,
|
| 276 |
+
"learning_rate": 1.7317372621240026e-05,
|
| 277 |
+
"loss": 0.6993,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.558011049723757,
|
| 282 |
+
"grad_norm": 91.50184631347656,
|
| 283 |
+
"learning_rate": 1.716390423572744e-05,
|
| 284 |
+
"loss": 0.8695,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.696132596685083,
|
| 289 |
+
"grad_norm": 13.931330680847168,
|
| 290 |
+
"learning_rate": 1.701043585021486e-05,
|
| 291 |
+
"loss": 0.7255,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.834254143646409,
|
| 296 |
+
"grad_norm": 79.41664123535156,
|
| 297 |
+
"learning_rate": 1.6856967464702273e-05,
|
| 298 |
+
"loss": 0.7118,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.972375690607735,
|
| 303 |
+
"grad_norm": 69.54364013671875,
|
| 304 |
+
"learning_rate": 1.670349907918969e-05,
|
| 305 |
+
"loss": 0.857,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 5.0,
|
| 310 |
+
"eval_accuracy": 0.8161512027491409,
|
| 311 |
+
"eval_loss": 0.6130379438400269,
|
| 312 |
+
"eval_macro_f1": 0.8141454538715812,
|
| 313 |
+
"eval_runtime": 0.4702,
|
| 314 |
+
"eval_samples_per_second": 1237.752,
|
| 315 |
+
"eval_steps_per_second": 78.689,
|
| 316 |
+
"eval_weighted_f1": 0.8154391746081736,
|
| 317 |
+
"step": 905
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.110497237569061,
|
| 321 |
+
"grad_norm": 17.945775985717773,
|
| 322 |
+
"learning_rate": 1.6550030693677105e-05,
|
| 323 |
+
"loss": 0.6807,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.248618784530387,
|
| 328 |
+
"grad_norm": 31.37040901184082,
|
| 329 |
+
"learning_rate": 1.639656230816452e-05,
|
| 330 |
+
"loss": 0.5006,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.386740331491713,
|
| 335 |
+
"grad_norm": 89.98050689697266,
|
| 336 |
+
"learning_rate": 1.6243093922651937e-05,
|
| 337 |
+
"loss": 0.5164,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.524861878453039,
|
| 342 |
+
"grad_norm": 69.73043060302734,
|
| 343 |
+
"learning_rate": 1.608962553713935e-05,
|
| 344 |
+
"loss": 0.4229,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.662983425414365,
|
| 349 |
+
"grad_norm": 2.5452558994293213,
|
| 350 |
+
"learning_rate": 1.5936157151626766e-05,
|
| 351 |
+
"loss": 0.5956,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.801104972375691,
|
| 356 |
+
"grad_norm": 18.021512985229492,
|
| 357 |
+
"learning_rate": 1.5782688766114183e-05,
|
| 358 |
+
"loss": 0.6233,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.939226519337017,
|
| 363 |
+
"grad_norm": 30.5696964263916,
|
| 364 |
+
"learning_rate": 1.5629220380601598e-05,
|
| 365 |
+
"loss": 0.4306,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 6.0,
|
| 370 |
+
"eval_accuracy": 0.8694158075601375,
|
| 371 |
+
"eval_loss": 0.5467572212219238,
|
| 372 |
+
"eval_macro_f1": 0.8688248539452187,
|
| 373 |
+
"eval_runtime": 0.4769,
|
| 374 |
+
"eval_samples_per_second": 1220.285,
|
| 375 |
+
"eval_steps_per_second": 77.578,
|
| 376 |
+
"eval_weighted_f1": 0.86920019102957,
|
| 377 |
+
"step": 1086
|
| 378 |
+
}
|
| 379 |
+
],
|
| 380 |
+
"logging_steps": 25,
|
| 381 |
+
"max_steps": 3620,
|
| 382 |
+
"num_input_tokens_seen": 0,
|
| 383 |
+
"num_train_epochs": 20,
|
| 384 |
+
"save_steps": 500,
|
| 385 |
+
"stateful_callbacks": {
|
| 386 |
+
"EarlyStoppingCallback": {
|
| 387 |
+
"args": {
|
| 388 |
+
"early_stopping_patience": 3,
|
| 389 |
+
"early_stopping_threshold": 0.0
|
| 390 |
+
},
|
| 391 |
+
"attributes": {
|
| 392 |
+
"early_stopping_patience_counter": 0
|
| 393 |
+
}
|
| 394 |
+
},
|
| 395 |
+
"TrainerControl": {
|
| 396 |
+
"args": {
|
| 397 |
+
"should_epoch_stop": false,
|
| 398 |
+
"should_evaluate": false,
|
| 399 |
+
"should_log": false,
|
| 400 |
+
"should_save": true,
|
| 401 |
+
"should_training_stop": false
|
| 402 |
+
},
|
| 403 |
+
"attributes": {}
|
| 404 |
+
}
|
| 405 |
+
},
|
| 406 |
+
"total_flos": 0.0,
|
| 407 |
+
"train_batch_size": 8,
|
| 408 |
+
"trial_name": null,
|
| 409 |
+
"trial_params": null
|
| 410 |
+
}
|
ood-vifactcheck-phobert-source-source-02-s42-20260826T151325Z-00ca7096/trainer/checkpoint-1086/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-source-source-02-s42-20260826T151325Z-00ca7096/trainer/checkpoint-1629/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-source-source-02-s42-20260826T151325Z-00ca7096/trainer/checkpoint-1629/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-source-source-02-s42-20260826T151325Z-00ca7096/trainer/checkpoint-1629/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-s42-20260826T151325Z-00ca7096/trainer/checkpoint-1629/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-s42-20260826T151325Z-00ca7096/trainer/checkpoint-1629/trainer_state.json
ADDED
|
@@ -0,0 +1,597 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1086,
|
| 3 |
+
"best_metric": 0.8688248539452187,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-source-source-02-s42-20260826T151325Z-00ca7096/trainer/checkpoint-1086",
|
| 5 |
+
"epoch": 9.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1629,
|
| 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": 8.235882759094238,
|
| 15 |
+
"learning_rate": 1.3259668508287293e-06,
|
| 16 |
+
"loss": 4.4025,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.27624309392265195,
|
| 21 |
+
"grad_norm": 7.7691545486450195,
|
| 22 |
+
"learning_rate": 2.707182320441989e-06,
|
| 23 |
+
"loss": 4.4426,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.4143646408839779,
|
| 28 |
+
"grad_norm": 8.282538414001465,
|
| 29 |
+
"learning_rate": 4.088397790055249e-06,
|
| 30 |
+
"loss": 4.3814,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5524861878453039,
|
| 35 |
+
"grad_norm": 6.672089099884033,
|
| 36 |
+
"learning_rate": 5.469613259668509e-06,
|
| 37 |
+
"loss": 4.3904,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6906077348066298,
|
| 42 |
+
"grad_norm": 6.964391708374023,
|
| 43 |
+
"learning_rate": 6.850828729281769e-06,
|
| 44 |
+
"loss": 4.3592,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8287292817679558,
|
| 49 |
+
"grad_norm": 9.143580436706543,
|
| 50 |
+
"learning_rate": 8.232044198895029e-06,
|
| 51 |
+
"loss": 4.2893,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9668508287292817,
|
| 56 |
+
"grad_norm": 20.1214599609375,
|
| 57 |
+
"learning_rate": 9.613259668508288e-06,
|
| 58 |
+
"loss": 4.1041,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.570446735395189,
|
| 64 |
+
"eval_loss": 0.9425908923149109,
|
| 65 |
+
"eval_macro_f1": 0.5624859767487972,
|
| 66 |
+
"eval_runtime": 0.4425,
|
| 67 |
+
"eval_samples_per_second": 1315.313,
|
| 68 |
+
"eval_steps_per_second": 83.62,
|
| 69 |
+
"eval_weighted_f1": 0.5655470497527066,
|
| 70 |
+
"step": 181
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.1049723756906078,
|
| 74 |
+
"grad_norm": 15.14535903930664,
|
| 75 |
+
"learning_rate": 1.0994475138121547e-05,
|
| 76 |
+
"loss": 3.7434,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.2430939226519337,
|
| 81 |
+
"grad_norm": 29.24941635131836,
|
| 82 |
+
"learning_rate": 1.2375690607734806e-05,
|
| 83 |
+
"loss": 3.3674,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.3812154696132597,
|
| 88 |
+
"grad_norm": 29.76244354248047,
|
| 89 |
+
"learning_rate": 1.3756906077348067e-05,
|
| 90 |
+
"loss": 3.2089,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.5193370165745856,
|
| 95 |
+
"grad_norm": 36.082275390625,
|
| 96 |
+
"learning_rate": 1.5138121546961326e-05,
|
| 97 |
+
"loss": 2.7138,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.6574585635359116,
|
| 102 |
+
"grad_norm": 31.891477584838867,
|
| 103 |
+
"learning_rate": 1.6519337016574586e-05,
|
| 104 |
+
"loss": 2.7224,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7955801104972375,
|
| 109 |
+
"grad_norm": 31.563203811645508,
|
| 110 |
+
"learning_rate": 1.7900552486187846e-05,
|
| 111 |
+
"loss": 2.2393,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.9337016574585635,
|
| 116 |
+
"grad_norm": 53.246212005615234,
|
| 117 |
+
"learning_rate": 1.9281767955801107e-05,
|
| 118 |
+
"loss": 2.2948,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.8402061855670103,
|
| 124 |
+
"eval_loss": 0.4632522165775299,
|
| 125 |
+
"eval_macro_f1": 0.8372743995427464,
|
| 126 |
+
"eval_runtime": 0.4452,
|
| 127 |
+
"eval_samples_per_second": 1307.306,
|
| 128 |
+
"eval_steps_per_second": 83.111,
|
| 129 |
+
"eval_weighted_f1": 0.8378945019163432,
|
| 130 |
+
"step": 362
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.0718232044198897,
|
| 134 |
+
"grad_norm": 50.67064666748047,
|
| 135 |
+
"learning_rate": 1.9926335174953962e-05,
|
| 136 |
+
"loss": 1.8427,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.2099447513812156,
|
| 141 |
+
"grad_norm": 59.71549606323242,
|
| 142 |
+
"learning_rate": 1.9772866789441377e-05,
|
| 143 |
+
"loss": 1.6347,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.3480662983425415,
|
| 148 |
+
"grad_norm": 53.55520248413086,
|
| 149 |
+
"learning_rate": 1.9619398403928794e-05,
|
| 150 |
+
"loss": 1.7446,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.4861878453038675,
|
| 155 |
+
"grad_norm": 26.377918243408203,
|
| 156 |
+
"learning_rate": 1.946593001841621e-05,
|
| 157 |
+
"loss": 1.6044,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.6243093922651934,
|
| 162 |
+
"grad_norm": 74.2736587524414,
|
| 163 |
+
"learning_rate": 1.9312461632903623e-05,
|
| 164 |
+
"loss": 1.5534,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.7624309392265194,
|
| 169 |
+
"grad_norm": 61.817203521728516,
|
| 170 |
+
"learning_rate": 1.915899324739104e-05,
|
| 171 |
+
"loss": 1.6098,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.9005524861878453,
|
| 176 |
+
"grad_norm": 39.13019943237305,
|
| 177 |
+
"learning_rate": 1.9005524861878455e-05,
|
| 178 |
+
"loss": 1.654,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 3.0,
|
| 183 |
+
"eval_accuracy": 0.845360824742268,
|
| 184 |
+
"eval_loss": 0.45580676198005676,
|
| 185 |
+
"eval_macro_f1": 0.8427808567735955,
|
| 186 |
+
"eval_runtime": 0.4167,
|
| 187 |
+
"eval_samples_per_second": 1396.818,
|
| 188 |
+
"eval_steps_per_second": 88.801,
|
| 189 |
+
"eval_weighted_f1": 0.8435995679422058,
|
| 190 |
+
"step": 543
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.0386740331491713,
|
| 194 |
+
"grad_norm": 49.5412712097168,
|
| 195 |
+
"learning_rate": 1.885205647636587e-05,
|
| 196 |
+
"loss": 1.268,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.1767955801104972,
|
| 201 |
+
"grad_norm": 27.812705993652344,
|
| 202 |
+
"learning_rate": 1.8698588090853287e-05,
|
| 203 |
+
"loss": 1.1552,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.314917127071823,
|
| 208 |
+
"grad_norm": 131.9677276611328,
|
| 209 |
+
"learning_rate": 1.85451197053407e-05,
|
| 210 |
+
"loss": 1.1495,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.453038674033149,
|
| 215 |
+
"grad_norm": 38.567710876464844,
|
| 216 |
+
"learning_rate": 1.8391651319828116e-05,
|
| 217 |
+
"loss": 1.1873,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.591160220994475,
|
| 222 |
+
"grad_norm": 31.839149475097656,
|
| 223 |
+
"learning_rate": 1.823818293431553e-05,
|
| 224 |
+
"loss": 1.0613,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.729281767955801,
|
| 229 |
+
"grad_norm": 53.34172821044922,
|
| 230 |
+
"learning_rate": 1.8084714548802948e-05,
|
| 231 |
+
"loss": 0.9166,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.867403314917127,
|
| 236 |
+
"grad_norm": 8.729403495788574,
|
| 237 |
+
"learning_rate": 1.7931246163290362e-05,
|
| 238 |
+
"loss": 1.1275,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 4.0,
|
| 243 |
+
"eval_accuracy": 0.8384879725085911,
|
| 244 |
+
"eval_loss": 0.5181735754013062,
|
| 245 |
+
"eval_macro_f1": 0.8361727325756103,
|
| 246 |
+
"eval_runtime": 0.4146,
|
| 247 |
+
"eval_samples_per_second": 1403.765,
|
| 248 |
+
"eval_steps_per_second": 89.243,
|
| 249 |
+
"eval_weighted_f1": 0.8371518889652836,
|
| 250 |
+
"step": 724
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 4.005524861878453,
|
| 254 |
+
"grad_norm": 20.16485023498535,
|
| 255 |
+
"learning_rate": 1.7777777777777777e-05,
|
| 256 |
+
"loss": 1.0104,
|
| 257 |
+
"step": 725
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.143646408839779,
|
| 261 |
+
"grad_norm": 36.119789123535156,
|
| 262 |
+
"learning_rate": 1.7624309392265194e-05,
|
| 263 |
+
"loss": 0.6214,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.281767955801105,
|
| 268 |
+
"grad_norm": 21.186372756958008,
|
| 269 |
+
"learning_rate": 1.747084100675261e-05,
|
| 270 |
+
"loss": 0.7556,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.419889502762431,
|
| 275 |
+
"grad_norm": 80.20263671875,
|
| 276 |
+
"learning_rate": 1.7317372621240026e-05,
|
| 277 |
+
"loss": 0.6993,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.558011049723757,
|
| 282 |
+
"grad_norm": 91.50184631347656,
|
| 283 |
+
"learning_rate": 1.716390423572744e-05,
|
| 284 |
+
"loss": 0.8695,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.696132596685083,
|
| 289 |
+
"grad_norm": 13.931330680847168,
|
| 290 |
+
"learning_rate": 1.701043585021486e-05,
|
| 291 |
+
"loss": 0.7255,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.834254143646409,
|
| 296 |
+
"grad_norm": 79.41664123535156,
|
| 297 |
+
"learning_rate": 1.6856967464702273e-05,
|
| 298 |
+
"loss": 0.7118,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.972375690607735,
|
| 303 |
+
"grad_norm": 69.54364013671875,
|
| 304 |
+
"learning_rate": 1.670349907918969e-05,
|
| 305 |
+
"loss": 0.857,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 5.0,
|
| 310 |
+
"eval_accuracy": 0.8161512027491409,
|
| 311 |
+
"eval_loss": 0.6130379438400269,
|
| 312 |
+
"eval_macro_f1": 0.8141454538715812,
|
| 313 |
+
"eval_runtime": 0.4702,
|
| 314 |
+
"eval_samples_per_second": 1237.752,
|
| 315 |
+
"eval_steps_per_second": 78.689,
|
| 316 |
+
"eval_weighted_f1": 0.8154391746081736,
|
| 317 |
+
"step": 905
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.110497237569061,
|
| 321 |
+
"grad_norm": 17.945775985717773,
|
| 322 |
+
"learning_rate": 1.6550030693677105e-05,
|
| 323 |
+
"loss": 0.6807,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.248618784530387,
|
| 328 |
+
"grad_norm": 31.37040901184082,
|
| 329 |
+
"learning_rate": 1.639656230816452e-05,
|
| 330 |
+
"loss": 0.5006,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.386740331491713,
|
| 335 |
+
"grad_norm": 89.98050689697266,
|
| 336 |
+
"learning_rate": 1.6243093922651937e-05,
|
| 337 |
+
"loss": 0.5164,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.524861878453039,
|
| 342 |
+
"grad_norm": 69.73043060302734,
|
| 343 |
+
"learning_rate": 1.608962553713935e-05,
|
| 344 |
+
"loss": 0.4229,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.662983425414365,
|
| 349 |
+
"grad_norm": 2.5452558994293213,
|
| 350 |
+
"learning_rate": 1.5936157151626766e-05,
|
| 351 |
+
"loss": 0.5956,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.801104972375691,
|
| 356 |
+
"grad_norm": 18.021512985229492,
|
| 357 |
+
"learning_rate": 1.5782688766114183e-05,
|
| 358 |
+
"loss": 0.6233,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.939226519337017,
|
| 363 |
+
"grad_norm": 30.5696964263916,
|
| 364 |
+
"learning_rate": 1.5629220380601598e-05,
|
| 365 |
+
"loss": 0.4306,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 6.0,
|
| 370 |
+
"eval_accuracy": 0.8694158075601375,
|
| 371 |
+
"eval_loss": 0.5467572212219238,
|
| 372 |
+
"eval_macro_f1": 0.8688248539452187,
|
| 373 |
+
"eval_runtime": 0.4769,
|
| 374 |
+
"eval_samples_per_second": 1220.285,
|
| 375 |
+
"eval_steps_per_second": 77.578,
|
| 376 |
+
"eval_weighted_f1": 0.86920019102957,
|
| 377 |
+
"step": 1086
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 6.077348066298343,
|
| 381 |
+
"grad_norm": 7.524197101593018,
|
| 382 |
+
"learning_rate": 1.5475751995089012e-05,
|
| 383 |
+
"loss": 0.5234,
|
| 384 |
+
"step": 1100
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 6.2154696132596685,
|
| 388 |
+
"grad_norm": 77.01425170898438,
|
| 389 |
+
"learning_rate": 1.532228360957643e-05,
|
| 390 |
+
"loss": 0.3889,
|
| 391 |
+
"step": 1125
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 6.3535911602209945,
|
| 395 |
+
"grad_norm": 154.5743408203125,
|
| 396 |
+
"learning_rate": 1.5168815224063844e-05,
|
| 397 |
+
"loss": 0.4862,
|
| 398 |
+
"step": 1150
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.49171270718232,
|
| 402 |
+
"grad_norm": 11.330290794372559,
|
| 403 |
+
"learning_rate": 1.5015346838551258e-05,
|
| 404 |
+
"loss": 0.462,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.629834254143646,
|
| 409 |
+
"grad_norm": 98.41529846191406,
|
| 410 |
+
"learning_rate": 1.4861878453038676e-05,
|
| 411 |
+
"loss": 0.3809,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.767955801104972,
|
| 416 |
+
"grad_norm": 5.612950801849365,
|
| 417 |
+
"learning_rate": 1.470841006752609e-05,
|
| 418 |
+
"loss": 0.3465,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.906077348066298,
|
| 423 |
+
"grad_norm": 7.889017105102539,
|
| 424 |
+
"learning_rate": 1.4554941682013506e-05,
|
| 425 |
+
"loss": 0.3639,
|
| 426 |
+
"step": 1250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 7.0,
|
| 430 |
+
"eval_accuracy": 0.8281786941580757,
|
| 431 |
+
"eval_loss": 0.8006986975669861,
|
| 432 |
+
"eval_macro_f1": 0.8275514385059887,
|
| 433 |
+
"eval_runtime": 0.4485,
|
| 434 |
+
"eval_samples_per_second": 1297.777,
|
| 435 |
+
"eval_steps_per_second": 82.505,
|
| 436 |
+
"eval_weighted_f1": 0.8281940452258545,
|
| 437 |
+
"step": 1267
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 7.044198895027624,
|
| 441 |
+
"grad_norm": 1.9094690084457397,
|
| 442 |
+
"learning_rate": 1.4401473296500922e-05,
|
| 443 |
+
"loss": 0.3677,
|
| 444 |
+
"step": 1275
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 7.18232044198895,
|
| 448 |
+
"grad_norm": 168.1240997314453,
|
| 449 |
+
"learning_rate": 1.4248004910988338e-05,
|
| 450 |
+
"loss": 0.3471,
|
| 451 |
+
"step": 1300
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 7.320441988950276,
|
| 455 |
+
"grad_norm": 1.873695969581604,
|
| 456 |
+
"learning_rate": 1.4094536525475753e-05,
|
| 457 |
+
"loss": 0.3006,
|
| 458 |
+
"step": 1325
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 7.458563535911602,
|
| 462 |
+
"grad_norm": 0.673685610294342,
|
| 463 |
+
"learning_rate": 1.394106813996317e-05,
|
| 464 |
+
"loss": 0.277,
|
| 465 |
+
"step": 1350
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 7.596685082872928,
|
| 469 |
+
"grad_norm": 0.4115293622016907,
|
| 470 |
+
"learning_rate": 1.3787599754450585e-05,
|
| 471 |
+
"loss": 0.3148,
|
| 472 |
+
"step": 1375
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 7.734806629834254,
|
| 476 |
+
"grad_norm": 2.08632230758667,
|
| 477 |
+
"learning_rate": 1.3634131368937999e-05,
|
| 478 |
+
"loss": 0.1682,
|
| 479 |
+
"step": 1400
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 7.87292817679558,
|
| 483 |
+
"grad_norm": 1.732498288154602,
|
| 484 |
+
"learning_rate": 1.3480662983425417e-05,
|
| 485 |
+
"loss": 0.4065,
|
| 486 |
+
"step": 1425
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 8.0,
|
| 490 |
+
"eval_accuracy": 0.8505154639175257,
|
| 491 |
+
"eval_loss": 0.7524002194404602,
|
| 492 |
+
"eval_macro_f1": 0.8500380658436214,
|
| 493 |
+
"eval_runtime": 0.4394,
|
| 494 |
+
"eval_samples_per_second": 1324.555,
|
| 495 |
+
"eval_steps_per_second": 84.207,
|
| 496 |
+
"eval_weighted_f1": 0.8505489743752918,
|
| 497 |
+
"step": 1448
|
| 498 |
+
},
|
| 499 |
+
{
|
| 500 |
+
"epoch": 8.011049723756907,
|
| 501 |
+
"grad_norm": 0.2235078066587448,
|
| 502 |
+
"learning_rate": 1.3327194597912831e-05,
|
| 503 |
+
"loss": 0.2752,
|
| 504 |
+
"step": 1450
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 8.149171270718233,
|
| 508 |
+
"grad_norm": 0.2353101223707199,
|
| 509 |
+
"learning_rate": 1.3173726212400245e-05,
|
| 510 |
+
"loss": 0.192,
|
| 511 |
+
"step": 1475
|
| 512 |
+
},
|
| 513 |
+
{
|
| 514 |
+
"epoch": 8.287292817679559,
|
| 515 |
+
"grad_norm": 0.21296977996826172,
|
| 516 |
+
"learning_rate": 1.3020257826887662e-05,
|
| 517 |
+
"loss": 0.1918,
|
| 518 |
+
"step": 1500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 8.425414364640885,
|
| 522 |
+
"grad_norm": 0.39459121227264404,
|
| 523 |
+
"learning_rate": 1.2866789441375078e-05,
|
| 524 |
+
"loss": 0.1346,
|
| 525 |
+
"step": 1525
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 8.56353591160221,
|
| 529 |
+
"grad_norm": 0.4998416304588318,
|
| 530 |
+
"learning_rate": 1.2713321055862494e-05,
|
| 531 |
+
"loss": 0.1853,
|
| 532 |
+
"step": 1550
|
| 533 |
+
},
|
| 534 |
+
{
|
| 535 |
+
"epoch": 8.701657458563536,
|
| 536 |
+
"grad_norm": 0.18831989169120789,
|
| 537 |
+
"learning_rate": 1.2559852670349908e-05,
|
| 538 |
+
"loss": 0.1461,
|
| 539 |
+
"step": 1575
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 8.839779005524862,
|
| 543 |
+
"grad_norm": 0.2321416139602661,
|
| 544 |
+
"learning_rate": 1.2406384284837326e-05,
|
| 545 |
+
"loss": 0.3288,
|
| 546 |
+
"step": 1600
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 8.977900552486188,
|
| 550 |
+
"grad_norm": 0.24179372191429138,
|
| 551 |
+
"learning_rate": 1.225291589932474e-05,
|
| 552 |
+
"loss": 0.3007,
|
| 553 |
+
"step": 1625
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 9.0,
|
| 557 |
+
"eval_accuracy": 0.845360824742268,
|
| 558 |
+
"eval_loss": 0.8242730498313904,
|
| 559 |
+
"eval_macro_f1": 0.8439134921724922,
|
| 560 |
+
"eval_runtime": 0.4788,
|
| 561 |
+
"eval_samples_per_second": 1215.524,
|
| 562 |
+
"eval_steps_per_second": 77.276,
|
| 563 |
+
"eval_weighted_f1": 0.8446933131075129,
|
| 564 |
+
"step": 1629
|
| 565 |
+
}
|
| 566 |
+
],
|
| 567 |
+
"logging_steps": 25,
|
| 568 |
+
"max_steps": 3620,
|
| 569 |
+
"num_input_tokens_seen": 0,
|
| 570 |
+
"num_train_epochs": 20,
|
| 571 |
+
"save_steps": 500,
|
| 572 |
+
"stateful_callbacks": {
|
| 573 |
+
"EarlyStoppingCallback": {
|
| 574 |
+
"args": {
|
| 575 |
+
"early_stopping_patience": 3,
|
| 576 |
+
"early_stopping_threshold": 0.0
|
| 577 |
+
},
|
| 578 |
+
"attributes": {
|
| 579 |
+
"early_stopping_patience_counter": 3
|
| 580 |
+
}
|
| 581 |
+
},
|
| 582 |
+
"TrainerControl": {
|
| 583 |
+
"args": {
|
| 584 |
+
"should_epoch_stop": false,
|
| 585 |
+
"should_evaluate": false,
|
| 586 |
+
"should_log": false,
|
| 587 |
+
"should_save": true,
|
| 588 |
+
"should_training_stop": true
|
| 589 |
+
},
|
| 590 |
+
"attributes": {}
|
| 591 |
+
}
|
| 592 |
+
},
|
| 593 |
+
"total_flos": 0.0,
|
| 594 |
+
"train_batch_size": 8,
|
| 595 |
+
"trial_name": null,
|
| 596 |
+
"trial_params": null
|
| 597 |
+
}
|
ood-vifactcheck-phobert-source-source-02-s42-20260826T151325Z-00ca7096/trainer/checkpoint-1629/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-source-source-03-s42-20260826T151649Z-00ca7096/trainer/checkpoint-1464/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-source-source-03-s42-20260826T151649Z-00ca7096/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-s42-20260826T151649Z-00ca7096/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-s42-20260826T151649Z-00ca7096/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-s42-20260826T151649Z-00ca7096/trainer/checkpoint-1464/trainer_state.json
ADDED
|
@@ -0,0 +1,537 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1464,
|
| 3 |
+
"best_metric": 0.8658216687964972,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-source-source-03-s42-20260826T151649Z-00ca7096/trainer/checkpoint-1464",
|
| 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": 9.540091514587402,
|
| 15 |
+
"learning_rate": 1.3114754098360657e-06,
|
| 16 |
+
"loss": 4.4265,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.27434842249657065,
|
| 21 |
+
"grad_norm": 7.004621982574463,
|
| 22 |
+
"learning_rate": 2.677595628415301e-06,
|
| 23 |
+
"loss": 4.3974,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.411522633744856,
|
| 28 |
+
"grad_norm": 5.635507583618164,
|
| 29 |
+
"learning_rate": 4.043715846994536e-06,
|
| 30 |
+
"loss": 4.4071,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5486968449931413,
|
| 35 |
+
"grad_norm": 8.972846031188965,
|
| 36 |
+
"learning_rate": 5.409836065573772e-06,
|
| 37 |
+
"loss": 4.3508,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6858710562414266,
|
| 42 |
+
"grad_norm": 7.308180809020996,
|
| 43 |
+
"learning_rate": 6.775956284153006e-06,
|
| 44 |
+
"loss": 4.3224,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.823045267489712,
|
| 49 |
+
"grad_norm": 9.94659423828125,
|
| 50 |
+
"learning_rate": 8.14207650273224e-06,
|
| 51 |
+
"loss": 4.2761,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9602194787379973,
|
| 56 |
+
"grad_norm": 12.845196723937988,
|
| 57 |
+
"learning_rate": 9.508196721311476e-06,
|
| 58 |
+
"loss": 4.1545,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.5574043261231281,
|
| 64 |
+
"eval_loss": 0.9540154337882996,
|
| 65 |
+
"eval_macro_f1": 0.5295010108297173,
|
| 66 |
+
"eval_runtime": 0.4751,
|
| 67 |
+
"eval_samples_per_second": 1264.914,
|
| 68 |
+
"eval_steps_per_second": 79.978,
|
| 69 |
+
"eval_weighted_f1": 0.5327022775590124,
|
| 70 |
+
"step": 183
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.093278463648834,
|
| 74 |
+
"grad_norm": 24.179080963134766,
|
| 75 |
+
"learning_rate": 1.0874316939890712e-05,
|
| 76 |
+
"loss": 3.6165,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.2304526748971194,
|
| 81 |
+
"grad_norm": 39.38579177856445,
|
| 82 |
+
"learning_rate": 1.2240437158469946e-05,
|
| 83 |
+
"loss": 3.2323,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.3676268861454046,
|
| 88 |
+
"grad_norm": 32.005470275878906,
|
| 89 |
+
"learning_rate": 1.3606557377049181e-05,
|
| 90 |
+
"loss": 3.0398,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.50480109739369,
|
| 95 |
+
"grad_norm": 35.23550796508789,
|
| 96 |
+
"learning_rate": 1.4972677595628417e-05,
|
| 97 |
+
"loss": 3.0094,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.6419753086419753,
|
| 102 |
+
"grad_norm": 28.283357620239258,
|
| 103 |
+
"learning_rate": 1.633879781420765e-05,
|
| 104 |
+
"loss": 2.6878,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7791495198902605,
|
| 109 |
+
"grad_norm": 45.30286407470703,
|
| 110 |
+
"learning_rate": 1.7704918032786887e-05,
|
| 111 |
+
"loss": 2.6862,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.916323731138546,
|
| 116 |
+
"grad_norm": 20.715063095092773,
|
| 117 |
+
"learning_rate": 1.9071038251366122e-05,
|
| 118 |
+
"loss": 2.1346,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.8236272878535774,
|
| 124 |
+
"eval_loss": 0.47288697957992554,
|
| 125 |
+
"eval_macro_f1": 0.8235667369793007,
|
| 126 |
+
"eval_runtime": 0.4215,
|
| 127 |
+
"eval_samples_per_second": 1425.946,
|
| 128 |
+
"eval_steps_per_second": 90.16,
|
| 129 |
+
"eval_weighted_f1": 0.823587567393166,
|
| 130 |
+
"step": 366
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.049382716049383,
|
| 134 |
+
"grad_norm": 38.181583404541016,
|
| 135 |
+
"learning_rate": 1.995142683667274e-05,
|
| 136 |
+
"loss": 1.9687,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.186556927297668,
|
| 141 |
+
"grad_norm": 47.709129333496094,
|
| 142 |
+
"learning_rate": 1.9799635701275045e-05,
|
| 143 |
+
"loss": 1.609,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.3237311385459534,
|
| 148 |
+
"grad_norm": 39.530494689941406,
|
| 149 |
+
"learning_rate": 1.9647844565877354e-05,
|
| 150 |
+
"loss": 1.769,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.460905349794239,
|
| 155 |
+
"grad_norm": 39.43211364746094,
|
| 156 |
+
"learning_rate": 1.9496053430479663e-05,
|
| 157 |
+
"loss": 1.7113,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.598079561042524,
|
| 162 |
+
"grad_norm": 43.74077606201172,
|
| 163 |
+
"learning_rate": 1.934426229508197e-05,
|
| 164 |
+
"loss": 1.5929,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.7352537722908092,
|
| 169 |
+
"grad_norm": 39.347652435302734,
|
| 170 |
+
"learning_rate": 1.9192471159684274e-05,
|
| 171 |
+
"loss": 1.5992,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.8724279835390947,
|
| 176 |
+
"grad_norm": 35.88862609863281,
|
| 177 |
+
"learning_rate": 1.9040680024286583e-05,
|
| 178 |
+
"loss": 1.7626,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 3.0,
|
| 183 |
+
"eval_accuracy": 0.8552412645590682,
|
| 184 |
+
"eval_loss": 0.408374547958374,
|
| 185 |
+
"eval_macro_f1": 0.8559032398750395,
|
| 186 |
+
"eval_runtime": 0.4087,
|
| 187 |
+
"eval_samples_per_second": 1470.618,
|
| 188 |
+
"eval_steps_per_second": 92.984,
|
| 189 |
+
"eval_weighted_f1": 0.855575761505065,
|
| 190 |
+
"step": 549
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.0054869684499312,
|
| 194 |
+
"grad_norm": 37.14410400390625,
|
| 195 |
+
"learning_rate": 1.888888888888889e-05,
|
| 196 |
+
"loss": 1.4598,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.1426611796982167,
|
| 201 |
+
"grad_norm": 38.32145309448242,
|
| 202 |
+
"learning_rate": 1.8737097753491198e-05,
|
| 203 |
+
"loss": 1.2316,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.279835390946502,
|
| 208 |
+
"grad_norm": 40.79185104370117,
|
| 209 |
+
"learning_rate": 1.8585306618093504e-05,
|
| 210 |
+
"loss": 0.9759,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.4170096021947876,
|
| 215 |
+
"grad_norm": 24.409038543701172,
|
| 216 |
+
"learning_rate": 1.8433515482695813e-05,
|
| 217 |
+
"loss": 0.9108,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.5541838134430725,
|
| 222 |
+
"grad_norm": 75.9214859008789,
|
| 223 |
+
"learning_rate": 1.828172434729812e-05,
|
| 224 |
+
"loss": 1.1125,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.691358024691358,
|
| 229 |
+
"grad_norm": 24.009607315063477,
|
| 230 |
+
"learning_rate": 1.8129933211900424e-05,
|
| 231 |
+
"loss": 1.153,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.8285322359396434,
|
| 236 |
+
"grad_norm": 35.08176040649414,
|
| 237 |
+
"learning_rate": 1.7978142076502733e-05,
|
| 238 |
+
"loss": 1.3368,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.9657064471879284,
|
| 243 |
+
"grad_norm": 33.363094329833984,
|
| 244 |
+
"learning_rate": 1.7826350941105042e-05,
|
| 245 |
+
"loss": 1.1569,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 4.0,
|
| 250 |
+
"eval_accuracy": 0.8519134775374376,
|
| 251 |
+
"eval_loss": 0.4405615031719208,
|
| 252 |
+
"eval_macro_f1": 0.8520966541201226,
|
| 253 |
+
"eval_runtime": 0.4349,
|
| 254 |
+
"eval_samples_per_second": 1381.888,
|
| 255 |
+
"eval_steps_per_second": 87.374,
|
| 256 |
+
"eval_weighted_f1": 0.851633137685227,
|
| 257 |
+
"step": 732
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.098765432098766,
|
| 261 |
+
"grad_norm": 37.453216552734375,
|
| 262 |
+
"learning_rate": 1.7674559805707348e-05,
|
| 263 |
+
"loss": 0.83,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.23593964334705,
|
| 268 |
+
"grad_norm": 35.347137451171875,
|
| 269 |
+
"learning_rate": 1.7522768670309657e-05,
|
| 270 |
+
"loss": 0.723,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.373113854595336,
|
| 275 |
+
"grad_norm": 36.2155876159668,
|
| 276 |
+
"learning_rate": 1.7370977534911962e-05,
|
| 277 |
+
"loss": 0.7777,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.510288065843621,
|
| 282 |
+
"grad_norm": 38.25359344482422,
|
| 283 |
+
"learning_rate": 1.7219186399514268e-05,
|
| 284 |
+
"loss": 0.6556,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.647462277091907,
|
| 289 |
+
"grad_norm": 69.14533233642578,
|
| 290 |
+
"learning_rate": 1.7067395264116577e-05,
|
| 291 |
+
"loss": 0.8692,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.784636488340192,
|
| 296 |
+
"grad_norm": 12.52321720123291,
|
| 297 |
+
"learning_rate": 1.6915604128718883e-05,
|
| 298 |
+
"loss": 0.8442,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.921810699588478,
|
| 303 |
+
"grad_norm": 32.477474212646484,
|
| 304 |
+
"learning_rate": 1.6763812993321192e-05,
|
| 305 |
+
"loss": 0.7072,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 5.0,
|
| 310 |
+
"eval_accuracy": 0.8569051580698835,
|
| 311 |
+
"eval_loss": 0.4720706641674042,
|
| 312 |
+
"eval_macro_f1": 0.8572179129600631,
|
| 313 |
+
"eval_runtime": 0.4497,
|
| 314 |
+
"eval_samples_per_second": 1336.492,
|
| 315 |
+
"eval_steps_per_second": 84.504,
|
| 316 |
+
"eval_weighted_f1": 0.856788740570396,
|
| 317 |
+
"step": 915
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.054869684499314,
|
| 321 |
+
"grad_norm": 48.532066345214844,
|
| 322 |
+
"learning_rate": 1.66120218579235e-05,
|
| 323 |
+
"loss": 0.7139,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.1920438957476,
|
| 328 |
+
"grad_norm": 1.913240909576416,
|
| 329 |
+
"learning_rate": 1.6460230722525807e-05,
|
| 330 |
+
"loss": 0.5688,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.329218106995885,
|
| 335 |
+
"grad_norm": 13.59510326385498,
|
| 336 |
+
"learning_rate": 1.6308439587128112e-05,
|
| 337 |
+
"loss": 0.4244,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.4663923182441705,
|
| 342 |
+
"grad_norm": 103.00779724121094,
|
| 343 |
+
"learning_rate": 1.615664845173042e-05,
|
| 344 |
+
"loss": 0.3592,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.603566529492455,
|
| 349 |
+
"grad_norm": 26.168684005737305,
|
| 350 |
+
"learning_rate": 1.6004857316332727e-05,
|
| 351 |
+
"loss": 0.5435,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.7407407407407405,
|
| 356 |
+
"grad_norm": 19.721546173095703,
|
| 357 |
+
"learning_rate": 1.5853066180935036e-05,
|
| 358 |
+
"loss": 0.5746,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.877914951989026,
|
| 363 |
+
"grad_norm": 86.65348052978516,
|
| 364 |
+
"learning_rate": 1.570127504553734e-05,
|
| 365 |
+
"loss": 0.7417,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 6.0,
|
| 370 |
+
"eval_accuracy": 0.8569051580698835,
|
| 371 |
+
"eval_loss": 0.5869454145431519,
|
| 372 |
+
"eval_macro_f1": 0.8568009493221235,
|
| 373 |
+
"eval_runtime": 0.3844,
|
| 374 |
+
"eval_samples_per_second": 1563.431,
|
| 375 |
+
"eval_steps_per_second": 98.853,
|
| 376 |
+
"eval_weighted_f1": 0.8566164852264386,
|
| 377 |
+
"step": 1098
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 6.0109739368998625,
|
| 381 |
+
"grad_norm": 64.31432342529297,
|
| 382 |
+
"learning_rate": 1.5549483910139647e-05,
|
| 383 |
+
"loss": 0.6296,
|
| 384 |
+
"step": 1100
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 6.148148148148148,
|
| 388 |
+
"grad_norm": 2.6430726051330566,
|
| 389 |
+
"learning_rate": 1.5397692774741956e-05,
|
| 390 |
+
"loss": 0.3162,
|
| 391 |
+
"step": 1125
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 6.285322359396433,
|
| 395 |
+
"grad_norm": 148.27394104003906,
|
| 396 |
+
"learning_rate": 1.5245901639344264e-05,
|
| 397 |
+
"loss": 0.5446,
|
| 398 |
+
"step": 1150
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.422496570644719,
|
| 402 |
+
"grad_norm": 114.31352996826172,
|
| 403 |
+
"learning_rate": 1.5094110503946571e-05,
|
| 404 |
+
"loss": 0.4956,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.559670781893004,
|
| 409 |
+
"grad_norm": 0.42057859897613525,
|
| 410 |
+
"learning_rate": 1.4942319368548878e-05,
|
| 411 |
+
"loss": 0.3221,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.69684499314129,
|
| 416 |
+
"grad_norm": 0.4278319180011749,
|
| 417 |
+
"learning_rate": 1.4790528233151184e-05,
|
| 418 |
+
"loss": 0.2884,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.834019204389575,
|
| 423 |
+
"grad_norm": 485.1934509277344,
|
| 424 |
+
"learning_rate": 1.4638737097753493e-05,
|
| 425 |
+
"loss": 0.2623,
|
| 426 |
+
"step": 1250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 6.97119341563786,
|
| 430 |
+
"grad_norm": 6.358796119689941,
|
| 431 |
+
"learning_rate": 1.44869459623558e-05,
|
| 432 |
+
"loss": 0.3269,
|
| 433 |
+
"step": 1275
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 7.0,
|
| 437 |
+
"eval_accuracy": 0.8602329450915142,
|
| 438 |
+
"eval_loss": 0.6258249878883362,
|
| 439 |
+
"eval_macro_f1": 0.8604473515864366,
|
| 440 |
+
"eval_runtime": 0.4274,
|
| 441 |
+
"eval_samples_per_second": 1406.173,
|
| 442 |
+
"eval_steps_per_second": 88.909,
|
| 443 |
+
"eval_weighted_f1": 0.8603024788936598,
|
| 444 |
+
"step": 1281
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 7.104252400548697,
|
| 448 |
+
"grad_norm": 10.11790943145752,
|
| 449 |
+
"learning_rate": 1.4335154826958106e-05,
|
| 450 |
+
"loss": 0.302,
|
| 451 |
+
"step": 1300
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 7.2414266117969825,
|
| 455 |
+
"grad_norm": 174.15176391601562,
|
| 456 |
+
"learning_rate": 1.4183363691560415e-05,
|
| 457 |
+
"loss": 0.3166,
|
| 458 |
+
"step": 1325
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 7.378600823045267,
|
| 462 |
+
"grad_norm": 2.7166993618011475,
|
| 463 |
+
"learning_rate": 1.403157255616272e-05,
|
| 464 |
+
"loss": 0.4992,
|
| 465 |
+
"step": 1350
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 7.5157750342935525,
|
| 469 |
+
"grad_norm": 13.237764358520508,
|
| 470 |
+
"learning_rate": 1.3879781420765028e-05,
|
| 471 |
+
"loss": 0.3626,
|
| 472 |
+
"step": 1375
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 7.652949245541838,
|
| 476 |
+
"grad_norm": 0.34104979038238525,
|
| 477 |
+
"learning_rate": 1.3727990285367337e-05,
|
| 478 |
+
"loss": 0.3417,
|
| 479 |
+
"step": 1400
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 7.790123456790123,
|
| 483 |
+
"grad_norm": 120.06942749023438,
|
| 484 |
+
"learning_rate": 1.3576199149969643e-05,
|
| 485 |
+
"loss": 0.1955,
|
| 486 |
+
"step": 1425
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 7.927297668038409,
|
| 490 |
+
"grad_norm": 39.646305084228516,
|
| 491 |
+
"learning_rate": 1.342440801457195e-05,
|
| 492 |
+
"loss": 0.2581,
|
| 493 |
+
"step": 1450
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 8.0,
|
| 497 |
+
"eval_accuracy": 0.8652246256239601,
|
| 498 |
+
"eval_loss": 0.6848629713058472,
|
| 499 |
+
"eval_macro_f1": 0.8658216687964972,
|
| 500 |
+
"eval_runtime": 0.4104,
|
| 501 |
+
"eval_samples_per_second": 1464.414,
|
| 502 |
+
"eval_steps_per_second": 92.592,
|
| 503 |
+
"eval_weighted_f1": 0.8653723135418402,
|
| 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": 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-03-s42-20260826T151649Z-00ca7096/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-s42-20260826T151649Z-00ca7096/trainer/checkpoint-2013/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-source-source-03-s42-20260826T151649Z-00ca7096/trainer/checkpoint-2013/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-source-source-03-s42-20260826T151649Z-00ca7096/trainer/checkpoint-2013/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-s42-20260826T151649Z-00ca7096/trainer/checkpoint-2013/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-s42-20260826T151649Z-00ca7096/trainer/checkpoint-2013/trainer_state.json
ADDED
|
@@ -0,0 +1,724 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1464,
|
| 3 |
+
"best_metric": 0.8658216687964972,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-source-source-03-s42-20260826T151649Z-00ca7096/trainer/checkpoint-1464",
|
| 5 |
+
"epoch": 11.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 2013,
|
| 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": 9.540091514587402,
|
| 15 |
+
"learning_rate": 1.3114754098360657e-06,
|
| 16 |
+
"loss": 4.4265,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.27434842249657065,
|
| 21 |
+
"grad_norm": 7.004621982574463,
|
| 22 |
+
"learning_rate": 2.677595628415301e-06,
|
| 23 |
+
"loss": 4.3974,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.411522633744856,
|
| 28 |
+
"grad_norm": 5.635507583618164,
|
| 29 |
+
"learning_rate": 4.043715846994536e-06,
|
| 30 |
+
"loss": 4.4071,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5486968449931413,
|
| 35 |
+
"grad_norm": 8.972846031188965,
|
| 36 |
+
"learning_rate": 5.409836065573772e-06,
|
| 37 |
+
"loss": 4.3508,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6858710562414266,
|
| 42 |
+
"grad_norm": 7.308180809020996,
|
| 43 |
+
"learning_rate": 6.775956284153006e-06,
|
| 44 |
+
"loss": 4.3224,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.823045267489712,
|
| 49 |
+
"grad_norm": 9.94659423828125,
|
| 50 |
+
"learning_rate": 8.14207650273224e-06,
|
| 51 |
+
"loss": 4.2761,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9602194787379973,
|
| 56 |
+
"grad_norm": 12.845196723937988,
|
| 57 |
+
"learning_rate": 9.508196721311476e-06,
|
| 58 |
+
"loss": 4.1545,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.5574043261231281,
|
| 64 |
+
"eval_loss": 0.9540154337882996,
|
| 65 |
+
"eval_macro_f1": 0.5295010108297173,
|
| 66 |
+
"eval_runtime": 0.4751,
|
| 67 |
+
"eval_samples_per_second": 1264.914,
|
| 68 |
+
"eval_steps_per_second": 79.978,
|
| 69 |
+
"eval_weighted_f1": 0.5327022775590124,
|
| 70 |
+
"step": 183
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.093278463648834,
|
| 74 |
+
"grad_norm": 24.179080963134766,
|
| 75 |
+
"learning_rate": 1.0874316939890712e-05,
|
| 76 |
+
"loss": 3.6165,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.2304526748971194,
|
| 81 |
+
"grad_norm": 39.38579177856445,
|
| 82 |
+
"learning_rate": 1.2240437158469946e-05,
|
| 83 |
+
"loss": 3.2323,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.3676268861454046,
|
| 88 |
+
"grad_norm": 32.005470275878906,
|
| 89 |
+
"learning_rate": 1.3606557377049181e-05,
|
| 90 |
+
"loss": 3.0398,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.50480109739369,
|
| 95 |
+
"grad_norm": 35.23550796508789,
|
| 96 |
+
"learning_rate": 1.4972677595628417e-05,
|
| 97 |
+
"loss": 3.0094,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.6419753086419753,
|
| 102 |
+
"grad_norm": 28.283357620239258,
|
| 103 |
+
"learning_rate": 1.633879781420765e-05,
|
| 104 |
+
"loss": 2.6878,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7791495198902605,
|
| 109 |
+
"grad_norm": 45.30286407470703,
|
| 110 |
+
"learning_rate": 1.7704918032786887e-05,
|
| 111 |
+
"loss": 2.6862,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.916323731138546,
|
| 116 |
+
"grad_norm": 20.715063095092773,
|
| 117 |
+
"learning_rate": 1.9071038251366122e-05,
|
| 118 |
+
"loss": 2.1346,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.8236272878535774,
|
| 124 |
+
"eval_loss": 0.47288697957992554,
|
| 125 |
+
"eval_macro_f1": 0.8235667369793007,
|
| 126 |
+
"eval_runtime": 0.4215,
|
| 127 |
+
"eval_samples_per_second": 1425.946,
|
| 128 |
+
"eval_steps_per_second": 90.16,
|
| 129 |
+
"eval_weighted_f1": 0.823587567393166,
|
| 130 |
+
"step": 366
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.049382716049383,
|
| 134 |
+
"grad_norm": 38.181583404541016,
|
| 135 |
+
"learning_rate": 1.995142683667274e-05,
|
| 136 |
+
"loss": 1.9687,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.186556927297668,
|
| 141 |
+
"grad_norm": 47.709129333496094,
|
| 142 |
+
"learning_rate": 1.9799635701275045e-05,
|
| 143 |
+
"loss": 1.609,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.3237311385459534,
|
| 148 |
+
"grad_norm": 39.530494689941406,
|
| 149 |
+
"learning_rate": 1.9647844565877354e-05,
|
| 150 |
+
"loss": 1.769,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.460905349794239,
|
| 155 |
+
"grad_norm": 39.43211364746094,
|
| 156 |
+
"learning_rate": 1.9496053430479663e-05,
|
| 157 |
+
"loss": 1.7113,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.598079561042524,
|
| 162 |
+
"grad_norm": 43.74077606201172,
|
| 163 |
+
"learning_rate": 1.934426229508197e-05,
|
| 164 |
+
"loss": 1.5929,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.7352537722908092,
|
| 169 |
+
"grad_norm": 39.347652435302734,
|
| 170 |
+
"learning_rate": 1.9192471159684274e-05,
|
| 171 |
+
"loss": 1.5992,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.8724279835390947,
|
| 176 |
+
"grad_norm": 35.88862609863281,
|
| 177 |
+
"learning_rate": 1.9040680024286583e-05,
|
| 178 |
+
"loss": 1.7626,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 3.0,
|
| 183 |
+
"eval_accuracy": 0.8552412645590682,
|
| 184 |
+
"eval_loss": 0.408374547958374,
|
| 185 |
+
"eval_macro_f1": 0.8559032398750395,
|
| 186 |
+
"eval_runtime": 0.4087,
|
| 187 |
+
"eval_samples_per_second": 1470.618,
|
| 188 |
+
"eval_steps_per_second": 92.984,
|
| 189 |
+
"eval_weighted_f1": 0.855575761505065,
|
| 190 |
+
"step": 549
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.0054869684499312,
|
| 194 |
+
"grad_norm": 37.14410400390625,
|
| 195 |
+
"learning_rate": 1.888888888888889e-05,
|
| 196 |
+
"loss": 1.4598,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.1426611796982167,
|
| 201 |
+
"grad_norm": 38.32145309448242,
|
| 202 |
+
"learning_rate": 1.8737097753491198e-05,
|
| 203 |
+
"loss": 1.2316,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.279835390946502,
|
| 208 |
+
"grad_norm": 40.79185104370117,
|
| 209 |
+
"learning_rate": 1.8585306618093504e-05,
|
| 210 |
+
"loss": 0.9759,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.4170096021947876,
|
| 215 |
+
"grad_norm": 24.409038543701172,
|
| 216 |
+
"learning_rate": 1.8433515482695813e-05,
|
| 217 |
+
"loss": 0.9108,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.5541838134430725,
|
| 222 |
+
"grad_norm": 75.9214859008789,
|
| 223 |
+
"learning_rate": 1.828172434729812e-05,
|
| 224 |
+
"loss": 1.1125,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.691358024691358,
|
| 229 |
+
"grad_norm": 24.009607315063477,
|
| 230 |
+
"learning_rate": 1.8129933211900424e-05,
|
| 231 |
+
"loss": 1.153,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.8285322359396434,
|
| 236 |
+
"grad_norm": 35.08176040649414,
|
| 237 |
+
"learning_rate": 1.7978142076502733e-05,
|
| 238 |
+
"loss": 1.3368,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.9657064471879284,
|
| 243 |
+
"grad_norm": 33.363094329833984,
|
| 244 |
+
"learning_rate": 1.7826350941105042e-05,
|
| 245 |
+
"loss": 1.1569,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 4.0,
|
| 250 |
+
"eval_accuracy": 0.8519134775374376,
|
| 251 |
+
"eval_loss": 0.4405615031719208,
|
| 252 |
+
"eval_macro_f1": 0.8520966541201226,
|
| 253 |
+
"eval_runtime": 0.4349,
|
| 254 |
+
"eval_samples_per_second": 1381.888,
|
| 255 |
+
"eval_steps_per_second": 87.374,
|
| 256 |
+
"eval_weighted_f1": 0.851633137685227,
|
| 257 |
+
"step": 732
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.098765432098766,
|
| 261 |
+
"grad_norm": 37.453216552734375,
|
| 262 |
+
"learning_rate": 1.7674559805707348e-05,
|
| 263 |
+
"loss": 0.83,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.23593964334705,
|
| 268 |
+
"grad_norm": 35.347137451171875,
|
| 269 |
+
"learning_rate": 1.7522768670309657e-05,
|
| 270 |
+
"loss": 0.723,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.373113854595336,
|
| 275 |
+
"grad_norm": 36.2155876159668,
|
| 276 |
+
"learning_rate": 1.7370977534911962e-05,
|
| 277 |
+
"loss": 0.7777,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.510288065843621,
|
| 282 |
+
"grad_norm": 38.25359344482422,
|
| 283 |
+
"learning_rate": 1.7219186399514268e-05,
|
| 284 |
+
"loss": 0.6556,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.647462277091907,
|
| 289 |
+
"grad_norm": 69.14533233642578,
|
| 290 |
+
"learning_rate": 1.7067395264116577e-05,
|
| 291 |
+
"loss": 0.8692,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.784636488340192,
|
| 296 |
+
"grad_norm": 12.52321720123291,
|
| 297 |
+
"learning_rate": 1.6915604128718883e-05,
|
| 298 |
+
"loss": 0.8442,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.921810699588478,
|
| 303 |
+
"grad_norm": 32.477474212646484,
|
| 304 |
+
"learning_rate": 1.6763812993321192e-05,
|
| 305 |
+
"loss": 0.7072,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 5.0,
|
| 310 |
+
"eval_accuracy": 0.8569051580698835,
|
| 311 |
+
"eval_loss": 0.4720706641674042,
|
| 312 |
+
"eval_macro_f1": 0.8572179129600631,
|
| 313 |
+
"eval_runtime": 0.4497,
|
| 314 |
+
"eval_samples_per_second": 1336.492,
|
| 315 |
+
"eval_steps_per_second": 84.504,
|
| 316 |
+
"eval_weighted_f1": 0.856788740570396,
|
| 317 |
+
"step": 915
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.054869684499314,
|
| 321 |
+
"grad_norm": 48.532066345214844,
|
| 322 |
+
"learning_rate": 1.66120218579235e-05,
|
| 323 |
+
"loss": 0.7139,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.1920438957476,
|
| 328 |
+
"grad_norm": 1.913240909576416,
|
| 329 |
+
"learning_rate": 1.6460230722525807e-05,
|
| 330 |
+
"loss": 0.5688,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.329218106995885,
|
| 335 |
+
"grad_norm": 13.59510326385498,
|
| 336 |
+
"learning_rate": 1.6308439587128112e-05,
|
| 337 |
+
"loss": 0.4244,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.4663923182441705,
|
| 342 |
+
"grad_norm": 103.00779724121094,
|
| 343 |
+
"learning_rate": 1.615664845173042e-05,
|
| 344 |
+
"loss": 0.3592,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.603566529492455,
|
| 349 |
+
"grad_norm": 26.168684005737305,
|
| 350 |
+
"learning_rate": 1.6004857316332727e-05,
|
| 351 |
+
"loss": 0.5435,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.7407407407407405,
|
| 356 |
+
"grad_norm": 19.721546173095703,
|
| 357 |
+
"learning_rate": 1.5853066180935036e-05,
|
| 358 |
+
"loss": 0.5746,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.877914951989026,
|
| 363 |
+
"grad_norm": 86.65348052978516,
|
| 364 |
+
"learning_rate": 1.570127504553734e-05,
|
| 365 |
+
"loss": 0.7417,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 6.0,
|
| 370 |
+
"eval_accuracy": 0.8569051580698835,
|
| 371 |
+
"eval_loss": 0.5869454145431519,
|
| 372 |
+
"eval_macro_f1": 0.8568009493221235,
|
| 373 |
+
"eval_runtime": 0.3844,
|
| 374 |
+
"eval_samples_per_second": 1563.431,
|
| 375 |
+
"eval_steps_per_second": 98.853,
|
| 376 |
+
"eval_weighted_f1": 0.8566164852264386,
|
| 377 |
+
"step": 1098
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 6.0109739368998625,
|
| 381 |
+
"grad_norm": 64.31432342529297,
|
| 382 |
+
"learning_rate": 1.5549483910139647e-05,
|
| 383 |
+
"loss": 0.6296,
|
| 384 |
+
"step": 1100
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 6.148148148148148,
|
| 388 |
+
"grad_norm": 2.6430726051330566,
|
| 389 |
+
"learning_rate": 1.5397692774741956e-05,
|
| 390 |
+
"loss": 0.3162,
|
| 391 |
+
"step": 1125
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 6.285322359396433,
|
| 395 |
+
"grad_norm": 148.27394104003906,
|
| 396 |
+
"learning_rate": 1.5245901639344264e-05,
|
| 397 |
+
"loss": 0.5446,
|
| 398 |
+
"step": 1150
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.422496570644719,
|
| 402 |
+
"grad_norm": 114.31352996826172,
|
| 403 |
+
"learning_rate": 1.5094110503946571e-05,
|
| 404 |
+
"loss": 0.4956,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.559670781893004,
|
| 409 |
+
"grad_norm": 0.42057859897613525,
|
| 410 |
+
"learning_rate": 1.4942319368548878e-05,
|
| 411 |
+
"loss": 0.3221,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.69684499314129,
|
| 416 |
+
"grad_norm": 0.4278319180011749,
|
| 417 |
+
"learning_rate": 1.4790528233151184e-05,
|
| 418 |
+
"loss": 0.2884,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.834019204389575,
|
| 423 |
+
"grad_norm": 485.1934509277344,
|
| 424 |
+
"learning_rate": 1.4638737097753493e-05,
|
| 425 |
+
"loss": 0.2623,
|
| 426 |
+
"step": 1250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 6.97119341563786,
|
| 430 |
+
"grad_norm": 6.358796119689941,
|
| 431 |
+
"learning_rate": 1.44869459623558e-05,
|
| 432 |
+
"loss": 0.3269,
|
| 433 |
+
"step": 1275
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 7.0,
|
| 437 |
+
"eval_accuracy": 0.8602329450915142,
|
| 438 |
+
"eval_loss": 0.6258249878883362,
|
| 439 |
+
"eval_macro_f1": 0.8604473515864366,
|
| 440 |
+
"eval_runtime": 0.4274,
|
| 441 |
+
"eval_samples_per_second": 1406.173,
|
| 442 |
+
"eval_steps_per_second": 88.909,
|
| 443 |
+
"eval_weighted_f1": 0.8603024788936598,
|
| 444 |
+
"step": 1281
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 7.104252400548697,
|
| 448 |
+
"grad_norm": 10.11790943145752,
|
| 449 |
+
"learning_rate": 1.4335154826958106e-05,
|
| 450 |
+
"loss": 0.302,
|
| 451 |
+
"step": 1300
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 7.2414266117969825,
|
| 455 |
+
"grad_norm": 174.15176391601562,
|
| 456 |
+
"learning_rate": 1.4183363691560415e-05,
|
| 457 |
+
"loss": 0.3166,
|
| 458 |
+
"step": 1325
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 7.378600823045267,
|
| 462 |
+
"grad_norm": 2.7166993618011475,
|
| 463 |
+
"learning_rate": 1.403157255616272e-05,
|
| 464 |
+
"loss": 0.4992,
|
| 465 |
+
"step": 1350
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 7.5157750342935525,
|
| 469 |
+
"grad_norm": 13.237764358520508,
|
| 470 |
+
"learning_rate": 1.3879781420765028e-05,
|
| 471 |
+
"loss": 0.3626,
|
| 472 |
+
"step": 1375
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 7.652949245541838,
|
| 476 |
+
"grad_norm": 0.34104979038238525,
|
| 477 |
+
"learning_rate": 1.3727990285367337e-05,
|
| 478 |
+
"loss": 0.3417,
|
| 479 |
+
"step": 1400
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 7.790123456790123,
|
| 483 |
+
"grad_norm": 120.06942749023438,
|
| 484 |
+
"learning_rate": 1.3576199149969643e-05,
|
| 485 |
+
"loss": 0.1955,
|
| 486 |
+
"step": 1425
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 7.927297668038409,
|
| 490 |
+
"grad_norm": 39.646305084228516,
|
| 491 |
+
"learning_rate": 1.342440801457195e-05,
|
| 492 |
+
"loss": 0.2581,
|
| 493 |
+
"step": 1450
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 8.0,
|
| 497 |
+
"eval_accuracy": 0.8652246256239601,
|
| 498 |
+
"eval_loss": 0.6848629713058472,
|
| 499 |
+
"eval_macro_f1": 0.8658216687964972,
|
| 500 |
+
"eval_runtime": 0.4104,
|
| 501 |
+
"eval_samples_per_second": 1464.414,
|
| 502 |
+
"eval_steps_per_second": 92.592,
|
| 503 |
+
"eval_weighted_f1": 0.8653723135418402,
|
| 504 |
+
"step": 1464
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 8.060356652949245,
|
| 508 |
+
"grad_norm": 4.535733222961426,
|
| 509 |
+
"learning_rate": 1.3272616879174259e-05,
|
| 510 |
+
"loss": 0.293,
|
| 511 |
+
"step": 1475
|
| 512 |
+
},
|
| 513 |
+
{
|
| 514 |
+
"epoch": 8.197530864197532,
|
| 515 |
+
"grad_norm": 0.2250947505235672,
|
| 516 |
+
"learning_rate": 1.3120825743776565e-05,
|
| 517 |
+
"loss": 0.1727,
|
| 518 |
+
"step": 1500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 8.334705075445816,
|
| 522 |
+
"grad_norm": 84.7310791015625,
|
| 523 |
+
"learning_rate": 1.2969034608378872e-05,
|
| 524 |
+
"loss": 0.2719,
|
| 525 |
+
"step": 1525
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 8.4718792866941,
|
| 529 |
+
"grad_norm": 0.23370251059532166,
|
| 530 |
+
"learning_rate": 1.2817243472981178e-05,
|
| 531 |
+
"loss": 0.2494,
|
| 532 |
+
"step": 1550
|
| 533 |
+
},
|
| 534 |
+
{
|
| 535 |
+
"epoch": 8.609053497942387,
|
| 536 |
+
"grad_norm": 236.5485076904297,
|
| 537 |
+
"learning_rate": 1.2665452337583487e-05,
|
| 538 |
+
"loss": 0.2405,
|
| 539 |
+
"step": 1575
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 8.746227709190672,
|
| 543 |
+
"grad_norm": 0.4279741048812866,
|
| 544 |
+
"learning_rate": 1.2513661202185794e-05,
|
| 545 |
+
"loss": 0.2004,
|
| 546 |
+
"step": 1600
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 8.883401920438958,
|
| 550 |
+
"grad_norm": 12.447382926940918,
|
| 551 |
+
"learning_rate": 1.23618700667881e-05,
|
| 552 |
+
"loss": 0.2088,
|
| 553 |
+
"step": 1625
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 9.0,
|
| 557 |
+
"eval_accuracy": 0.848585690515807,
|
| 558 |
+
"eval_loss": 0.7770034074783325,
|
| 559 |
+
"eval_macro_f1": 0.849335984732475,
|
| 560 |
+
"eval_runtime": 0.4242,
|
| 561 |
+
"eval_samples_per_second": 1416.847,
|
| 562 |
+
"eval_steps_per_second": 89.584,
|
| 563 |
+
"eval_weighted_f1": 0.8485937782796599,
|
| 564 |
+
"step": 1647
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"epoch": 9.016460905349794,
|
| 568 |
+
"grad_norm": 47.39944076538086,
|
| 569 |
+
"learning_rate": 1.2210078931390409e-05,
|
| 570 |
+
"loss": 0.2674,
|
| 571 |
+
"step": 1650
|
| 572 |
+
},
|
| 573 |
+
{
|
| 574 |
+
"epoch": 9.15363511659808,
|
| 575 |
+
"grad_norm": 0.18962512910366058,
|
| 576 |
+
"learning_rate": 1.2058287795992716e-05,
|
| 577 |
+
"loss": 0.2532,
|
| 578 |
+
"step": 1675
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 9.290809327846365,
|
| 582 |
+
"grad_norm": 1.6242234706878662,
|
| 583 |
+
"learning_rate": 1.1906496660595022e-05,
|
| 584 |
+
"loss": 0.2636,
|
| 585 |
+
"step": 1700
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 9.427983539094651,
|
| 589 |
+
"grad_norm": 0.321869432926178,
|
| 590 |
+
"learning_rate": 1.1754705525197329e-05,
|
| 591 |
+
"loss": 0.16,
|
| 592 |
+
"step": 1725
|
| 593 |
+
},
|
| 594 |
+
{
|
| 595 |
+
"epoch": 9.565157750342935,
|
| 596 |
+
"grad_norm": 153.10398864746094,
|
| 597 |
+
"learning_rate": 1.1602914389799638e-05,
|
| 598 |
+
"loss": 0.1382,
|
| 599 |
+
"step": 1750
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 9.70233196159122,
|
| 603 |
+
"grad_norm": 0.6122498512268066,
|
| 604 |
+
"learning_rate": 1.1451123254401944e-05,
|
| 605 |
+
"loss": 0.0756,
|
| 606 |
+
"step": 1775
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"epoch": 9.839506172839506,
|
| 610 |
+
"grad_norm": 8.670122146606445,
|
| 611 |
+
"learning_rate": 1.1299332119004251e-05,
|
| 612 |
+
"loss": 0.2291,
|
| 613 |
+
"step": 1800
|
| 614 |
+
},
|
| 615 |
+
{
|
| 616 |
+
"epoch": 9.976680384087791,
|
| 617 |
+
"grad_norm": 0.16853903234004974,
|
| 618 |
+
"learning_rate": 1.1147540983606557e-05,
|
| 619 |
+
"loss": 0.1129,
|
| 620 |
+
"step": 1825
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 10.0,
|
| 624 |
+
"eval_accuracy": 0.8402662229617305,
|
| 625 |
+
"eval_loss": 0.8923433423042297,
|
| 626 |
+
"eval_macro_f1": 0.8402263987116724,
|
| 627 |
+
"eval_runtime": 0.4345,
|
| 628 |
+
"eval_samples_per_second": 1383.122,
|
| 629 |
+
"eval_steps_per_second": 87.452,
|
| 630 |
+
"eval_weighted_f1": 0.8396345204190007,
|
| 631 |
+
"step": 1830
|
| 632 |
+
},
|
| 633 |
+
{
|
| 634 |
+
"epoch": 10.109739368998628,
|
| 635 |
+
"grad_norm": 8.139730453491211,
|
| 636 |
+
"learning_rate": 1.0995749848208866e-05,
|
| 637 |
+
"loss": 0.086,
|
| 638 |
+
"step": 1850
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 10.246913580246913,
|
| 642 |
+
"grad_norm": 82.33293914794922,
|
| 643 |
+
"learning_rate": 1.0843958712811173e-05,
|
| 644 |
+
"loss": 0.0708,
|
| 645 |
+
"step": 1875
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 10.3840877914952,
|
| 649 |
+
"grad_norm": 0.21831564605236053,
|
| 650 |
+
"learning_rate": 1.0692167577413479e-05,
|
| 651 |
+
"loss": 0.1522,
|
| 652 |
+
"step": 1900
|
| 653 |
+
},
|
| 654 |
+
{
|
| 655 |
+
"epoch": 10.521262002743484,
|
| 656 |
+
"grad_norm": 448.5440368652344,
|
| 657 |
+
"learning_rate": 1.0540376442015788e-05,
|
| 658 |
+
"loss": 0.1584,
|
| 659 |
+
"step": 1925
|
| 660 |
+
},
|
| 661 |
+
{
|
| 662 |
+
"epoch": 10.65843621399177,
|
| 663 |
+
"grad_norm": 0.26956242322921753,
|
| 664 |
+
"learning_rate": 1.0388585306618095e-05,
|
| 665 |
+
"loss": 0.1184,
|
| 666 |
+
"step": 1950
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"epoch": 10.795610425240055,
|
| 670 |
+
"grad_norm": 0.1106225773692131,
|
| 671 |
+
"learning_rate": 1.02367941712204e-05,
|
| 672 |
+
"loss": 0.1119,
|
| 673 |
+
"step": 1975
|
| 674 |
+
},
|
| 675 |
+
{
|
| 676 |
+
"epoch": 10.932784636488341,
|
| 677 |
+
"grad_norm": 0.10314441472291946,
|
| 678 |
+
"learning_rate": 1.008500303582271e-05,
|
| 679 |
+
"loss": 0.0847,
|
| 680 |
+
"step": 2000
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 11.0,
|
| 684 |
+
"eval_accuracy": 0.8386023294509152,
|
| 685 |
+
"eval_loss": 0.9175944924354553,
|
| 686 |
+
"eval_macro_f1": 0.8385863883061967,
|
| 687 |
+
"eval_runtime": 0.4351,
|
| 688 |
+
"eval_samples_per_second": 1381.368,
|
| 689 |
+
"eval_steps_per_second": 87.341,
|
| 690 |
+
"eval_weighted_f1": 0.8381946514408195,
|
| 691 |
+
"step": 2013
|
| 692 |
+
}
|
| 693 |
+
],
|
| 694 |
+
"logging_steps": 25,
|
| 695 |
+
"max_steps": 3660,
|
| 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-03-s42-20260826T151649Z-00ca7096/trainer/checkpoint-2013/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-source-source-04-s42-20260826T152033Z-00ca7096/trainer/checkpoint-1408/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-source-source-04-s42-20260826T152033Z-00ca7096/trainer/checkpoint-1408/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|