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-s44-20260826T160106Z-ab4112e3/trainer/checkpoint-1092/added_tokens.json +3 -0
- ood-vifactcheck-phobert-source-source-00-s44-20260826T160106Z-ab4112e3/trainer/checkpoint-1092/bpe.codes +0 -0
- ood-vifactcheck-phobert-source-source-00-s44-20260826T160106Z-ab4112e3/trainer/checkpoint-1092/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-source-source-00-s44-20260826T160106Z-ab4112e3/trainer/checkpoint-1092/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-source-source-00-s44-20260826T160106Z-ab4112e3/trainer/checkpoint-1092/trainer_state.json +410 -0
- ood-vifactcheck-phobert-source-source-00-s44-20260826T160106Z-ab4112e3/trainer/checkpoint-1092/vocab.txt +0 -0
- ood-vifactcheck-phobert-source-source-00-s44-20260826T160106Z-ab4112e3/trainer/checkpoint-546/added_tokens.json +3 -0
- ood-vifactcheck-phobert-source-source-00-s44-20260826T160106Z-ab4112e3/trainer/checkpoint-546/bpe.codes +0 -0
- ood-vifactcheck-phobert-source-source-00-s44-20260826T160106Z-ab4112e3/trainer/checkpoint-546/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-source-source-00-s44-20260826T160106Z-ab4112e3/trainer/checkpoint-546/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-source-source-00-s44-20260826T160106Z-ab4112e3/trainer/checkpoint-546/trainer_state.json +223 -0
- ood-vifactcheck-phobert-source-source-00-s44-20260826T160106Z-ab4112e3/trainer/checkpoint-546/vocab.txt +0 -0
- ood-vifactcheck-phobert-source-source-01-s44-20260826T160312Z-ab4112e3/trainer/checkpoint-1539/added_tokens.json +3 -0
- ood-vifactcheck-phobert-source-source-01-s44-20260826T160312Z-ab4112e3/trainer/checkpoint-1539/bpe.codes +0 -0
- ood-vifactcheck-phobert-source-source-01-s44-20260826T160312Z-ab4112e3/trainer/checkpoint-1539/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-source-source-01-s44-20260826T160312Z-ab4112e3/trainer/checkpoint-1539/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-source-source-01-s44-20260826T160312Z-ab4112e3/trainer/checkpoint-1539/trainer_state.json +569 -0
- ood-vifactcheck-phobert-source-source-01-s44-20260826T160312Z-ab4112e3/trainer/checkpoint-1539/vocab.txt +0 -0
- ood-vifactcheck-phobert-source-source-01-s44-20260826T160312Z-ab4112e3/trainer/checkpoint-2052/added_tokens.json +3 -0
- ood-vifactcheck-phobert-source-source-01-s44-20260826T160312Z-ab4112e3/trainer/checkpoint-2052/bpe.codes +0 -0
- ood-vifactcheck-phobert-source-source-01-s44-20260826T160312Z-ab4112e3/trainer/checkpoint-2052/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-source-source-01-s44-20260826T160312Z-ab4112e3/trainer/checkpoint-2052/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-source-source-01-s44-20260826T160312Z-ab4112e3/trainer/checkpoint-2052/trainer_state.json +749 -0
- ood-vifactcheck-phobert-source-source-01-s44-20260826T160312Z-ab4112e3/trainer/checkpoint-2052/vocab.txt +0 -0
- ood-vifactcheck-phobert-source-source-02-s44-20260826T160702Z-ab4112e3/trainer/checkpoint-1086/added_tokens.json +3 -0
- ood-vifactcheck-phobert-source-source-02-s44-20260826T160702Z-ab4112e3/trainer/checkpoint-1086/bpe.codes +0 -0
- ood-vifactcheck-phobert-source-source-02-s44-20260826T160702Z-ab4112e3/trainer/checkpoint-1086/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-source-source-02-s44-20260826T160702Z-ab4112e3/trainer/checkpoint-1086/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-source-source-02-s44-20260826T160702Z-ab4112e3/trainer/checkpoint-1086/trainer_state.json +410 -0
- ood-vifactcheck-phobert-source-source-02-s44-20260826T160702Z-ab4112e3/trainer/checkpoint-1086/vocab.txt +0 -0
- ood-vifactcheck-phobert-source-source-02-s44-20260826T160702Z-ab4112e3/trainer/checkpoint-1629/added_tokens.json +3 -0
- ood-vifactcheck-phobert-source-source-02-s44-20260826T160702Z-ab4112e3/trainer/checkpoint-1629/bpe.codes +0 -0
- ood-vifactcheck-phobert-source-source-02-s44-20260826T160702Z-ab4112e3/trainer/checkpoint-1629/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-source-source-02-s44-20260826T160702Z-ab4112e3/trainer/checkpoint-1629/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-source-source-02-s44-20260826T160702Z-ab4112e3/trainer/checkpoint-1629/trainer_state.json +597 -0
- ood-vifactcheck-phobert-source-source-02-s44-20260826T160702Z-ab4112e3/trainer/checkpoint-1629/vocab.txt +0 -0
- ood-vifactcheck-phobert-source-source-03-s44-20260826T161012Z-ab4112e3/trainer/checkpoint-1098/added_tokens.json +3 -0
- ood-vifactcheck-phobert-source-source-03-s44-20260826T161012Z-ab4112e3/trainer/checkpoint-1098/bpe.codes +0 -0
- ood-vifactcheck-phobert-source-source-03-s44-20260826T161012Z-ab4112e3/trainer/checkpoint-1098/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-source-source-03-s44-20260826T161012Z-ab4112e3/trainer/checkpoint-1098/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-source-source-03-s44-20260826T161012Z-ab4112e3/trainer/checkpoint-1098/trainer_state.json +410 -0
- ood-vifactcheck-phobert-source-source-03-s44-20260826T161012Z-ab4112e3/trainer/checkpoint-1098/vocab.txt +0 -0
- ood-vifactcheck-phobert-source-source-03-s44-20260826T161012Z-ab4112e3/trainer/checkpoint-1647/added_tokens.json +3 -0
- ood-vifactcheck-phobert-source-source-03-s44-20260826T161012Z-ab4112e3/trainer/checkpoint-1647/bpe.codes +0 -0
- ood-vifactcheck-phobert-source-source-03-s44-20260826T161012Z-ab4112e3/trainer/checkpoint-1647/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-source-source-03-s44-20260826T161012Z-ab4112e3/trainer/checkpoint-1647/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-source-source-03-s44-20260826T161012Z-ab4112e3/trainer/checkpoint-1647/trainer_state.json +597 -0
- ood-vifactcheck-phobert-source-source-03-s44-20260826T161012Z-ab4112e3/trainer/checkpoint-1647/vocab.txt +0 -0
- ood-vifactcheck-phobert-source-source-04-s44-20260826T161317Z-ab4112e3/trainer/checkpoint-1232/added_tokens.json +3 -0
- ood-vifactcheck-phobert-source-source-04-s44-20260826T161317Z-ab4112e3/trainer/checkpoint-1232/bpe.codes +0 -0
ood-vifactcheck-phobert-source-source-00-s44-20260826T160106Z-ab4112e3/trainer/checkpoint-1092/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-source-source-00-s44-20260826T160106Z-ab4112e3/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-s44-20260826T160106Z-ab4112e3/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-s44-20260826T160106Z-ab4112e3/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-s44-20260826T160106Z-ab4112e3/trainer/checkpoint-1092/trainer_state.json
ADDED
|
@@ -0,0 +1,410 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 546,
|
| 3 |
+
"best_metric": 0.8399379636589276,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-source-source-00-s44-20260826T160106Z-ab4112e3/trainer/checkpoint-546",
|
| 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": 6.436005592346191,
|
| 15 |
+
"learning_rate": 1.3186813186813187e-06,
|
| 16 |
+
"loss": 4.4287,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.27586206896551724,
|
| 21 |
+
"grad_norm": 7.078612804412842,
|
| 22 |
+
"learning_rate": 2.6923076923076923e-06,
|
| 23 |
+
"loss": 4.3796,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.41379310344827586,
|
| 28 |
+
"grad_norm": 8.72610855102539,
|
| 29 |
+
"learning_rate": 4.065934065934066e-06,
|
| 30 |
+
"loss": 4.3507,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5517241379310345,
|
| 35 |
+
"grad_norm": 9.247332572937012,
|
| 36 |
+
"learning_rate": 5.43956043956044e-06,
|
| 37 |
+
"loss": 4.3268,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6896551724137931,
|
| 42 |
+
"grad_norm": 13.744568824768066,
|
| 43 |
+
"learning_rate": 6.813186813186814e-06,
|
| 44 |
+
"loss": 4.2635,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8275862068965517,
|
| 49 |
+
"grad_norm": 12.254353523254395,
|
| 50 |
+
"learning_rate": 8.186813186813188e-06,
|
| 51 |
+
"loss": 4.1413,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9655172413793104,
|
| 56 |
+
"grad_norm": 15.760031700134277,
|
| 57 |
+
"learning_rate": 9.560439560439562e-06,
|
| 58 |
+
"loss": 4.0484,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.5933014354066986,
|
| 64 |
+
"eval_loss": 0.9013376235961914,
|
| 65 |
+
"eval_macro_f1": 0.5891153112663168,
|
| 66 |
+
"eval_runtime": 0.4458,
|
| 67 |
+
"eval_samples_per_second": 1406.533,
|
| 68 |
+
"eval_steps_per_second": 89.731,
|
| 69 |
+
"eval_weighted_f1": 0.5938081589361177,
|
| 70 |
+
"step": 182
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0993103448275863,
|
| 74 |
+
"grad_norm": 19.441864013671875,
|
| 75 |
+
"learning_rate": 1.0934065934065933e-05,
|
| 76 |
+
"loss": 3.5205,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.237241379310345,
|
| 81 |
+
"grad_norm": 27.32406234741211,
|
| 82 |
+
"learning_rate": 1.230769230769231e-05,
|
| 83 |
+
"loss": 3.3318,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.3751724137931034,
|
| 88 |
+
"grad_norm": 48.73912811279297,
|
| 89 |
+
"learning_rate": 1.3681318681318683e-05,
|
| 90 |
+
"loss": 3.1308,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.513103448275862,
|
| 95 |
+
"grad_norm": 30.94043731689453,
|
| 96 |
+
"learning_rate": 1.5054945054945056e-05,
|
| 97 |
+
"loss": 2.7108,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.6510344827586207,
|
| 102 |
+
"grad_norm": 33.952392578125,
|
| 103 |
+
"learning_rate": 1.642857142857143e-05,
|
| 104 |
+
"loss": 2.5379,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7889655172413792,
|
| 109 |
+
"grad_norm": 49.711551666259766,
|
| 110 |
+
"learning_rate": 1.78021978021978e-05,
|
| 111 |
+
"loss": 2.4233,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.926896551724138,
|
| 116 |
+
"grad_norm": 51.16925048828125,
|
| 117 |
+
"learning_rate": 1.9175824175824178e-05,
|
| 118 |
+
"loss": 2.5303,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.7862838915470495,
|
| 124 |
+
"eval_loss": 0.5505752563476562,
|
| 125 |
+
"eval_macro_f1": 0.7851859232883802,
|
| 126 |
+
"eval_runtime": 0.4481,
|
| 127 |
+
"eval_samples_per_second": 1399.335,
|
| 128 |
+
"eval_steps_per_second": 89.272,
|
| 129 |
+
"eval_weighted_f1": 0.7857234351971534,
|
| 130 |
+
"step": 364
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.0606896551724136,
|
| 134 |
+
"grad_norm": 28.892581939697266,
|
| 135 |
+
"learning_rate": 1.993894993894994e-05,
|
| 136 |
+
"loss": 1.8631,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.1986206896551725,
|
| 141 |
+
"grad_norm": 43.200462341308594,
|
| 142 |
+
"learning_rate": 1.9786324786324788e-05,
|
| 143 |
+
"loss": 1.764,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.336551724137931,
|
| 148 |
+
"grad_norm": 34.97949981689453,
|
| 149 |
+
"learning_rate": 1.9633699633699634e-05,
|
| 150 |
+
"loss": 1.4838,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.47448275862069,
|
| 155 |
+
"grad_norm": 46.3452033996582,
|
| 156 |
+
"learning_rate": 1.948107448107448e-05,
|
| 157 |
+
"loss": 1.5003,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.6124137931034483,
|
| 162 |
+
"grad_norm": 38.658203125,
|
| 163 |
+
"learning_rate": 1.932844932844933e-05,
|
| 164 |
+
"loss": 1.4092,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.750344827586207,
|
| 169 |
+
"grad_norm": 39.02288055419922,
|
| 170 |
+
"learning_rate": 1.9175824175824178e-05,
|
| 171 |
+
"loss": 1.6422,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.8882758620689657,
|
| 176 |
+
"grad_norm": 49.1451530456543,
|
| 177 |
+
"learning_rate": 1.9023199023199025e-05,
|
| 178 |
+
"loss": 1.5683,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 3.0,
|
| 183 |
+
"eval_accuracy": 0.8421052631578947,
|
| 184 |
+
"eval_loss": 0.4859905242919922,
|
| 185 |
+
"eval_macro_f1": 0.8399379636589276,
|
| 186 |
+
"eval_runtime": 0.4384,
|
| 187 |
+
"eval_samples_per_second": 1430.07,
|
| 188 |
+
"eval_steps_per_second": 91.233,
|
| 189 |
+
"eval_weighted_f1": 0.8406701982792217,
|
| 190 |
+
"step": 546
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.0220689655172412,
|
| 194 |
+
"grad_norm": 25.318910598754883,
|
| 195 |
+
"learning_rate": 1.887057387057387e-05,
|
| 196 |
+
"loss": 1.5493,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.16,
|
| 201 |
+
"grad_norm": 54.30934524536133,
|
| 202 |
+
"learning_rate": 1.8717948717948718e-05,
|
| 203 |
+
"loss": 1.0021,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.2979310344827586,
|
| 208 |
+
"grad_norm": 57.739986419677734,
|
| 209 |
+
"learning_rate": 1.8565323565323568e-05,
|
| 210 |
+
"loss": 1.0429,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.435862068965517,
|
| 215 |
+
"grad_norm": 49.789955139160156,
|
| 216 |
+
"learning_rate": 1.8412698412698415e-05,
|
| 217 |
+
"loss": 1.0258,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.573793103448276,
|
| 222 |
+
"grad_norm": 45.379608154296875,
|
| 223 |
+
"learning_rate": 1.826007326007326e-05,
|
| 224 |
+
"loss": 1.1068,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.7117241379310344,
|
| 229 |
+
"grad_norm": 48.28389358520508,
|
| 230 |
+
"learning_rate": 1.8107448107448108e-05,
|
| 231 |
+
"loss": 0.9938,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.849655172413793,
|
| 236 |
+
"grad_norm": 15.085484504699707,
|
| 237 |
+
"learning_rate": 1.7954822954822958e-05,
|
| 238 |
+
"loss": 1.2601,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.9875862068965517,
|
| 243 |
+
"grad_norm": 36.533302307128906,
|
| 244 |
+
"learning_rate": 1.78021978021978e-05,
|
| 245 |
+
"loss": 0.9784,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 4.0,
|
| 250 |
+
"eval_accuracy": 0.8181818181818182,
|
| 251 |
+
"eval_loss": 0.5944927334785461,
|
| 252 |
+
"eval_macro_f1": 0.8178638742613997,
|
| 253 |
+
"eval_runtime": 0.4417,
|
| 254 |
+
"eval_samples_per_second": 1419.588,
|
| 255 |
+
"eval_steps_per_second": 90.564,
|
| 256 |
+
"eval_weighted_f1": 0.8182883354186913,
|
| 257 |
+
"step": 728
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.121379310344827,
|
| 261 |
+
"grad_norm": 32.77482986450195,
|
| 262 |
+
"learning_rate": 1.764957264957265e-05,
|
| 263 |
+
"loss": 0.5782,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.259310344827586,
|
| 268 |
+
"grad_norm": 16.01955223083496,
|
| 269 |
+
"learning_rate": 1.74969474969475e-05,
|
| 270 |
+
"loss": 0.54,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.397241379310345,
|
| 275 |
+
"grad_norm": 33.46676254272461,
|
| 276 |
+
"learning_rate": 1.7344322344322345e-05,
|
| 277 |
+
"loss": 0.7993,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.535172413793103,
|
| 282 |
+
"grad_norm": 8.671110153198242,
|
| 283 |
+
"learning_rate": 1.7191697191697195e-05,
|
| 284 |
+
"loss": 0.731,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.673103448275862,
|
| 289 |
+
"grad_norm": 41.44632339477539,
|
| 290 |
+
"learning_rate": 1.703907203907204e-05,
|
| 291 |
+
"loss": 0.7804,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.811034482758621,
|
| 296 |
+
"grad_norm": 51.48649978637695,
|
| 297 |
+
"learning_rate": 1.688644688644689e-05,
|
| 298 |
+
"loss": 0.7902,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.94896551724138,
|
| 303 |
+
"grad_norm": 68.74658966064453,
|
| 304 |
+
"learning_rate": 1.6733821733821735e-05,
|
| 305 |
+
"loss": 1.0364,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 5.0,
|
| 310 |
+
"eval_accuracy": 0.8261562998405104,
|
| 311 |
+
"eval_loss": 0.5850275158882141,
|
| 312 |
+
"eval_macro_f1": 0.8248356969951032,
|
| 313 |
+
"eval_runtime": 0.4462,
|
| 314 |
+
"eval_samples_per_second": 1405.316,
|
| 315 |
+
"eval_steps_per_second": 89.653,
|
| 316 |
+
"eval_weighted_f1": 0.8252846066340737,
|
| 317 |
+
"step": 910
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.082758620689655,
|
| 321 |
+
"grad_norm": 8.805537223815918,
|
| 322 |
+
"learning_rate": 1.6581196581196585e-05,
|
| 323 |
+
"loss": 0.6446,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.220689655172414,
|
| 328 |
+
"grad_norm": 25.179691314697266,
|
| 329 |
+
"learning_rate": 1.642857142857143e-05,
|
| 330 |
+
"loss": 0.489,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.358620689655172,
|
| 335 |
+
"grad_norm": 23.378501892089844,
|
| 336 |
+
"learning_rate": 1.627594627594628e-05,
|
| 337 |
+
"loss": 0.4269,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.496551724137931,
|
| 342 |
+
"grad_norm": 44.56985855102539,
|
| 343 |
+
"learning_rate": 1.6123321123321125e-05,
|
| 344 |
+
"loss": 0.5008,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.63448275862069,
|
| 349 |
+
"grad_norm": 48.46287155151367,
|
| 350 |
+
"learning_rate": 1.5970695970695972e-05,
|
| 351 |
+
"loss": 0.5367,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.772413793103448,
|
| 356 |
+
"grad_norm": 86.95524597167969,
|
| 357 |
+
"learning_rate": 1.581807081807082e-05,
|
| 358 |
+
"loss": 0.7081,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.910344827586207,
|
| 363 |
+
"grad_norm": 231.86767578125,
|
| 364 |
+
"learning_rate": 1.5665445665445665e-05,
|
| 365 |
+
"loss": 0.7566,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 6.0,
|
| 370 |
+
"eval_accuracy": 0.8293460925039873,
|
| 371 |
+
"eval_loss": 0.6939406991004944,
|
| 372 |
+
"eval_macro_f1": 0.828785029746999,
|
| 373 |
+
"eval_runtime": 0.4423,
|
| 374 |
+
"eval_samples_per_second": 1417.504,
|
| 375 |
+
"eval_steps_per_second": 90.431,
|
| 376 |
+
"eval_weighted_f1": 0.8286529623003928,
|
| 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": 3
|
| 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": true
|
| 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-s44-20260826T160106Z-ab4112e3/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-s44-20260826T160106Z-ab4112e3/trainer/checkpoint-546/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-source-source-00-s44-20260826T160106Z-ab4112e3/trainer/checkpoint-546/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-source-source-00-s44-20260826T160106Z-ab4112e3/trainer/checkpoint-546/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-s44-20260826T160106Z-ab4112e3/trainer/checkpoint-546/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-s44-20260826T160106Z-ab4112e3/trainer/checkpoint-546/trainer_state.json
ADDED
|
@@ -0,0 +1,223 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 546,
|
| 3 |
+
"best_metric": 0.8399379636589276,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-source-source-00-s44-20260826T160106Z-ab4112e3/trainer/checkpoint-546",
|
| 5 |
+
"epoch": 3.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 546,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.13793103448275862,
|
| 14 |
+
"grad_norm": 6.436005592346191,
|
| 15 |
+
"learning_rate": 1.3186813186813187e-06,
|
| 16 |
+
"loss": 4.4287,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.27586206896551724,
|
| 21 |
+
"grad_norm": 7.078612804412842,
|
| 22 |
+
"learning_rate": 2.6923076923076923e-06,
|
| 23 |
+
"loss": 4.3796,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.41379310344827586,
|
| 28 |
+
"grad_norm": 8.72610855102539,
|
| 29 |
+
"learning_rate": 4.065934065934066e-06,
|
| 30 |
+
"loss": 4.3507,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5517241379310345,
|
| 35 |
+
"grad_norm": 9.247332572937012,
|
| 36 |
+
"learning_rate": 5.43956043956044e-06,
|
| 37 |
+
"loss": 4.3268,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6896551724137931,
|
| 42 |
+
"grad_norm": 13.744568824768066,
|
| 43 |
+
"learning_rate": 6.813186813186814e-06,
|
| 44 |
+
"loss": 4.2635,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8275862068965517,
|
| 49 |
+
"grad_norm": 12.254353523254395,
|
| 50 |
+
"learning_rate": 8.186813186813188e-06,
|
| 51 |
+
"loss": 4.1413,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9655172413793104,
|
| 56 |
+
"grad_norm": 15.760031700134277,
|
| 57 |
+
"learning_rate": 9.560439560439562e-06,
|
| 58 |
+
"loss": 4.0484,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.5933014354066986,
|
| 64 |
+
"eval_loss": 0.9013376235961914,
|
| 65 |
+
"eval_macro_f1": 0.5891153112663168,
|
| 66 |
+
"eval_runtime": 0.4458,
|
| 67 |
+
"eval_samples_per_second": 1406.533,
|
| 68 |
+
"eval_steps_per_second": 89.731,
|
| 69 |
+
"eval_weighted_f1": 0.5938081589361177,
|
| 70 |
+
"step": 182
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0993103448275863,
|
| 74 |
+
"grad_norm": 19.441864013671875,
|
| 75 |
+
"learning_rate": 1.0934065934065933e-05,
|
| 76 |
+
"loss": 3.5205,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.237241379310345,
|
| 81 |
+
"grad_norm": 27.32406234741211,
|
| 82 |
+
"learning_rate": 1.230769230769231e-05,
|
| 83 |
+
"loss": 3.3318,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.3751724137931034,
|
| 88 |
+
"grad_norm": 48.73912811279297,
|
| 89 |
+
"learning_rate": 1.3681318681318683e-05,
|
| 90 |
+
"loss": 3.1308,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.513103448275862,
|
| 95 |
+
"grad_norm": 30.94043731689453,
|
| 96 |
+
"learning_rate": 1.5054945054945056e-05,
|
| 97 |
+
"loss": 2.7108,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.6510344827586207,
|
| 102 |
+
"grad_norm": 33.952392578125,
|
| 103 |
+
"learning_rate": 1.642857142857143e-05,
|
| 104 |
+
"loss": 2.5379,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7889655172413792,
|
| 109 |
+
"grad_norm": 49.711551666259766,
|
| 110 |
+
"learning_rate": 1.78021978021978e-05,
|
| 111 |
+
"loss": 2.4233,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.926896551724138,
|
| 116 |
+
"grad_norm": 51.16925048828125,
|
| 117 |
+
"learning_rate": 1.9175824175824178e-05,
|
| 118 |
+
"loss": 2.5303,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.7862838915470495,
|
| 124 |
+
"eval_loss": 0.5505752563476562,
|
| 125 |
+
"eval_macro_f1": 0.7851859232883802,
|
| 126 |
+
"eval_runtime": 0.4481,
|
| 127 |
+
"eval_samples_per_second": 1399.335,
|
| 128 |
+
"eval_steps_per_second": 89.272,
|
| 129 |
+
"eval_weighted_f1": 0.7857234351971534,
|
| 130 |
+
"step": 364
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.0606896551724136,
|
| 134 |
+
"grad_norm": 28.892581939697266,
|
| 135 |
+
"learning_rate": 1.993894993894994e-05,
|
| 136 |
+
"loss": 1.8631,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.1986206896551725,
|
| 141 |
+
"grad_norm": 43.200462341308594,
|
| 142 |
+
"learning_rate": 1.9786324786324788e-05,
|
| 143 |
+
"loss": 1.764,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.336551724137931,
|
| 148 |
+
"grad_norm": 34.97949981689453,
|
| 149 |
+
"learning_rate": 1.9633699633699634e-05,
|
| 150 |
+
"loss": 1.4838,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.47448275862069,
|
| 155 |
+
"grad_norm": 46.3452033996582,
|
| 156 |
+
"learning_rate": 1.948107448107448e-05,
|
| 157 |
+
"loss": 1.5003,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.6124137931034483,
|
| 162 |
+
"grad_norm": 38.658203125,
|
| 163 |
+
"learning_rate": 1.932844932844933e-05,
|
| 164 |
+
"loss": 1.4092,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.750344827586207,
|
| 169 |
+
"grad_norm": 39.02288055419922,
|
| 170 |
+
"learning_rate": 1.9175824175824178e-05,
|
| 171 |
+
"loss": 1.6422,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.8882758620689657,
|
| 176 |
+
"grad_norm": 49.1451530456543,
|
| 177 |
+
"learning_rate": 1.9023199023199025e-05,
|
| 178 |
+
"loss": 1.5683,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 3.0,
|
| 183 |
+
"eval_accuracy": 0.8421052631578947,
|
| 184 |
+
"eval_loss": 0.4859905242919922,
|
| 185 |
+
"eval_macro_f1": 0.8399379636589276,
|
| 186 |
+
"eval_runtime": 0.4384,
|
| 187 |
+
"eval_samples_per_second": 1430.07,
|
| 188 |
+
"eval_steps_per_second": 91.233,
|
| 189 |
+
"eval_weighted_f1": 0.8406701982792217,
|
| 190 |
+
"step": 546
|
| 191 |
+
}
|
| 192 |
+
],
|
| 193 |
+
"logging_steps": 25,
|
| 194 |
+
"max_steps": 3640,
|
| 195 |
+
"num_input_tokens_seen": 0,
|
| 196 |
+
"num_train_epochs": 20,
|
| 197 |
+
"save_steps": 500,
|
| 198 |
+
"stateful_callbacks": {
|
| 199 |
+
"EarlyStoppingCallback": {
|
| 200 |
+
"args": {
|
| 201 |
+
"early_stopping_patience": 3,
|
| 202 |
+
"early_stopping_threshold": 0.0
|
| 203 |
+
},
|
| 204 |
+
"attributes": {
|
| 205 |
+
"early_stopping_patience_counter": 0
|
| 206 |
+
}
|
| 207 |
+
},
|
| 208 |
+
"TrainerControl": {
|
| 209 |
+
"args": {
|
| 210 |
+
"should_epoch_stop": false,
|
| 211 |
+
"should_evaluate": false,
|
| 212 |
+
"should_log": false,
|
| 213 |
+
"should_save": true,
|
| 214 |
+
"should_training_stop": false
|
| 215 |
+
},
|
| 216 |
+
"attributes": {}
|
| 217 |
+
}
|
| 218 |
+
},
|
| 219 |
+
"total_flos": 0.0,
|
| 220 |
+
"train_batch_size": 8,
|
| 221 |
+
"trial_name": null,
|
| 222 |
+
"trial_params": null
|
| 223 |
+
}
|
ood-vifactcheck-phobert-source-source-00-s44-20260826T160106Z-ab4112e3/trainer/checkpoint-546/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-source-source-01-s44-20260826T160312Z-ab4112e3/trainer/checkpoint-1539/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-source-source-01-s44-20260826T160312Z-ab4112e3/trainer/checkpoint-1539/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-source-source-01-s44-20260826T160312Z-ab4112e3/trainer/checkpoint-1539/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-s44-20260826T160312Z-ab4112e3/trainer/checkpoint-1539/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-s44-20260826T160312Z-ab4112e3/trainer/checkpoint-1539/trainer_state.json
ADDED
|
@@ -0,0 +1,569 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1539,
|
| 3 |
+
"best_metric": 0.8391483513916458,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-source-source-01-s44-20260826T160312Z-ab4112e3/trainer/checkpoint-1539",
|
| 5 |
+
"epoch": 9.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1539,
|
| 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": 7.84555721282959,
|
| 15 |
+
"learning_rate": 1.4035087719298246e-06,
|
| 16 |
+
"loss": 4.4161,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.2936857562408223,
|
| 21 |
+
"grad_norm": 7.153261661529541,
|
| 22 |
+
"learning_rate": 2.865497076023392e-06,
|
| 23 |
+
"loss": 4.4047,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.44052863436123346,
|
| 28 |
+
"grad_norm": 7.9243245124816895,
|
| 29 |
+
"learning_rate": 4.3274853801169596e-06,
|
| 30 |
+
"loss": 4.3514,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5873715124816447,
|
| 35 |
+
"grad_norm": 9.712228775024414,
|
| 36 |
+
"learning_rate": 5.789473684210527e-06,
|
| 37 |
+
"loss": 4.3165,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.7342143906020558,
|
| 42 |
+
"grad_norm": 10.181035041809082,
|
| 43 |
+
"learning_rate": 7.251461988304094e-06,
|
| 44 |
+
"loss": 4.2481,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8810572687224669,
|
| 49 |
+
"grad_norm": 12.949995040893555,
|
| 50 |
+
"learning_rate": 8.713450292397661e-06,
|
| 51 |
+
"loss": 4.173,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 1.0,
|
| 56 |
+
"eval_accuracy": 0.5569395017793595,
|
| 57 |
+
"eval_loss": 0.9317622780799866,
|
| 58 |
+
"eval_macro_f1": 0.5355200679762083,
|
| 59 |
+
"eval_runtime": 0.4479,
|
| 60 |
+
"eval_samples_per_second": 1254.744,
|
| 61 |
+
"eval_steps_per_second": 80.375,
|
| 62 |
+
"eval_weighted_f1": 0.5384912744952078,
|
| 63 |
+
"step": 171
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 1.0234948604992657,
|
| 67 |
+
"grad_norm": 19.33929443359375,
|
| 68 |
+
"learning_rate": 1.017543859649123e-05,
|
| 69 |
+
"loss": 3.7871,
|
| 70 |
+
"step": 175
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.170337738619677,
|
| 74 |
+
"grad_norm": 38.31132888793945,
|
| 75 |
+
"learning_rate": 1.1637426900584797e-05,
|
| 76 |
+
"loss": 3.5028,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.3171806167400881,
|
| 81 |
+
"grad_norm": 22.062904357910156,
|
| 82 |
+
"learning_rate": 1.3099415204678362e-05,
|
| 83 |
+
"loss": 3.2352,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.4640234948604993,
|
| 88 |
+
"grad_norm": 34.893985748291016,
|
| 89 |
+
"learning_rate": 1.4561403508771931e-05,
|
| 90 |
+
"loss": 2.8678,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.6108663729809103,
|
| 95 |
+
"grad_norm": 52.86626434326172,
|
| 96 |
+
"learning_rate": 1.60233918128655e-05,
|
| 97 |
+
"loss": 2.8973,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.7577092511013217,
|
| 102 |
+
"grad_norm": 35.33583450317383,
|
| 103 |
+
"learning_rate": 1.7485380116959064e-05,
|
| 104 |
+
"loss": 2.692,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.9045521292217327,
|
| 109 |
+
"grad_norm": 38.24336242675781,
|
| 110 |
+
"learning_rate": 1.894736842105263e-05,
|
| 111 |
+
"loss": 2.4724,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 2.0,
|
| 116 |
+
"eval_accuracy": 0.8042704626334519,
|
| 117 |
+
"eval_loss": 0.5339668393135071,
|
| 118 |
+
"eval_macro_f1": 0.8027260489211816,
|
| 119 |
+
"eval_runtime": 0.5085,
|
| 120 |
+
"eval_samples_per_second": 1105.194,
|
| 121 |
+
"eval_steps_per_second": 70.795,
|
| 122 |
+
"eval_weighted_f1": 0.8031343064311335,
|
| 123 |
+
"step": 342
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 2.0469897209985315,
|
| 127 |
+
"grad_norm": 23.2413272857666,
|
| 128 |
+
"learning_rate": 1.99545159194282e-05,
|
| 129 |
+
"loss": 2.1682,
|
| 130 |
+
"step": 350
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.193832599118943,
|
| 134 |
+
"grad_norm": 41.5785026550293,
|
| 135 |
+
"learning_rate": 1.9792072774528916e-05,
|
| 136 |
+
"loss": 1.9912,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.340675477239354,
|
| 141 |
+
"grad_norm": 46.24933624267578,
|
| 142 |
+
"learning_rate": 1.962962962962963e-05,
|
| 143 |
+
"loss": 1.8267,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.4875183553597653,
|
| 148 |
+
"grad_norm": 20.881942749023438,
|
| 149 |
+
"learning_rate": 1.9467186484730346e-05,
|
| 150 |
+
"loss": 1.6831,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.6343612334801763,
|
| 155 |
+
"grad_norm": 41.941307067871094,
|
| 156 |
+
"learning_rate": 1.930474333983106e-05,
|
| 157 |
+
"loss": 1.653,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.7812041116005872,
|
| 162 |
+
"grad_norm": 35.23773956298828,
|
| 163 |
+
"learning_rate": 1.9142300194931776e-05,
|
| 164 |
+
"loss": 1.5744,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.9280469897209986,
|
| 169 |
+
"grad_norm": 36.22048568725586,
|
| 170 |
+
"learning_rate": 1.897985705003249e-05,
|
| 171 |
+
"loss": 1.5683,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 3.0,
|
| 176 |
+
"eval_accuracy": 0.7864768683274022,
|
| 177 |
+
"eval_loss": 0.5891265273094177,
|
| 178 |
+
"eval_macro_f1": 0.7769388468589041,
|
| 179 |
+
"eval_runtime": 0.4443,
|
| 180 |
+
"eval_samples_per_second": 1264.786,
|
| 181 |
+
"eval_steps_per_second": 81.018,
|
| 182 |
+
"eval_weighted_f1": 0.778110405009241,
|
| 183 |
+
"step": 513
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 3.0704845814977975,
|
| 187 |
+
"grad_norm": 27.708154678344727,
|
| 188 |
+
"learning_rate": 1.8817413905133203e-05,
|
| 189 |
+
"loss": 1.4858,
|
| 190 |
+
"step": 525
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.2173274596182084,
|
| 194 |
+
"grad_norm": 31.59416961669922,
|
| 195 |
+
"learning_rate": 1.8654970760233918e-05,
|
| 196 |
+
"loss": 1.1723,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.36417033773862,
|
| 201 |
+
"grad_norm": 24.580162048339844,
|
| 202 |
+
"learning_rate": 1.8492527615334633e-05,
|
| 203 |
+
"loss": 1.2132,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.511013215859031,
|
| 208 |
+
"grad_norm": 10.629317283630371,
|
| 209 |
+
"learning_rate": 1.833008447043535e-05,
|
| 210 |
+
"loss": 1.0242,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.657856093979442,
|
| 215 |
+
"grad_norm": 37.01762008666992,
|
| 216 |
+
"learning_rate": 1.8167641325536064e-05,
|
| 217 |
+
"loss": 1.077,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.804698972099853,
|
| 222 |
+
"grad_norm": 8.71078109741211,
|
| 223 |
+
"learning_rate": 1.800519818063678e-05,
|
| 224 |
+
"loss": 1.0395,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.951541850220264,
|
| 229 |
+
"grad_norm": 46.896827697753906,
|
| 230 |
+
"learning_rate": 1.7842755035737494e-05,
|
| 231 |
+
"loss": 1.1302,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 4.0,
|
| 236 |
+
"eval_accuracy": 0.798932384341637,
|
| 237 |
+
"eval_loss": 0.5832245945930481,
|
| 238 |
+
"eval_macro_f1": 0.7982622607469184,
|
| 239 |
+
"eval_runtime": 0.4217,
|
| 240 |
+
"eval_samples_per_second": 1332.83,
|
| 241 |
+
"eval_steps_per_second": 85.377,
|
| 242 |
+
"eval_weighted_f1": 0.7985408167179564,
|
| 243 |
+
"step": 684
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 4.093979441997063,
|
| 247 |
+
"grad_norm": 38.06997299194336,
|
| 248 |
+
"learning_rate": 1.768031189083821e-05,
|
| 249 |
+
"loss": 0.8855,
|
| 250 |
+
"step": 700
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 4.240822320117474,
|
| 254 |
+
"grad_norm": 42.93487548828125,
|
| 255 |
+
"learning_rate": 1.751786874593892e-05,
|
| 256 |
+
"loss": 0.7338,
|
| 257 |
+
"step": 725
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.387665198237886,
|
| 261 |
+
"grad_norm": 71.83784484863281,
|
| 262 |
+
"learning_rate": 1.7355425601039636e-05,
|
| 263 |
+
"loss": 0.7202,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.534508076358296,
|
| 268 |
+
"grad_norm": 10.435612678527832,
|
| 269 |
+
"learning_rate": 1.719298245614035e-05,
|
| 270 |
+
"loss": 0.908,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.681350954478708,
|
| 275 |
+
"grad_norm": 20.794347763061523,
|
| 276 |
+
"learning_rate": 1.7030539311241066e-05,
|
| 277 |
+
"loss": 0.7732,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.828193832599119,
|
| 282 |
+
"grad_norm": 14.409199714660645,
|
| 283 |
+
"learning_rate": 1.686809616634178e-05,
|
| 284 |
+
"loss": 0.8153,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.975036710719531,
|
| 289 |
+
"grad_norm": 39.207069396972656,
|
| 290 |
+
"learning_rate": 1.6705653021442497e-05,
|
| 291 |
+
"loss": 0.6247,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 5.0,
|
| 296 |
+
"eval_accuracy": 0.8131672597864769,
|
| 297 |
+
"eval_loss": 0.6464099884033203,
|
| 298 |
+
"eval_macro_f1": 0.8121675130442756,
|
| 299 |
+
"eval_runtime": 0.3928,
|
| 300 |
+
"eval_samples_per_second": 1430.824,
|
| 301 |
+
"eval_steps_per_second": 91.654,
|
| 302 |
+
"eval_weighted_f1": 0.8122333432812133,
|
| 303 |
+
"step": 855
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 5.117474302496329,
|
| 307 |
+
"grad_norm": 5.910846710205078,
|
| 308 |
+
"learning_rate": 1.654320987654321e-05,
|
| 309 |
+
"loss": 0.5436,
|
| 310 |
+
"step": 875
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 5.26431718061674,
|
| 314 |
+
"grad_norm": 33.61992645263672,
|
| 315 |
+
"learning_rate": 1.6380766731643927e-05,
|
| 316 |
+
"loss": 0.4557,
|
| 317 |
+
"step": 900
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.411160058737151,
|
| 321 |
+
"grad_norm": 70.12264251708984,
|
| 322 |
+
"learning_rate": 1.621832358674464e-05,
|
| 323 |
+
"loss": 0.6485,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.558002936857562,
|
| 328 |
+
"grad_norm": 96.00413513183594,
|
| 329 |
+
"learning_rate": 1.6055880441845354e-05,
|
| 330 |
+
"loss": 0.3341,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.704845814977974,
|
| 335 |
+
"grad_norm": 35.97297286987305,
|
| 336 |
+
"learning_rate": 1.589343729694607e-05,
|
| 337 |
+
"loss": 0.4215,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.851688693098385,
|
| 342 |
+
"grad_norm": 0.4652848541736603,
|
| 343 |
+
"learning_rate": 1.5730994152046787e-05,
|
| 344 |
+
"loss": 0.557,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.998531571218796,
|
| 349 |
+
"grad_norm": 35.06332015991211,
|
| 350 |
+
"learning_rate": 1.55685510071475e-05,
|
| 351 |
+
"loss": 0.6435,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 6.0,
|
| 356 |
+
"eval_accuracy": 0.8238434163701067,
|
| 357 |
+
"eval_loss": 0.7540084719657898,
|
| 358 |
+
"eval_macro_f1": 0.8249699264971574,
|
| 359 |
+
"eval_runtime": 0.4296,
|
| 360 |
+
"eval_samples_per_second": 1308.316,
|
| 361 |
+
"eval_steps_per_second": 83.807,
|
| 362 |
+
"eval_weighted_f1": 0.8248304790057952,
|
| 363 |
+
"step": 1026
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 6.140969162995595,
|
| 367 |
+
"grad_norm": 28.07526397705078,
|
| 368 |
+
"learning_rate": 1.5406107862248214e-05,
|
| 369 |
+
"loss": 0.4453,
|
| 370 |
+
"step": 1050
|
| 371 |
+
},
|
| 372 |
+
{
|
| 373 |
+
"epoch": 6.2878120411160054,
|
| 374 |
+
"grad_norm": 99.6798324584961,
|
| 375 |
+
"learning_rate": 1.524366471734893e-05,
|
| 376 |
+
"loss": 0.5713,
|
| 377 |
+
"step": 1075
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 6.434654919236417,
|
| 381 |
+
"grad_norm": 146.7027130126953,
|
| 382 |
+
"learning_rate": 1.5081221572449645e-05,
|
| 383 |
+
"loss": 0.3374,
|
| 384 |
+
"step": 1100
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 6.581497797356828,
|
| 388 |
+
"grad_norm": 41.28255844116211,
|
| 389 |
+
"learning_rate": 1.4918778427550358e-05,
|
| 390 |
+
"loss": 0.3599,
|
| 391 |
+
"step": 1125
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 6.72834067547724,
|
| 395 |
+
"grad_norm": 1.6257884502410889,
|
| 396 |
+
"learning_rate": 1.4756335282651073e-05,
|
| 397 |
+
"loss": 0.5842,
|
| 398 |
+
"step": 1150
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.87518355359765,
|
| 402 |
+
"grad_norm": 72.80775451660156,
|
| 403 |
+
"learning_rate": 1.4593892137751788e-05,
|
| 404 |
+
"loss": 0.4317,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 7.0,
|
| 409 |
+
"eval_accuracy": 0.8380782918149466,
|
| 410 |
+
"eval_loss": 0.7543966174125671,
|
| 411 |
+
"eval_macro_f1": 0.837147361933059,
|
| 412 |
+
"eval_runtime": 0.4215,
|
| 413 |
+
"eval_samples_per_second": 1333.358,
|
| 414 |
+
"eval_steps_per_second": 85.411,
|
| 415 |
+
"eval_weighted_f1": 0.8371001506761083,
|
| 416 |
+
"step": 1197
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 7.0176211453744495,
|
| 420 |
+
"grad_norm": 11.848440170288086,
|
| 421 |
+
"learning_rate": 1.4431448992852503e-05,
|
| 422 |
+
"loss": 0.3042,
|
| 423 |
+
"step": 1200
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 7.164464023494861,
|
| 427 |
+
"grad_norm": 1.9709209203720093,
|
| 428 |
+
"learning_rate": 1.4269005847953217e-05,
|
| 429 |
+
"loss": 0.4397,
|
| 430 |
+
"step": 1225
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 7.311306901615271,
|
| 434 |
+
"grad_norm": 4.237395286560059,
|
| 435 |
+
"learning_rate": 1.4106562703053932e-05,
|
| 436 |
+
"loss": 0.2646,
|
| 437 |
+
"step": 1250
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 7.458149779735683,
|
| 441 |
+
"grad_norm": 0.33251532912254333,
|
| 442 |
+
"learning_rate": 1.3944119558154647e-05,
|
| 443 |
+
"loss": 0.3262,
|
| 444 |
+
"step": 1275
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 7.604992657856094,
|
| 448 |
+
"grad_norm": 232.72943115234375,
|
| 449 |
+
"learning_rate": 1.3781676413255362e-05,
|
| 450 |
+
"loss": 0.2437,
|
| 451 |
+
"step": 1300
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 7.751835535976505,
|
| 455 |
+
"grad_norm": 3.8374176025390625,
|
| 456 |
+
"learning_rate": 1.3619233268356076e-05,
|
| 457 |
+
"loss": 0.2939,
|
| 458 |
+
"step": 1325
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 7.898678414096916,
|
| 462 |
+
"grad_norm": 61.01853942871094,
|
| 463 |
+
"learning_rate": 1.345679012345679e-05,
|
| 464 |
+
"loss": 0.2948,
|
| 465 |
+
"step": 1350
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 8.0,
|
| 469 |
+
"eval_accuracy": 0.8220640569395018,
|
| 470 |
+
"eval_loss": 0.9075856804847717,
|
| 471 |
+
"eval_macro_f1": 0.8217997889293627,
|
| 472 |
+
"eval_runtime": 0.4173,
|
| 473 |
+
"eval_samples_per_second": 1346.668,
|
| 474 |
+
"eval_steps_per_second": 86.263,
|
| 475 |
+
"eval_weighted_f1": 0.8218397979985076,
|
| 476 |
+
"step": 1368
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 8.041116005873715,
|
| 480 |
+
"grad_norm": 0.22358790040016174,
|
| 481 |
+
"learning_rate": 1.3294346978557506e-05,
|
| 482 |
+
"loss": 0.2681,
|
| 483 |
+
"step": 1375
|
| 484 |
+
},
|
| 485 |
+
{
|
| 486 |
+
"epoch": 8.187958883994126,
|
| 487 |
+
"grad_norm": 0.24839718639850616,
|
| 488 |
+
"learning_rate": 1.3131903833658221e-05,
|
| 489 |
+
"loss": 0.2614,
|
| 490 |
+
"step": 1400
|
| 491 |
+
},
|
| 492 |
+
{
|
| 493 |
+
"epoch": 8.334801762114537,
|
| 494 |
+
"grad_norm": 192.28289794921875,
|
| 495 |
+
"learning_rate": 1.2969460688758935e-05,
|
| 496 |
+
"loss": 0.2763,
|
| 497 |
+
"step": 1425
|
| 498 |
+
},
|
| 499 |
+
{
|
| 500 |
+
"epoch": 8.481644640234949,
|
| 501 |
+
"grad_norm": 90.91962432861328,
|
| 502 |
+
"learning_rate": 1.280701754385965e-05,
|
| 503 |
+
"loss": 0.2009,
|
| 504 |
+
"step": 1450
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 8.62848751835536,
|
| 508 |
+
"grad_norm": 1.8026515245437622,
|
| 509 |
+
"learning_rate": 1.2644574398960365e-05,
|
| 510 |
+
"loss": 0.2735,
|
| 511 |
+
"step": 1475
|
| 512 |
+
},
|
| 513 |
+
{
|
| 514 |
+
"epoch": 8.775330396475772,
|
| 515 |
+
"grad_norm": 0.24513249099254608,
|
| 516 |
+
"learning_rate": 1.248213125406108e-05,
|
| 517 |
+
"loss": 0.108,
|
| 518 |
+
"step": 1500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 8.922173274596181,
|
| 522 |
+
"grad_norm": 74.591796875,
|
| 523 |
+
"learning_rate": 1.2319688109161793e-05,
|
| 524 |
+
"loss": 0.2123,
|
| 525 |
+
"step": 1525
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 9.0,
|
| 529 |
+
"eval_accuracy": 0.8398576512455516,
|
| 530 |
+
"eval_loss": 0.8600338697433472,
|
| 531 |
+
"eval_macro_f1": 0.8391483513916458,
|
| 532 |
+
"eval_runtime": 0.4135,
|
| 533 |
+
"eval_samples_per_second": 1359.247,
|
| 534 |
+
"eval_steps_per_second": 87.069,
|
| 535 |
+
"eval_weighted_f1": 0.8390432699066734,
|
| 536 |
+
"step": 1539
|
| 537 |
+
}
|
| 538 |
+
],
|
| 539 |
+
"logging_steps": 25,
|
| 540 |
+
"max_steps": 3420,
|
| 541 |
+
"num_input_tokens_seen": 0,
|
| 542 |
+
"num_train_epochs": 20,
|
| 543 |
+
"save_steps": 500,
|
| 544 |
+
"stateful_callbacks": {
|
| 545 |
+
"EarlyStoppingCallback": {
|
| 546 |
+
"args": {
|
| 547 |
+
"early_stopping_patience": 3,
|
| 548 |
+
"early_stopping_threshold": 0.0
|
| 549 |
+
},
|
| 550 |
+
"attributes": {
|
| 551 |
+
"early_stopping_patience_counter": 0
|
| 552 |
+
}
|
| 553 |
+
},
|
| 554 |
+
"TrainerControl": {
|
| 555 |
+
"args": {
|
| 556 |
+
"should_epoch_stop": false,
|
| 557 |
+
"should_evaluate": false,
|
| 558 |
+
"should_log": false,
|
| 559 |
+
"should_save": true,
|
| 560 |
+
"should_training_stop": false
|
| 561 |
+
},
|
| 562 |
+
"attributes": {}
|
| 563 |
+
}
|
| 564 |
+
},
|
| 565 |
+
"total_flos": 0.0,
|
| 566 |
+
"train_batch_size": 8,
|
| 567 |
+
"trial_name": null,
|
| 568 |
+
"trial_params": null
|
| 569 |
+
}
|
ood-vifactcheck-phobert-source-source-01-s44-20260826T160312Z-ab4112e3/trainer/checkpoint-1539/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-source-source-01-s44-20260826T160312Z-ab4112e3/trainer/checkpoint-2052/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-source-source-01-s44-20260826T160312Z-ab4112e3/trainer/checkpoint-2052/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-source-source-01-s44-20260826T160312Z-ab4112e3/trainer/checkpoint-2052/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-s44-20260826T160312Z-ab4112e3/trainer/checkpoint-2052/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-s44-20260826T160312Z-ab4112e3/trainer/checkpoint-2052/trainer_state.json
ADDED
|
@@ -0,0 +1,749 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1539,
|
| 3 |
+
"best_metric": 0.8391483513916458,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-source-source-01-s44-20260826T160312Z-ab4112e3/trainer/checkpoint-1539",
|
| 5 |
+
"epoch": 12.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 2052,
|
| 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": 7.84555721282959,
|
| 15 |
+
"learning_rate": 1.4035087719298246e-06,
|
| 16 |
+
"loss": 4.4161,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.2936857562408223,
|
| 21 |
+
"grad_norm": 7.153261661529541,
|
| 22 |
+
"learning_rate": 2.865497076023392e-06,
|
| 23 |
+
"loss": 4.4047,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.44052863436123346,
|
| 28 |
+
"grad_norm": 7.9243245124816895,
|
| 29 |
+
"learning_rate": 4.3274853801169596e-06,
|
| 30 |
+
"loss": 4.3514,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5873715124816447,
|
| 35 |
+
"grad_norm": 9.712228775024414,
|
| 36 |
+
"learning_rate": 5.789473684210527e-06,
|
| 37 |
+
"loss": 4.3165,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.7342143906020558,
|
| 42 |
+
"grad_norm": 10.181035041809082,
|
| 43 |
+
"learning_rate": 7.251461988304094e-06,
|
| 44 |
+
"loss": 4.2481,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8810572687224669,
|
| 49 |
+
"grad_norm": 12.949995040893555,
|
| 50 |
+
"learning_rate": 8.713450292397661e-06,
|
| 51 |
+
"loss": 4.173,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 1.0,
|
| 56 |
+
"eval_accuracy": 0.5569395017793595,
|
| 57 |
+
"eval_loss": 0.9317622780799866,
|
| 58 |
+
"eval_macro_f1": 0.5355200679762083,
|
| 59 |
+
"eval_runtime": 0.4479,
|
| 60 |
+
"eval_samples_per_second": 1254.744,
|
| 61 |
+
"eval_steps_per_second": 80.375,
|
| 62 |
+
"eval_weighted_f1": 0.5384912744952078,
|
| 63 |
+
"step": 171
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 1.0234948604992657,
|
| 67 |
+
"grad_norm": 19.33929443359375,
|
| 68 |
+
"learning_rate": 1.017543859649123e-05,
|
| 69 |
+
"loss": 3.7871,
|
| 70 |
+
"step": 175
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.170337738619677,
|
| 74 |
+
"grad_norm": 38.31132888793945,
|
| 75 |
+
"learning_rate": 1.1637426900584797e-05,
|
| 76 |
+
"loss": 3.5028,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.3171806167400881,
|
| 81 |
+
"grad_norm": 22.062904357910156,
|
| 82 |
+
"learning_rate": 1.3099415204678362e-05,
|
| 83 |
+
"loss": 3.2352,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.4640234948604993,
|
| 88 |
+
"grad_norm": 34.893985748291016,
|
| 89 |
+
"learning_rate": 1.4561403508771931e-05,
|
| 90 |
+
"loss": 2.8678,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.6108663729809103,
|
| 95 |
+
"grad_norm": 52.86626434326172,
|
| 96 |
+
"learning_rate": 1.60233918128655e-05,
|
| 97 |
+
"loss": 2.8973,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.7577092511013217,
|
| 102 |
+
"grad_norm": 35.33583450317383,
|
| 103 |
+
"learning_rate": 1.7485380116959064e-05,
|
| 104 |
+
"loss": 2.692,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.9045521292217327,
|
| 109 |
+
"grad_norm": 38.24336242675781,
|
| 110 |
+
"learning_rate": 1.894736842105263e-05,
|
| 111 |
+
"loss": 2.4724,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 2.0,
|
| 116 |
+
"eval_accuracy": 0.8042704626334519,
|
| 117 |
+
"eval_loss": 0.5339668393135071,
|
| 118 |
+
"eval_macro_f1": 0.8027260489211816,
|
| 119 |
+
"eval_runtime": 0.5085,
|
| 120 |
+
"eval_samples_per_second": 1105.194,
|
| 121 |
+
"eval_steps_per_second": 70.795,
|
| 122 |
+
"eval_weighted_f1": 0.8031343064311335,
|
| 123 |
+
"step": 342
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 2.0469897209985315,
|
| 127 |
+
"grad_norm": 23.2413272857666,
|
| 128 |
+
"learning_rate": 1.99545159194282e-05,
|
| 129 |
+
"loss": 2.1682,
|
| 130 |
+
"step": 350
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.193832599118943,
|
| 134 |
+
"grad_norm": 41.5785026550293,
|
| 135 |
+
"learning_rate": 1.9792072774528916e-05,
|
| 136 |
+
"loss": 1.9912,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.340675477239354,
|
| 141 |
+
"grad_norm": 46.24933624267578,
|
| 142 |
+
"learning_rate": 1.962962962962963e-05,
|
| 143 |
+
"loss": 1.8267,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.4875183553597653,
|
| 148 |
+
"grad_norm": 20.881942749023438,
|
| 149 |
+
"learning_rate": 1.9467186484730346e-05,
|
| 150 |
+
"loss": 1.6831,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.6343612334801763,
|
| 155 |
+
"grad_norm": 41.941307067871094,
|
| 156 |
+
"learning_rate": 1.930474333983106e-05,
|
| 157 |
+
"loss": 1.653,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.7812041116005872,
|
| 162 |
+
"grad_norm": 35.23773956298828,
|
| 163 |
+
"learning_rate": 1.9142300194931776e-05,
|
| 164 |
+
"loss": 1.5744,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.9280469897209986,
|
| 169 |
+
"grad_norm": 36.22048568725586,
|
| 170 |
+
"learning_rate": 1.897985705003249e-05,
|
| 171 |
+
"loss": 1.5683,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 3.0,
|
| 176 |
+
"eval_accuracy": 0.7864768683274022,
|
| 177 |
+
"eval_loss": 0.5891265273094177,
|
| 178 |
+
"eval_macro_f1": 0.7769388468589041,
|
| 179 |
+
"eval_runtime": 0.4443,
|
| 180 |
+
"eval_samples_per_second": 1264.786,
|
| 181 |
+
"eval_steps_per_second": 81.018,
|
| 182 |
+
"eval_weighted_f1": 0.778110405009241,
|
| 183 |
+
"step": 513
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 3.0704845814977975,
|
| 187 |
+
"grad_norm": 27.708154678344727,
|
| 188 |
+
"learning_rate": 1.8817413905133203e-05,
|
| 189 |
+
"loss": 1.4858,
|
| 190 |
+
"step": 525
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.2173274596182084,
|
| 194 |
+
"grad_norm": 31.59416961669922,
|
| 195 |
+
"learning_rate": 1.8654970760233918e-05,
|
| 196 |
+
"loss": 1.1723,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.36417033773862,
|
| 201 |
+
"grad_norm": 24.580162048339844,
|
| 202 |
+
"learning_rate": 1.8492527615334633e-05,
|
| 203 |
+
"loss": 1.2132,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.511013215859031,
|
| 208 |
+
"grad_norm": 10.629317283630371,
|
| 209 |
+
"learning_rate": 1.833008447043535e-05,
|
| 210 |
+
"loss": 1.0242,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.657856093979442,
|
| 215 |
+
"grad_norm": 37.01762008666992,
|
| 216 |
+
"learning_rate": 1.8167641325536064e-05,
|
| 217 |
+
"loss": 1.077,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.804698972099853,
|
| 222 |
+
"grad_norm": 8.71078109741211,
|
| 223 |
+
"learning_rate": 1.800519818063678e-05,
|
| 224 |
+
"loss": 1.0395,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.951541850220264,
|
| 229 |
+
"grad_norm": 46.896827697753906,
|
| 230 |
+
"learning_rate": 1.7842755035737494e-05,
|
| 231 |
+
"loss": 1.1302,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 4.0,
|
| 236 |
+
"eval_accuracy": 0.798932384341637,
|
| 237 |
+
"eval_loss": 0.5832245945930481,
|
| 238 |
+
"eval_macro_f1": 0.7982622607469184,
|
| 239 |
+
"eval_runtime": 0.4217,
|
| 240 |
+
"eval_samples_per_second": 1332.83,
|
| 241 |
+
"eval_steps_per_second": 85.377,
|
| 242 |
+
"eval_weighted_f1": 0.7985408167179564,
|
| 243 |
+
"step": 684
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 4.093979441997063,
|
| 247 |
+
"grad_norm": 38.06997299194336,
|
| 248 |
+
"learning_rate": 1.768031189083821e-05,
|
| 249 |
+
"loss": 0.8855,
|
| 250 |
+
"step": 700
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 4.240822320117474,
|
| 254 |
+
"grad_norm": 42.93487548828125,
|
| 255 |
+
"learning_rate": 1.751786874593892e-05,
|
| 256 |
+
"loss": 0.7338,
|
| 257 |
+
"step": 725
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.387665198237886,
|
| 261 |
+
"grad_norm": 71.83784484863281,
|
| 262 |
+
"learning_rate": 1.7355425601039636e-05,
|
| 263 |
+
"loss": 0.7202,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.534508076358296,
|
| 268 |
+
"grad_norm": 10.435612678527832,
|
| 269 |
+
"learning_rate": 1.719298245614035e-05,
|
| 270 |
+
"loss": 0.908,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.681350954478708,
|
| 275 |
+
"grad_norm": 20.794347763061523,
|
| 276 |
+
"learning_rate": 1.7030539311241066e-05,
|
| 277 |
+
"loss": 0.7732,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.828193832599119,
|
| 282 |
+
"grad_norm": 14.409199714660645,
|
| 283 |
+
"learning_rate": 1.686809616634178e-05,
|
| 284 |
+
"loss": 0.8153,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.975036710719531,
|
| 289 |
+
"grad_norm": 39.207069396972656,
|
| 290 |
+
"learning_rate": 1.6705653021442497e-05,
|
| 291 |
+
"loss": 0.6247,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 5.0,
|
| 296 |
+
"eval_accuracy": 0.8131672597864769,
|
| 297 |
+
"eval_loss": 0.6464099884033203,
|
| 298 |
+
"eval_macro_f1": 0.8121675130442756,
|
| 299 |
+
"eval_runtime": 0.3928,
|
| 300 |
+
"eval_samples_per_second": 1430.824,
|
| 301 |
+
"eval_steps_per_second": 91.654,
|
| 302 |
+
"eval_weighted_f1": 0.8122333432812133,
|
| 303 |
+
"step": 855
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 5.117474302496329,
|
| 307 |
+
"grad_norm": 5.910846710205078,
|
| 308 |
+
"learning_rate": 1.654320987654321e-05,
|
| 309 |
+
"loss": 0.5436,
|
| 310 |
+
"step": 875
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 5.26431718061674,
|
| 314 |
+
"grad_norm": 33.61992645263672,
|
| 315 |
+
"learning_rate": 1.6380766731643927e-05,
|
| 316 |
+
"loss": 0.4557,
|
| 317 |
+
"step": 900
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.411160058737151,
|
| 321 |
+
"grad_norm": 70.12264251708984,
|
| 322 |
+
"learning_rate": 1.621832358674464e-05,
|
| 323 |
+
"loss": 0.6485,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.558002936857562,
|
| 328 |
+
"grad_norm": 96.00413513183594,
|
| 329 |
+
"learning_rate": 1.6055880441845354e-05,
|
| 330 |
+
"loss": 0.3341,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.704845814977974,
|
| 335 |
+
"grad_norm": 35.97297286987305,
|
| 336 |
+
"learning_rate": 1.589343729694607e-05,
|
| 337 |
+
"loss": 0.4215,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.851688693098385,
|
| 342 |
+
"grad_norm": 0.4652848541736603,
|
| 343 |
+
"learning_rate": 1.5730994152046787e-05,
|
| 344 |
+
"loss": 0.557,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.998531571218796,
|
| 349 |
+
"grad_norm": 35.06332015991211,
|
| 350 |
+
"learning_rate": 1.55685510071475e-05,
|
| 351 |
+
"loss": 0.6435,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 6.0,
|
| 356 |
+
"eval_accuracy": 0.8238434163701067,
|
| 357 |
+
"eval_loss": 0.7540084719657898,
|
| 358 |
+
"eval_macro_f1": 0.8249699264971574,
|
| 359 |
+
"eval_runtime": 0.4296,
|
| 360 |
+
"eval_samples_per_second": 1308.316,
|
| 361 |
+
"eval_steps_per_second": 83.807,
|
| 362 |
+
"eval_weighted_f1": 0.8248304790057952,
|
| 363 |
+
"step": 1026
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 6.140969162995595,
|
| 367 |
+
"grad_norm": 28.07526397705078,
|
| 368 |
+
"learning_rate": 1.5406107862248214e-05,
|
| 369 |
+
"loss": 0.4453,
|
| 370 |
+
"step": 1050
|
| 371 |
+
},
|
| 372 |
+
{
|
| 373 |
+
"epoch": 6.2878120411160054,
|
| 374 |
+
"grad_norm": 99.6798324584961,
|
| 375 |
+
"learning_rate": 1.524366471734893e-05,
|
| 376 |
+
"loss": 0.5713,
|
| 377 |
+
"step": 1075
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 6.434654919236417,
|
| 381 |
+
"grad_norm": 146.7027130126953,
|
| 382 |
+
"learning_rate": 1.5081221572449645e-05,
|
| 383 |
+
"loss": 0.3374,
|
| 384 |
+
"step": 1100
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 6.581497797356828,
|
| 388 |
+
"grad_norm": 41.28255844116211,
|
| 389 |
+
"learning_rate": 1.4918778427550358e-05,
|
| 390 |
+
"loss": 0.3599,
|
| 391 |
+
"step": 1125
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 6.72834067547724,
|
| 395 |
+
"grad_norm": 1.6257884502410889,
|
| 396 |
+
"learning_rate": 1.4756335282651073e-05,
|
| 397 |
+
"loss": 0.5842,
|
| 398 |
+
"step": 1150
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.87518355359765,
|
| 402 |
+
"grad_norm": 72.80775451660156,
|
| 403 |
+
"learning_rate": 1.4593892137751788e-05,
|
| 404 |
+
"loss": 0.4317,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 7.0,
|
| 409 |
+
"eval_accuracy": 0.8380782918149466,
|
| 410 |
+
"eval_loss": 0.7543966174125671,
|
| 411 |
+
"eval_macro_f1": 0.837147361933059,
|
| 412 |
+
"eval_runtime": 0.4215,
|
| 413 |
+
"eval_samples_per_second": 1333.358,
|
| 414 |
+
"eval_steps_per_second": 85.411,
|
| 415 |
+
"eval_weighted_f1": 0.8371001506761083,
|
| 416 |
+
"step": 1197
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 7.0176211453744495,
|
| 420 |
+
"grad_norm": 11.848440170288086,
|
| 421 |
+
"learning_rate": 1.4431448992852503e-05,
|
| 422 |
+
"loss": 0.3042,
|
| 423 |
+
"step": 1200
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 7.164464023494861,
|
| 427 |
+
"grad_norm": 1.9709209203720093,
|
| 428 |
+
"learning_rate": 1.4269005847953217e-05,
|
| 429 |
+
"loss": 0.4397,
|
| 430 |
+
"step": 1225
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 7.311306901615271,
|
| 434 |
+
"grad_norm": 4.237395286560059,
|
| 435 |
+
"learning_rate": 1.4106562703053932e-05,
|
| 436 |
+
"loss": 0.2646,
|
| 437 |
+
"step": 1250
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 7.458149779735683,
|
| 441 |
+
"grad_norm": 0.33251532912254333,
|
| 442 |
+
"learning_rate": 1.3944119558154647e-05,
|
| 443 |
+
"loss": 0.3262,
|
| 444 |
+
"step": 1275
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 7.604992657856094,
|
| 448 |
+
"grad_norm": 232.72943115234375,
|
| 449 |
+
"learning_rate": 1.3781676413255362e-05,
|
| 450 |
+
"loss": 0.2437,
|
| 451 |
+
"step": 1300
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 7.751835535976505,
|
| 455 |
+
"grad_norm": 3.8374176025390625,
|
| 456 |
+
"learning_rate": 1.3619233268356076e-05,
|
| 457 |
+
"loss": 0.2939,
|
| 458 |
+
"step": 1325
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 7.898678414096916,
|
| 462 |
+
"grad_norm": 61.01853942871094,
|
| 463 |
+
"learning_rate": 1.345679012345679e-05,
|
| 464 |
+
"loss": 0.2948,
|
| 465 |
+
"step": 1350
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 8.0,
|
| 469 |
+
"eval_accuracy": 0.8220640569395018,
|
| 470 |
+
"eval_loss": 0.9075856804847717,
|
| 471 |
+
"eval_macro_f1": 0.8217997889293627,
|
| 472 |
+
"eval_runtime": 0.4173,
|
| 473 |
+
"eval_samples_per_second": 1346.668,
|
| 474 |
+
"eval_steps_per_second": 86.263,
|
| 475 |
+
"eval_weighted_f1": 0.8218397979985076,
|
| 476 |
+
"step": 1368
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 8.041116005873715,
|
| 480 |
+
"grad_norm": 0.22358790040016174,
|
| 481 |
+
"learning_rate": 1.3294346978557506e-05,
|
| 482 |
+
"loss": 0.2681,
|
| 483 |
+
"step": 1375
|
| 484 |
+
},
|
| 485 |
+
{
|
| 486 |
+
"epoch": 8.187958883994126,
|
| 487 |
+
"grad_norm": 0.24839718639850616,
|
| 488 |
+
"learning_rate": 1.3131903833658221e-05,
|
| 489 |
+
"loss": 0.2614,
|
| 490 |
+
"step": 1400
|
| 491 |
+
},
|
| 492 |
+
{
|
| 493 |
+
"epoch": 8.334801762114537,
|
| 494 |
+
"grad_norm": 192.28289794921875,
|
| 495 |
+
"learning_rate": 1.2969460688758935e-05,
|
| 496 |
+
"loss": 0.2763,
|
| 497 |
+
"step": 1425
|
| 498 |
+
},
|
| 499 |
+
{
|
| 500 |
+
"epoch": 8.481644640234949,
|
| 501 |
+
"grad_norm": 90.91962432861328,
|
| 502 |
+
"learning_rate": 1.280701754385965e-05,
|
| 503 |
+
"loss": 0.2009,
|
| 504 |
+
"step": 1450
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 8.62848751835536,
|
| 508 |
+
"grad_norm": 1.8026515245437622,
|
| 509 |
+
"learning_rate": 1.2644574398960365e-05,
|
| 510 |
+
"loss": 0.2735,
|
| 511 |
+
"step": 1475
|
| 512 |
+
},
|
| 513 |
+
{
|
| 514 |
+
"epoch": 8.775330396475772,
|
| 515 |
+
"grad_norm": 0.24513249099254608,
|
| 516 |
+
"learning_rate": 1.248213125406108e-05,
|
| 517 |
+
"loss": 0.108,
|
| 518 |
+
"step": 1500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 8.922173274596181,
|
| 522 |
+
"grad_norm": 74.591796875,
|
| 523 |
+
"learning_rate": 1.2319688109161793e-05,
|
| 524 |
+
"loss": 0.2123,
|
| 525 |
+
"step": 1525
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 9.0,
|
| 529 |
+
"eval_accuracy": 0.8398576512455516,
|
| 530 |
+
"eval_loss": 0.8600338697433472,
|
| 531 |
+
"eval_macro_f1": 0.8391483513916458,
|
| 532 |
+
"eval_runtime": 0.4135,
|
| 533 |
+
"eval_samples_per_second": 1359.247,
|
| 534 |
+
"eval_steps_per_second": 87.069,
|
| 535 |
+
"eval_weighted_f1": 0.8390432699066734,
|
| 536 |
+
"step": 1539
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 9.06461086637298,
|
| 540 |
+
"grad_norm": 6.55684232711792,
|
| 541 |
+
"learning_rate": 1.2157244964262509e-05,
|
| 542 |
+
"loss": 0.0563,
|
| 543 |
+
"step": 1550
|
| 544 |
+
},
|
| 545 |
+
{
|
| 546 |
+
"epoch": 9.211453744493392,
|
| 547 |
+
"grad_norm": 203.81256103515625,
|
| 548 |
+
"learning_rate": 1.1994801819363224e-05,
|
| 549 |
+
"loss": 0.1337,
|
| 550 |
+
"step": 1575
|
| 551 |
+
},
|
| 552 |
+
{
|
| 553 |
+
"epoch": 9.358296622613803,
|
| 554 |
+
"grad_norm": 0.15012528002262115,
|
| 555 |
+
"learning_rate": 1.1832358674463939e-05,
|
| 556 |
+
"loss": 0.2038,
|
| 557 |
+
"step": 1600
|
| 558 |
+
},
|
| 559 |
+
{
|
| 560 |
+
"epoch": 9.505139500734215,
|
| 561 |
+
"grad_norm": 0.14163698256015778,
|
| 562 |
+
"learning_rate": 1.1669915529564652e-05,
|
| 563 |
+
"loss": 0.1833,
|
| 564 |
+
"step": 1625
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"epoch": 9.651982378854626,
|
| 568 |
+
"grad_norm": 0.4035705626010895,
|
| 569 |
+
"learning_rate": 1.1507472384665367e-05,
|
| 570 |
+
"loss": 0.145,
|
| 571 |
+
"step": 1650
|
| 572 |
+
},
|
| 573 |
+
{
|
| 574 |
+
"epoch": 9.798825256975036,
|
| 575 |
+
"grad_norm": 1.7409424781799316,
|
| 576 |
+
"learning_rate": 1.1345029239766083e-05,
|
| 577 |
+
"loss": 0.218,
|
| 578 |
+
"step": 1675
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 9.945668135095447,
|
| 582 |
+
"grad_norm": 85.52859497070312,
|
| 583 |
+
"learning_rate": 1.11825860948668e-05,
|
| 584 |
+
"loss": 0.2671,
|
| 585 |
+
"step": 1700
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 10.0,
|
| 589 |
+
"eval_accuracy": 0.8327402135231317,
|
| 590 |
+
"eval_loss": 0.9421208500862122,
|
| 591 |
+
"eval_macro_f1": 0.833096720881097,
|
| 592 |
+
"eval_runtime": 0.4112,
|
| 593 |
+
"eval_samples_per_second": 1366.831,
|
| 594 |
+
"eval_steps_per_second": 87.555,
|
| 595 |
+
"eval_weighted_f1": 0.83299911180292,
|
| 596 |
+
"step": 1710
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 10.088105726872246,
|
| 600 |
+
"grad_norm": 0.18047431111335754,
|
| 601 |
+
"learning_rate": 1.1020142949967511e-05,
|
| 602 |
+
"loss": 0.2654,
|
| 603 |
+
"step": 1725
|
| 604 |
+
},
|
| 605 |
+
{
|
| 606 |
+
"epoch": 10.234948604992658,
|
| 607 |
+
"grad_norm": 0.21616902947425842,
|
| 608 |
+
"learning_rate": 1.0857699805068226e-05,
|
| 609 |
+
"loss": 0.2162,
|
| 610 |
+
"step": 1750
|
| 611 |
+
},
|
| 612 |
+
{
|
| 613 |
+
"epoch": 10.38179148311307,
|
| 614 |
+
"grad_norm": 0.13655419647693634,
|
| 615 |
+
"learning_rate": 1.0695256660168943e-05,
|
| 616 |
+
"loss": 0.0749,
|
| 617 |
+
"step": 1775
|
| 618 |
+
},
|
| 619 |
+
{
|
| 620 |
+
"epoch": 10.52863436123348,
|
| 621 |
+
"grad_norm": 0.1132725328207016,
|
| 622 |
+
"learning_rate": 1.0532813515269658e-05,
|
| 623 |
+
"loss": 0.1316,
|
| 624 |
+
"step": 1800
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"epoch": 10.675477239353892,
|
| 628 |
+
"grad_norm": 0.12083166092634201,
|
| 629 |
+
"learning_rate": 1.037037037037037e-05,
|
| 630 |
+
"loss": 0.1238,
|
| 631 |
+
"step": 1825
|
| 632 |
+
},
|
| 633 |
+
{
|
| 634 |
+
"epoch": 10.822320117474302,
|
| 635 |
+
"grad_norm": 0.6282656788825989,
|
| 636 |
+
"learning_rate": 1.0207927225471085e-05,
|
| 637 |
+
"loss": 0.2175,
|
| 638 |
+
"step": 1850
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 10.969162995594713,
|
| 642 |
+
"grad_norm": 270.02667236328125,
|
| 643 |
+
"learning_rate": 1.0045484080571802e-05,
|
| 644 |
+
"loss": 0.2554,
|
| 645 |
+
"step": 1875
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 11.0,
|
| 649 |
+
"eval_accuracy": 0.8274021352313167,
|
| 650 |
+
"eval_loss": 0.9721890091896057,
|
| 651 |
+
"eval_macro_f1": 0.8275243275243275,
|
| 652 |
+
"eval_runtime": 0.5207,
|
| 653 |
+
"eval_samples_per_second": 1079.268,
|
| 654 |
+
"eval_steps_per_second": 69.135,
|
| 655 |
+
"eval_weighted_f1": 0.8274250462862561,
|
| 656 |
+
"step": 1881
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 11.111600587371512,
|
| 660 |
+
"grad_norm": 342.13433837890625,
|
| 661 |
+
"learning_rate": 9.883040935672515e-06,
|
| 662 |
+
"loss": 0.1666,
|
| 663 |
+
"step": 1900
|
| 664 |
+
},
|
| 665 |
+
{
|
| 666 |
+
"epoch": 11.258443465491924,
|
| 667 |
+
"grad_norm": 0.12337096035480499,
|
| 668 |
+
"learning_rate": 9.720597790773229e-06,
|
| 669 |
+
"loss": 0.071,
|
| 670 |
+
"step": 1925
|
| 671 |
+
},
|
| 672 |
+
{
|
| 673 |
+
"epoch": 11.405286343612335,
|
| 674 |
+
"grad_norm": 4.025733947753906,
|
| 675 |
+
"learning_rate": 9.558154645873946e-06,
|
| 676 |
+
"loss": 0.1661,
|
| 677 |
+
"step": 1950
|
| 678 |
+
},
|
| 679 |
+
{
|
| 680 |
+
"epoch": 11.552129221732747,
|
| 681 |
+
"grad_norm": 0.10743283480405807,
|
| 682 |
+
"learning_rate": 9.395711500974659e-06,
|
| 683 |
+
"loss": 0.0464,
|
| 684 |
+
"step": 1975
|
| 685 |
+
},
|
| 686 |
+
{
|
| 687 |
+
"epoch": 11.698972099853156,
|
| 688 |
+
"grad_norm": 4.657911777496338,
|
| 689 |
+
"learning_rate": 9.233268356075374e-06,
|
| 690 |
+
"loss": 0.2062,
|
| 691 |
+
"step": 2000
|
| 692 |
+
},
|
| 693 |
+
{
|
| 694 |
+
"epoch": 11.845814977973568,
|
| 695 |
+
"grad_norm": 0.10279026627540588,
|
| 696 |
+
"learning_rate": 9.07082521117609e-06,
|
| 697 |
+
"loss": 0.1156,
|
| 698 |
+
"step": 2025
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 11.99265785609398,
|
| 702 |
+
"grad_norm": 0.11668312549591064,
|
| 703 |
+
"learning_rate": 8.908382066276805e-06,
|
| 704 |
+
"loss": 0.1259,
|
| 705 |
+
"step": 2050
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"epoch": 12.0,
|
| 709 |
+
"eval_accuracy": 0.8291814946619217,
|
| 710 |
+
"eval_loss": 1.0024290084838867,
|
| 711 |
+
"eval_macro_f1": 0.8286282720735,
|
| 712 |
+
"eval_runtime": 0.4095,
|
| 713 |
+
"eval_samples_per_second": 1372.283,
|
| 714 |
+
"eval_steps_per_second": 87.904,
|
| 715 |
+
"eval_weighted_f1": 0.8284554881506117,
|
| 716 |
+
"step": 2052
|
| 717 |
+
}
|
| 718 |
+
],
|
| 719 |
+
"logging_steps": 25,
|
| 720 |
+
"max_steps": 3420,
|
| 721 |
+
"num_input_tokens_seen": 0,
|
| 722 |
+
"num_train_epochs": 20,
|
| 723 |
+
"save_steps": 500,
|
| 724 |
+
"stateful_callbacks": {
|
| 725 |
+
"EarlyStoppingCallback": {
|
| 726 |
+
"args": {
|
| 727 |
+
"early_stopping_patience": 3,
|
| 728 |
+
"early_stopping_threshold": 0.0
|
| 729 |
+
},
|
| 730 |
+
"attributes": {
|
| 731 |
+
"early_stopping_patience_counter": 3
|
| 732 |
+
}
|
| 733 |
+
},
|
| 734 |
+
"TrainerControl": {
|
| 735 |
+
"args": {
|
| 736 |
+
"should_epoch_stop": false,
|
| 737 |
+
"should_evaluate": false,
|
| 738 |
+
"should_log": false,
|
| 739 |
+
"should_save": true,
|
| 740 |
+
"should_training_stop": true
|
| 741 |
+
},
|
| 742 |
+
"attributes": {}
|
| 743 |
+
}
|
| 744 |
+
},
|
| 745 |
+
"total_flos": 0.0,
|
| 746 |
+
"train_batch_size": 8,
|
| 747 |
+
"trial_name": null,
|
| 748 |
+
"trial_params": null
|
| 749 |
+
}
|
ood-vifactcheck-phobert-source-source-01-s44-20260826T160312Z-ab4112e3/trainer/checkpoint-2052/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-source-source-02-s44-20260826T160702Z-ab4112e3/trainer/checkpoint-1086/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-source-source-02-s44-20260826T160702Z-ab4112e3/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-s44-20260826T160702Z-ab4112e3/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-s44-20260826T160702Z-ab4112e3/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-s44-20260826T160702Z-ab4112e3/trainer/checkpoint-1086/trainer_state.json
ADDED
|
@@ -0,0 +1,410 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1086,
|
| 3 |
+
"best_metric": 0.8406598784091951,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-source-source-02-s44-20260826T160702Z-ab4112e3/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": 6.044527530670166,
|
| 15 |
+
"learning_rate": 1.3259668508287293e-06,
|
| 16 |
+
"loss": 4.4406,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.27624309392265195,
|
| 21 |
+
"grad_norm": 7.572896957397461,
|
| 22 |
+
"learning_rate": 2.707182320441989e-06,
|
| 23 |
+
"loss": 4.3821,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.4143646408839779,
|
| 28 |
+
"grad_norm": 7.113949775695801,
|
| 29 |
+
"learning_rate": 4.088397790055249e-06,
|
| 30 |
+
"loss": 4.3729,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5524861878453039,
|
| 35 |
+
"grad_norm": 6.951432228088379,
|
| 36 |
+
"learning_rate": 5.469613259668509e-06,
|
| 37 |
+
"loss": 4.2956,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6906077348066298,
|
| 42 |
+
"grad_norm": 8.143998146057129,
|
| 43 |
+
"learning_rate": 6.850828729281769e-06,
|
| 44 |
+
"loss": 4.2841,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8287292817679558,
|
| 49 |
+
"grad_norm": 18.067462921142578,
|
| 50 |
+
"learning_rate": 8.232044198895029e-06,
|
| 51 |
+
"loss": 4.0918,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9668508287292817,
|
| 56 |
+
"grad_norm": 22.35405731201172,
|
| 57 |
+
"learning_rate": 9.613259668508288e-06,
|
| 58 |
+
"loss": 3.9472,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.5807560137457045,
|
| 64 |
+
"eval_loss": 0.8864754438400269,
|
| 65 |
+
"eval_macro_f1": 0.5673594939188857,
|
| 66 |
+
"eval_runtime": 0.404,
|
| 67 |
+
"eval_samples_per_second": 1440.51,
|
| 68 |
+
"eval_steps_per_second": 91.579,
|
| 69 |
+
"eval_weighted_f1": 0.572302371883713,
|
| 70 |
+
"step": 181
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.1049723756906078,
|
| 74 |
+
"grad_norm": 25.207218170166016,
|
| 75 |
+
"learning_rate": 1.0994475138121547e-05,
|
| 76 |
+
"loss": 3.653,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.2430939226519337,
|
| 81 |
+
"grad_norm": 16.948280334472656,
|
| 82 |
+
"learning_rate": 1.2375690607734806e-05,
|
| 83 |
+
"loss": 3.4973,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.3812154696132597,
|
| 88 |
+
"grad_norm": 49.38298797607422,
|
| 89 |
+
"learning_rate": 1.3756906077348067e-05,
|
| 90 |
+
"loss": 3.2864,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.5193370165745856,
|
| 95 |
+
"grad_norm": 43.83489990234375,
|
| 96 |
+
"learning_rate": 1.5138121546961326e-05,
|
| 97 |
+
"loss": 2.9535,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.6574585635359116,
|
| 102 |
+
"grad_norm": 27.498220443725586,
|
| 103 |
+
"learning_rate": 1.6519337016574586e-05,
|
| 104 |
+
"loss": 2.496,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7955801104972375,
|
| 109 |
+
"grad_norm": 45.16746139526367,
|
| 110 |
+
"learning_rate": 1.7900552486187846e-05,
|
| 111 |
+
"loss": 2.4388,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.9337016574585635,
|
| 116 |
+
"grad_norm": 30.030315399169922,
|
| 117 |
+
"learning_rate": 1.9281767955801107e-05,
|
| 118 |
+
"loss": 2.3232,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.7869415807560137,
|
| 124 |
+
"eval_loss": 0.5499523282051086,
|
| 125 |
+
"eval_macro_f1": 0.7847817494176726,
|
| 126 |
+
"eval_runtime": 0.383,
|
| 127 |
+
"eval_samples_per_second": 1519.528,
|
| 128 |
+
"eval_steps_per_second": 96.602,
|
| 129 |
+
"eval_weighted_f1": 0.7861326943126823,
|
| 130 |
+
"step": 362
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.0718232044198897,
|
| 134 |
+
"grad_norm": 18.682743072509766,
|
| 135 |
+
"learning_rate": 1.9926335174953962e-05,
|
| 136 |
+
"loss": 1.9575,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.2099447513812156,
|
| 141 |
+
"grad_norm": 33.75157165527344,
|
| 142 |
+
"learning_rate": 1.9772866789441377e-05,
|
| 143 |
+
"loss": 1.7536,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.3480662983425415,
|
| 148 |
+
"grad_norm": 50.764705657958984,
|
| 149 |
+
"learning_rate": 1.9619398403928794e-05,
|
| 150 |
+
"loss": 1.6032,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.4861878453038675,
|
| 155 |
+
"grad_norm": 61.391056060791016,
|
| 156 |
+
"learning_rate": 1.946593001841621e-05,
|
| 157 |
+
"loss": 1.8642,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.6243093922651934,
|
| 162 |
+
"grad_norm": 31.565210342407227,
|
| 163 |
+
"learning_rate": 1.9312461632903623e-05,
|
| 164 |
+
"loss": 1.7283,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.7624309392265194,
|
| 169 |
+
"grad_norm": 34.41781234741211,
|
| 170 |
+
"learning_rate": 1.915899324739104e-05,
|
| 171 |
+
"loss": 1.6703,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.9005524861878453,
|
| 176 |
+
"grad_norm": 36.9642219543457,
|
| 177 |
+
"learning_rate": 1.9005524861878455e-05,
|
| 178 |
+
"loss": 1.5379,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 3.0,
|
| 183 |
+
"eval_accuracy": 0.8006872852233677,
|
| 184 |
+
"eval_loss": 0.5332517623901367,
|
| 185 |
+
"eval_macro_f1": 0.7970787408095045,
|
| 186 |
+
"eval_runtime": 0.4492,
|
| 187 |
+
"eval_samples_per_second": 1295.696,
|
| 188 |
+
"eval_steps_per_second": 82.372,
|
| 189 |
+
"eval_weighted_f1": 0.798666263369714,
|
| 190 |
+
"step": 543
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.0386740331491713,
|
| 194 |
+
"grad_norm": 21.794315338134766,
|
| 195 |
+
"learning_rate": 1.885205647636587e-05,
|
| 196 |
+
"loss": 1.5075,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.1767955801104972,
|
| 201 |
+
"grad_norm": 32.58504104614258,
|
| 202 |
+
"learning_rate": 1.8698588090853287e-05,
|
| 203 |
+
"loss": 1.0903,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.314917127071823,
|
| 208 |
+
"grad_norm": 31.611621856689453,
|
| 209 |
+
"learning_rate": 1.85451197053407e-05,
|
| 210 |
+
"loss": 0.9415,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.453038674033149,
|
| 215 |
+
"grad_norm": 29.14419937133789,
|
| 216 |
+
"learning_rate": 1.8391651319828116e-05,
|
| 217 |
+
"loss": 1.1406,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.591160220994475,
|
| 222 |
+
"grad_norm": 45.6563606262207,
|
| 223 |
+
"learning_rate": 1.823818293431553e-05,
|
| 224 |
+
"loss": 1.1686,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.729281767955801,
|
| 229 |
+
"grad_norm": 4.467889785766602,
|
| 230 |
+
"learning_rate": 1.8084714548802948e-05,
|
| 231 |
+
"loss": 0.9831,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.867403314917127,
|
| 236 |
+
"grad_norm": 89.3020248413086,
|
| 237 |
+
"learning_rate": 1.7931246163290362e-05,
|
| 238 |
+
"loss": 1.2411,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 4.0,
|
| 243 |
+
"eval_accuracy": 0.8247422680412371,
|
| 244 |
+
"eval_loss": 0.517365038394928,
|
| 245 |
+
"eval_macro_f1": 0.8222884518956111,
|
| 246 |
+
"eval_runtime": 0.4849,
|
| 247 |
+
"eval_samples_per_second": 1200.262,
|
| 248 |
+
"eval_steps_per_second": 76.305,
|
| 249 |
+
"eval_weighted_f1": 0.823348778311946,
|
| 250 |
+
"step": 724
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 4.005524861878453,
|
| 254 |
+
"grad_norm": 67.46331787109375,
|
| 255 |
+
"learning_rate": 1.7777777777777777e-05,
|
| 256 |
+
"loss": 1.0388,
|
| 257 |
+
"step": 725
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.143646408839779,
|
| 261 |
+
"grad_norm": 20.47144317626953,
|
| 262 |
+
"learning_rate": 1.7624309392265194e-05,
|
| 263 |
+
"loss": 0.634,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.281767955801105,
|
| 268 |
+
"grad_norm": 108.18099212646484,
|
| 269 |
+
"learning_rate": 1.747084100675261e-05,
|
| 270 |
+
"loss": 0.7141,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.419889502762431,
|
| 275 |
+
"grad_norm": 113.71337890625,
|
| 276 |
+
"learning_rate": 1.7317372621240026e-05,
|
| 277 |
+
"loss": 0.8237,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.558011049723757,
|
| 282 |
+
"grad_norm": 49.596534729003906,
|
| 283 |
+
"learning_rate": 1.716390423572744e-05,
|
| 284 |
+
"loss": 0.6921,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.696132596685083,
|
| 289 |
+
"grad_norm": 42.291099548339844,
|
| 290 |
+
"learning_rate": 1.701043585021486e-05,
|
| 291 |
+
"loss": 0.8426,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.834254143646409,
|
| 296 |
+
"grad_norm": 53.41448974609375,
|
| 297 |
+
"learning_rate": 1.6856967464702273e-05,
|
| 298 |
+
"loss": 0.7432,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.972375690607735,
|
| 303 |
+
"grad_norm": 20.884723663330078,
|
| 304 |
+
"learning_rate": 1.670349907918969e-05,
|
| 305 |
+
"loss": 0.6546,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 5.0,
|
| 310 |
+
"eval_accuracy": 0.8127147766323024,
|
| 311 |
+
"eval_loss": 0.6466218829154968,
|
| 312 |
+
"eval_macro_f1": 0.8103961419622757,
|
| 313 |
+
"eval_runtime": 0.4249,
|
| 314 |
+
"eval_samples_per_second": 1369.574,
|
| 315 |
+
"eval_steps_per_second": 87.069,
|
| 316 |
+
"eval_weighted_f1": 0.8115402855105713,
|
| 317 |
+
"step": 905
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.110497237569061,
|
| 321 |
+
"grad_norm": 33.517417907714844,
|
| 322 |
+
"learning_rate": 1.6550030693677105e-05,
|
| 323 |
+
"loss": 0.5304,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.248618784530387,
|
| 328 |
+
"grad_norm": 6.461974143981934,
|
| 329 |
+
"learning_rate": 1.639656230816452e-05,
|
| 330 |
+
"loss": 0.6021,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.386740331491713,
|
| 335 |
+
"grad_norm": 76.11135864257812,
|
| 336 |
+
"learning_rate": 1.6243093922651937e-05,
|
| 337 |
+
"loss": 0.6793,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.524861878453039,
|
| 342 |
+
"grad_norm": 52.50167465209961,
|
| 343 |
+
"learning_rate": 1.608962553713935e-05,
|
| 344 |
+
"loss": 0.5689,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.662983425414365,
|
| 349 |
+
"grad_norm": 4.192610740661621,
|
| 350 |
+
"learning_rate": 1.5936157151626766e-05,
|
| 351 |
+
"loss": 0.6667,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.801104972375691,
|
| 356 |
+
"grad_norm": 0.6879556775093079,
|
| 357 |
+
"learning_rate": 1.5782688766114183e-05,
|
| 358 |
+
"loss": 0.5259,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.939226519337017,
|
| 363 |
+
"grad_norm": 167.03099060058594,
|
| 364 |
+
"learning_rate": 1.5629220380601598e-05,
|
| 365 |
+
"loss": 0.5718,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 6.0,
|
| 370 |
+
"eval_accuracy": 0.8419243986254296,
|
| 371 |
+
"eval_loss": 0.6402910947799683,
|
| 372 |
+
"eval_macro_f1": 0.8406598784091951,
|
| 373 |
+
"eval_runtime": 0.4241,
|
| 374 |
+
"eval_samples_per_second": 1372.219,
|
| 375 |
+
"eval_steps_per_second": 87.237,
|
| 376 |
+
"eval_weighted_f1": 0.8411055752570561,
|
| 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-s44-20260826T160702Z-ab4112e3/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-s44-20260826T160702Z-ab4112e3/trainer/checkpoint-1629/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-source-source-02-s44-20260826T160702Z-ab4112e3/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-s44-20260826T160702Z-ab4112e3/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-s44-20260826T160702Z-ab4112e3/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-s44-20260826T160702Z-ab4112e3/trainer/checkpoint-1629/trainer_state.json
ADDED
|
@@ -0,0 +1,597 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1086,
|
| 3 |
+
"best_metric": 0.8406598784091951,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-source-source-02-s44-20260826T160702Z-ab4112e3/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": 6.044527530670166,
|
| 15 |
+
"learning_rate": 1.3259668508287293e-06,
|
| 16 |
+
"loss": 4.4406,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.27624309392265195,
|
| 21 |
+
"grad_norm": 7.572896957397461,
|
| 22 |
+
"learning_rate": 2.707182320441989e-06,
|
| 23 |
+
"loss": 4.3821,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.4143646408839779,
|
| 28 |
+
"grad_norm": 7.113949775695801,
|
| 29 |
+
"learning_rate": 4.088397790055249e-06,
|
| 30 |
+
"loss": 4.3729,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5524861878453039,
|
| 35 |
+
"grad_norm": 6.951432228088379,
|
| 36 |
+
"learning_rate": 5.469613259668509e-06,
|
| 37 |
+
"loss": 4.2956,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6906077348066298,
|
| 42 |
+
"grad_norm": 8.143998146057129,
|
| 43 |
+
"learning_rate": 6.850828729281769e-06,
|
| 44 |
+
"loss": 4.2841,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8287292817679558,
|
| 49 |
+
"grad_norm": 18.067462921142578,
|
| 50 |
+
"learning_rate": 8.232044198895029e-06,
|
| 51 |
+
"loss": 4.0918,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9668508287292817,
|
| 56 |
+
"grad_norm": 22.35405731201172,
|
| 57 |
+
"learning_rate": 9.613259668508288e-06,
|
| 58 |
+
"loss": 3.9472,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.5807560137457045,
|
| 64 |
+
"eval_loss": 0.8864754438400269,
|
| 65 |
+
"eval_macro_f1": 0.5673594939188857,
|
| 66 |
+
"eval_runtime": 0.404,
|
| 67 |
+
"eval_samples_per_second": 1440.51,
|
| 68 |
+
"eval_steps_per_second": 91.579,
|
| 69 |
+
"eval_weighted_f1": 0.572302371883713,
|
| 70 |
+
"step": 181
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.1049723756906078,
|
| 74 |
+
"grad_norm": 25.207218170166016,
|
| 75 |
+
"learning_rate": 1.0994475138121547e-05,
|
| 76 |
+
"loss": 3.653,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.2430939226519337,
|
| 81 |
+
"grad_norm": 16.948280334472656,
|
| 82 |
+
"learning_rate": 1.2375690607734806e-05,
|
| 83 |
+
"loss": 3.4973,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.3812154696132597,
|
| 88 |
+
"grad_norm": 49.38298797607422,
|
| 89 |
+
"learning_rate": 1.3756906077348067e-05,
|
| 90 |
+
"loss": 3.2864,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.5193370165745856,
|
| 95 |
+
"grad_norm": 43.83489990234375,
|
| 96 |
+
"learning_rate": 1.5138121546961326e-05,
|
| 97 |
+
"loss": 2.9535,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.6574585635359116,
|
| 102 |
+
"grad_norm": 27.498220443725586,
|
| 103 |
+
"learning_rate": 1.6519337016574586e-05,
|
| 104 |
+
"loss": 2.496,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7955801104972375,
|
| 109 |
+
"grad_norm": 45.16746139526367,
|
| 110 |
+
"learning_rate": 1.7900552486187846e-05,
|
| 111 |
+
"loss": 2.4388,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.9337016574585635,
|
| 116 |
+
"grad_norm": 30.030315399169922,
|
| 117 |
+
"learning_rate": 1.9281767955801107e-05,
|
| 118 |
+
"loss": 2.3232,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.7869415807560137,
|
| 124 |
+
"eval_loss": 0.5499523282051086,
|
| 125 |
+
"eval_macro_f1": 0.7847817494176726,
|
| 126 |
+
"eval_runtime": 0.383,
|
| 127 |
+
"eval_samples_per_second": 1519.528,
|
| 128 |
+
"eval_steps_per_second": 96.602,
|
| 129 |
+
"eval_weighted_f1": 0.7861326943126823,
|
| 130 |
+
"step": 362
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.0718232044198897,
|
| 134 |
+
"grad_norm": 18.682743072509766,
|
| 135 |
+
"learning_rate": 1.9926335174953962e-05,
|
| 136 |
+
"loss": 1.9575,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.2099447513812156,
|
| 141 |
+
"grad_norm": 33.75157165527344,
|
| 142 |
+
"learning_rate": 1.9772866789441377e-05,
|
| 143 |
+
"loss": 1.7536,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.3480662983425415,
|
| 148 |
+
"grad_norm": 50.764705657958984,
|
| 149 |
+
"learning_rate": 1.9619398403928794e-05,
|
| 150 |
+
"loss": 1.6032,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.4861878453038675,
|
| 155 |
+
"grad_norm": 61.391056060791016,
|
| 156 |
+
"learning_rate": 1.946593001841621e-05,
|
| 157 |
+
"loss": 1.8642,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.6243093922651934,
|
| 162 |
+
"grad_norm": 31.565210342407227,
|
| 163 |
+
"learning_rate": 1.9312461632903623e-05,
|
| 164 |
+
"loss": 1.7283,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.7624309392265194,
|
| 169 |
+
"grad_norm": 34.41781234741211,
|
| 170 |
+
"learning_rate": 1.915899324739104e-05,
|
| 171 |
+
"loss": 1.6703,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.9005524861878453,
|
| 176 |
+
"grad_norm": 36.9642219543457,
|
| 177 |
+
"learning_rate": 1.9005524861878455e-05,
|
| 178 |
+
"loss": 1.5379,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 3.0,
|
| 183 |
+
"eval_accuracy": 0.8006872852233677,
|
| 184 |
+
"eval_loss": 0.5332517623901367,
|
| 185 |
+
"eval_macro_f1": 0.7970787408095045,
|
| 186 |
+
"eval_runtime": 0.4492,
|
| 187 |
+
"eval_samples_per_second": 1295.696,
|
| 188 |
+
"eval_steps_per_second": 82.372,
|
| 189 |
+
"eval_weighted_f1": 0.798666263369714,
|
| 190 |
+
"step": 543
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.0386740331491713,
|
| 194 |
+
"grad_norm": 21.794315338134766,
|
| 195 |
+
"learning_rate": 1.885205647636587e-05,
|
| 196 |
+
"loss": 1.5075,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.1767955801104972,
|
| 201 |
+
"grad_norm": 32.58504104614258,
|
| 202 |
+
"learning_rate": 1.8698588090853287e-05,
|
| 203 |
+
"loss": 1.0903,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.314917127071823,
|
| 208 |
+
"grad_norm": 31.611621856689453,
|
| 209 |
+
"learning_rate": 1.85451197053407e-05,
|
| 210 |
+
"loss": 0.9415,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.453038674033149,
|
| 215 |
+
"grad_norm": 29.14419937133789,
|
| 216 |
+
"learning_rate": 1.8391651319828116e-05,
|
| 217 |
+
"loss": 1.1406,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.591160220994475,
|
| 222 |
+
"grad_norm": 45.6563606262207,
|
| 223 |
+
"learning_rate": 1.823818293431553e-05,
|
| 224 |
+
"loss": 1.1686,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.729281767955801,
|
| 229 |
+
"grad_norm": 4.467889785766602,
|
| 230 |
+
"learning_rate": 1.8084714548802948e-05,
|
| 231 |
+
"loss": 0.9831,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.867403314917127,
|
| 236 |
+
"grad_norm": 89.3020248413086,
|
| 237 |
+
"learning_rate": 1.7931246163290362e-05,
|
| 238 |
+
"loss": 1.2411,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 4.0,
|
| 243 |
+
"eval_accuracy": 0.8247422680412371,
|
| 244 |
+
"eval_loss": 0.517365038394928,
|
| 245 |
+
"eval_macro_f1": 0.8222884518956111,
|
| 246 |
+
"eval_runtime": 0.4849,
|
| 247 |
+
"eval_samples_per_second": 1200.262,
|
| 248 |
+
"eval_steps_per_second": 76.305,
|
| 249 |
+
"eval_weighted_f1": 0.823348778311946,
|
| 250 |
+
"step": 724
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 4.005524861878453,
|
| 254 |
+
"grad_norm": 67.46331787109375,
|
| 255 |
+
"learning_rate": 1.7777777777777777e-05,
|
| 256 |
+
"loss": 1.0388,
|
| 257 |
+
"step": 725
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.143646408839779,
|
| 261 |
+
"grad_norm": 20.47144317626953,
|
| 262 |
+
"learning_rate": 1.7624309392265194e-05,
|
| 263 |
+
"loss": 0.634,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.281767955801105,
|
| 268 |
+
"grad_norm": 108.18099212646484,
|
| 269 |
+
"learning_rate": 1.747084100675261e-05,
|
| 270 |
+
"loss": 0.7141,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.419889502762431,
|
| 275 |
+
"grad_norm": 113.71337890625,
|
| 276 |
+
"learning_rate": 1.7317372621240026e-05,
|
| 277 |
+
"loss": 0.8237,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.558011049723757,
|
| 282 |
+
"grad_norm": 49.596534729003906,
|
| 283 |
+
"learning_rate": 1.716390423572744e-05,
|
| 284 |
+
"loss": 0.6921,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.696132596685083,
|
| 289 |
+
"grad_norm": 42.291099548339844,
|
| 290 |
+
"learning_rate": 1.701043585021486e-05,
|
| 291 |
+
"loss": 0.8426,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.834254143646409,
|
| 296 |
+
"grad_norm": 53.41448974609375,
|
| 297 |
+
"learning_rate": 1.6856967464702273e-05,
|
| 298 |
+
"loss": 0.7432,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.972375690607735,
|
| 303 |
+
"grad_norm": 20.884723663330078,
|
| 304 |
+
"learning_rate": 1.670349907918969e-05,
|
| 305 |
+
"loss": 0.6546,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 5.0,
|
| 310 |
+
"eval_accuracy": 0.8127147766323024,
|
| 311 |
+
"eval_loss": 0.6466218829154968,
|
| 312 |
+
"eval_macro_f1": 0.8103961419622757,
|
| 313 |
+
"eval_runtime": 0.4249,
|
| 314 |
+
"eval_samples_per_second": 1369.574,
|
| 315 |
+
"eval_steps_per_second": 87.069,
|
| 316 |
+
"eval_weighted_f1": 0.8115402855105713,
|
| 317 |
+
"step": 905
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.110497237569061,
|
| 321 |
+
"grad_norm": 33.517417907714844,
|
| 322 |
+
"learning_rate": 1.6550030693677105e-05,
|
| 323 |
+
"loss": 0.5304,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.248618784530387,
|
| 328 |
+
"grad_norm": 6.461974143981934,
|
| 329 |
+
"learning_rate": 1.639656230816452e-05,
|
| 330 |
+
"loss": 0.6021,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.386740331491713,
|
| 335 |
+
"grad_norm": 76.11135864257812,
|
| 336 |
+
"learning_rate": 1.6243093922651937e-05,
|
| 337 |
+
"loss": 0.6793,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.524861878453039,
|
| 342 |
+
"grad_norm": 52.50167465209961,
|
| 343 |
+
"learning_rate": 1.608962553713935e-05,
|
| 344 |
+
"loss": 0.5689,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.662983425414365,
|
| 349 |
+
"grad_norm": 4.192610740661621,
|
| 350 |
+
"learning_rate": 1.5936157151626766e-05,
|
| 351 |
+
"loss": 0.6667,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.801104972375691,
|
| 356 |
+
"grad_norm": 0.6879556775093079,
|
| 357 |
+
"learning_rate": 1.5782688766114183e-05,
|
| 358 |
+
"loss": 0.5259,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.939226519337017,
|
| 363 |
+
"grad_norm": 167.03099060058594,
|
| 364 |
+
"learning_rate": 1.5629220380601598e-05,
|
| 365 |
+
"loss": 0.5718,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 6.0,
|
| 370 |
+
"eval_accuracy": 0.8419243986254296,
|
| 371 |
+
"eval_loss": 0.6402910947799683,
|
| 372 |
+
"eval_macro_f1": 0.8406598784091951,
|
| 373 |
+
"eval_runtime": 0.4241,
|
| 374 |
+
"eval_samples_per_second": 1372.219,
|
| 375 |
+
"eval_steps_per_second": 87.237,
|
| 376 |
+
"eval_weighted_f1": 0.8411055752570561,
|
| 377 |
+
"step": 1086
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 6.077348066298343,
|
| 381 |
+
"grad_norm": 0.6399435997009277,
|
| 382 |
+
"learning_rate": 1.5475751995089012e-05,
|
| 383 |
+
"loss": 0.3542,
|
| 384 |
+
"step": 1100
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 6.2154696132596685,
|
| 388 |
+
"grad_norm": 0.48149237036705017,
|
| 389 |
+
"learning_rate": 1.532228360957643e-05,
|
| 390 |
+
"loss": 0.3222,
|
| 391 |
+
"step": 1125
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 6.3535911602209945,
|
| 395 |
+
"grad_norm": 53.789710998535156,
|
| 396 |
+
"learning_rate": 1.5168815224063844e-05,
|
| 397 |
+
"loss": 0.4171,
|
| 398 |
+
"step": 1150
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.49171270718232,
|
| 402 |
+
"grad_norm": 0.9558493494987488,
|
| 403 |
+
"learning_rate": 1.5015346838551258e-05,
|
| 404 |
+
"loss": 0.372,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.629834254143646,
|
| 409 |
+
"grad_norm": 21.95255470275879,
|
| 410 |
+
"learning_rate": 1.4861878453038676e-05,
|
| 411 |
+
"loss": 0.4194,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.767955801104972,
|
| 416 |
+
"grad_norm": 78.58480072021484,
|
| 417 |
+
"learning_rate": 1.470841006752609e-05,
|
| 418 |
+
"loss": 0.5042,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.906077348066298,
|
| 423 |
+
"grad_norm": 17.510019302368164,
|
| 424 |
+
"learning_rate": 1.4554941682013506e-05,
|
| 425 |
+
"loss": 0.3986,
|
| 426 |
+
"step": 1250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 7.0,
|
| 430 |
+
"eval_accuracy": 0.8316151202749141,
|
| 431 |
+
"eval_loss": 0.8065578937530518,
|
| 432 |
+
"eval_macro_f1": 0.8300560490102974,
|
| 433 |
+
"eval_runtime": 0.4326,
|
| 434 |
+
"eval_samples_per_second": 1345.443,
|
| 435 |
+
"eval_steps_per_second": 85.535,
|
| 436 |
+
"eval_weighted_f1": 0.8308818010320604,
|
| 437 |
+
"step": 1267
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 7.044198895027624,
|
| 441 |
+
"grad_norm": 1.7133725881576538,
|
| 442 |
+
"learning_rate": 1.4401473296500922e-05,
|
| 443 |
+
"loss": 0.3648,
|
| 444 |
+
"step": 1275
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 7.18232044198895,
|
| 448 |
+
"grad_norm": 225.0423583984375,
|
| 449 |
+
"learning_rate": 1.4248004910988338e-05,
|
| 450 |
+
"loss": 0.3277,
|
| 451 |
+
"step": 1300
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 7.320441988950276,
|
| 455 |
+
"grad_norm": 0.575197160243988,
|
| 456 |
+
"learning_rate": 1.4094536525475753e-05,
|
| 457 |
+
"loss": 0.3068,
|
| 458 |
+
"step": 1325
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 7.458563535911602,
|
| 462 |
+
"grad_norm": 2.1740214824676514,
|
| 463 |
+
"learning_rate": 1.394106813996317e-05,
|
| 464 |
+
"loss": 0.4025,
|
| 465 |
+
"step": 1350
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 7.596685082872928,
|
| 469 |
+
"grad_norm": 0.3489013612270355,
|
| 470 |
+
"learning_rate": 1.3787599754450585e-05,
|
| 471 |
+
"loss": 0.2831,
|
| 472 |
+
"step": 1375
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 7.734806629834254,
|
| 476 |
+
"grad_norm": 192.26475524902344,
|
| 477 |
+
"learning_rate": 1.3634131368937999e-05,
|
| 478 |
+
"loss": 0.4622,
|
| 479 |
+
"step": 1400
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 7.87292817679558,
|
| 483 |
+
"grad_norm": 89.55827331542969,
|
| 484 |
+
"learning_rate": 1.3480662983425417e-05,
|
| 485 |
+
"loss": 0.1985,
|
| 486 |
+
"step": 1425
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 8.0,
|
| 490 |
+
"eval_accuracy": 0.8264604810996563,
|
| 491 |
+
"eval_loss": 0.892669141292572,
|
| 492 |
+
"eval_macro_f1": 0.8223510359402196,
|
| 493 |
+
"eval_runtime": 0.4953,
|
| 494 |
+
"eval_samples_per_second": 1175.1,
|
| 495 |
+
"eval_steps_per_second": 74.706,
|
| 496 |
+
"eval_weighted_f1": 0.8234622252208713,
|
| 497 |
+
"step": 1448
|
| 498 |
+
},
|
| 499 |
+
{
|
| 500 |
+
"epoch": 8.011049723756907,
|
| 501 |
+
"grad_norm": 0.7826727032661438,
|
| 502 |
+
"learning_rate": 1.3327194597912831e-05,
|
| 503 |
+
"loss": 0.4191,
|
| 504 |
+
"step": 1450
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 8.149171270718233,
|
| 508 |
+
"grad_norm": 0.24642150104045868,
|
| 509 |
+
"learning_rate": 1.3173726212400245e-05,
|
| 510 |
+
"loss": 0.2722,
|
| 511 |
+
"step": 1475
|
| 512 |
+
},
|
| 513 |
+
{
|
| 514 |
+
"epoch": 8.287292817679559,
|
| 515 |
+
"grad_norm": 0.24304021894931793,
|
| 516 |
+
"learning_rate": 1.3020257826887662e-05,
|
| 517 |
+
"loss": 0.2508,
|
| 518 |
+
"step": 1500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 8.425414364640885,
|
| 522 |
+
"grad_norm": 0.21780814230442047,
|
| 523 |
+
"learning_rate": 1.2866789441375078e-05,
|
| 524 |
+
"loss": 0.1802,
|
| 525 |
+
"step": 1525
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 8.56353591160221,
|
| 529 |
+
"grad_norm": 0.28572791814804077,
|
| 530 |
+
"learning_rate": 1.2713321055862494e-05,
|
| 531 |
+
"loss": 0.2262,
|
| 532 |
+
"step": 1550
|
| 533 |
+
},
|
| 534 |
+
{
|
| 535 |
+
"epoch": 8.701657458563536,
|
| 536 |
+
"grad_norm": 93.37821960449219,
|
| 537 |
+
"learning_rate": 1.2559852670349908e-05,
|
| 538 |
+
"loss": 0.2345,
|
| 539 |
+
"step": 1575
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 8.839779005524862,
|
| 543 |
+
"grad_norm": 3.6706416606903076,
|
| 544 |
+
"learning_rate": 1.2406384284837326e-05,
|
| 545 |
+
"loss": 0.1305,
|
| 546 |
+
"step": 1600
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 8.977900552486188,
|
| 550 |
+
"grad_norm": 151.74456787109375,
|
| 551 |
+
"learning_rate": 1.225291589932474e-05,
|
| 552 |
+
"loss": 0.3204,
|
| 553 |
+
"step": 1625
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 9.0,
|
| 557 |
+
"eval_accuracy": 0.8195876288659794,
|
| 558 |
+
"eval_loss": 0.997809886932373,
|
| 559 |
+
"eval_macro_f1": 0.8169357999149741,
|
| 560 |
+
"eval_runtime": 0.4116,
|
| 561 |
+
"eval_samples_per_second": 1413.927,
|
| 562 |
+
"eval_steps_per_second": 89.889,
|
| 563 |
+
"eval_weighted_f1": 0.8180848350450222,
|
| 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-s44-20260826T160702Z-ab4112e3/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-s44-20260826T161012Z-ab4112e3/trainer/checkpoint-1098/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-source-source-03-s44-20260826T161012Z-ab4112e3/trainer/checkpoint-1098/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-source-source-03-s44-20260826T161012Z-ab4112e3/trainer/checkpoint-1098/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-s44-20260826T161012Z-ab4112e3/trainer/checkpoint-1098/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-s44-20260826T161012Z-ab4112e3/trainer/checkpoint-1098/trainer_state.json
ADDED
|
@@ -0,0 +1,410 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1098,
|
| 3 |
+
"best_metric": 0.8667731384709021,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-source-source-03-s44-20260826T161012Z-ab4112e3/trainer/checkpoint-1098",
|
| 5 |
+
"epoch": 6.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1098,
|
| 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": 9.14084529876709,
|
| 15 |
+
"learning_rate": 1.3114754098360657e-06,
|
| 16 |
+
"loss": 4.4565,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.27434842249657065,
|
| 21 |
+
"grad_norm": 6.638696670532227,
|
| 22 |
+
"learning_rate": 2.677595628415301e-06,
|
| 23 |
+
"loss": 4.3952,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.411522633744856,
|
| 28 |
+
"grad_norm": 6.47113037109375,
|
| 29 |
+
"learning_rate": 4.043715846994536e-06,
|
| 30 |
+
"loss": 4.3677,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5486968449931413,
|
| 35 |
+
"grad_norm": 7.433169364929199,
|
| 36 |
+
"learning_rate": 5.409836065573772e-06,
|
| 37 |
+
"loss": 4.3446,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6858710562414266,
|
| 42 |
+
"grad_norm": 8.611000061035156,
|
| 43 |
+
"learning_rate": 6.775956284153006e-06,
|
| 44 |
+
"loss": 4.2631,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.823045267489712,
|
| 49 |
+
"grad_norm": 14.969688415527344,
|
| 50 |
+
"learning_rate": 8.14207650273224e-06,
|
| 51 |
+
"loss": 4.2146,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9602194787379973,
|
| 56 |
+
"grad_norm": 15.003164291381836,
|
| 57 |
+
"learning_rate": 9.508196721311476e-06,
|
| 58 |
+
"loss": 3.9667,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.5723793677204659,
|
| 64 |
+
"eval_loss": 0.9121304750442505,
|
| 65 |
+
"eval_macro_f1": 0.5530641979463095,
|
| 66 |
+
"eval_runtime": 0.4551,
|
| 67 |
+
"eval_samples_per_second": 1320.641,
|
| 68 |
+
"eval_steps_per_second": 83.501,
|
| 69 |
+
"eval_weighted_f1": 0.5571217836229532,
|
| 70 |
+
"step": 183
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.093278463648834,
|
| 74 |
+
"grad_norm": 27.919403076171875,
|
| 75 |
+
"learning_rate": 1.0874316939890712e-05,
|
| 76 |
+
"loss": 3.5152,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.2304526748971194,
|
| 81 |
+
"grad_norm": 40.56401443481445,
|
| 82 |
+
"learning_rate": 1.2240437158469946e-05,
|
| 83 |
+
"loss": 3.2635,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.3676268861454046,
|
| 88 |
+
"grad_norm": 36.111392974853516,
|
| 89 |
+
"learning_rate": 1.3606557377049181e-05,
|
| 90 |
+
"loss": 3.0643,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.50480109739369,
|
| 95 |
+
"grad_norm": 37.799720764160156,
|
| 96 |
+
"learning_rate": 1.4972677595628417e-05,
|
| 97 |
+
"loss": 2.7515,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.6419753086419753,
|
| 102 |
+
"grad_norm": 27.418241500854492,
|
| 103 |
+
"learning_rate": 1.633879781420765e-05,
|
| 104 |
+
"loss": 2.5363,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7791495198902605,
|
| 109 |
+
"grad_norm": 38.169376373291016,
|
| 110 |
+
"learning_rate": 1.7704918032786887e-05,
|
| 111 |
+
"loss": 2.4116,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.916323731138546,
|
| 116 |
+
"grad_norm": 50.547508239746094,
|
| 117 |
+
"learning_rate": 1.9071038251366122e-05,
|
| 118 |
+
"loss": 2.364,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.8202995008319468,
|
| 124 |
+
"eval_loss": 0.5121298432350159,
|
| 125 |
+
"eval_macro_f1": 0.819253145951842,
|
| 126 |
+
"eval_runtime": 0.4602,
|
| 127 |
+
"eval_samples_per_second": 1306.067,
|
| 128 |
+
"eval_steps_per_second": 82.58,
|
| 129 |
+
"eval_weighted_f1": 0.8191082804521627,
|
| 130 |
+
"step": 366
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.049382716049383,
|
| 134 |
+
"grad_norm": 31.986751556396484,
|
| 135 |
+
"learning_rate": 1.995142683667274e-05,
|
| 136 |
+
"loss": 1.9526,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.186556927297668,
|
| 141 |
+
"grad_norm": 56.468162536621094,
|
| 142 |
+
"learning_rate": 1.9799635701275045e-05,
|
| 143 |
+
"loss": 1.8796,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.3237311385459534,
|
| 148 |
+
"grad_norm": 40.80320358276367,
|
| 149 |
+
"learning_rate": 1.9647844565877354e-05,
|
| 150 |
+
"loss": 1.6696,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.460905349794239,
|
| 155 |
+
"grad_norm": 26.488534927368164,
|
| 156 |
+
"learning_rate": 1.9496053430479663e-05,
|
| 157 |
+
"loss": 1.7246,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.598079561042524,
|
| 162 |
+
"grad_norm": 41.13813018798828,
|
| 163 |
+
"learning_rate": 1.934426229508197e-05,
|
| 164 |
+
"loss": 1.7269,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.7352537722908092,
|
| 169 |
+
"grad_norm": 43.186031341552734,
|
| 170 |
+
"learning_rate": 1.9192471159684274e-05,
|
| 171 |
+
"loss": 1.7226,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.8724279835390947,
|
| 176 |
+
"grad_norm": 45.34514617919922,
|
| 177 |
+
"learning_rate": 1.9040680024286583e-05,
|
| 178 |
+
"loss": 1.4887,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 3.0,
|
| 183 |
+
"eval_accuracy": 0.8186356073211315,
|
| 184 |
+
"eval_loss": 0.48486408591270447,
|
| 185 |
+
"eval_macro_f1": 0.8209874858482937,
|
| 186 |
+
"eval_runtime": 0.4278,
|
| 187 |
+
"eval_samples_per_second": 1404.899,
|
| 188 |
+
"eval_steps_per_second": 88.829,
|
| 189 |
+
"eval_weighted_f1": 0.8211503490462685,
|
| 190 |
+
"step": 549
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.0054869684499312,
|
| 194 |
+
"grad_norm": 32.80615997314453,
|
| 195 |
+
"learning_rate": 1.888888888888889e-05,
|
| 196 |
+
"loss": 1.4462,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.1426611796982167,
|
| 201 |
+
"grad_norm": 46.36381149291992,
|
| 202 |
+
"learning_rate": 1.8737097753491198e-05,
|
| 203 |
+
"loss": 1.3822,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.279835390946502,
|
| 208 |
+
"grad_norm": 37.848907470703125,
|
| 209 |
+
"learning_rate": 1.8585306618093504e-05,
|
| 210 |
+
"loss": 1.1065,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.4170096021947876,
|
| 215 |
+
"grad_norm": 52.04416275024414,
|
| 216 |
+
"learning_rate": 1.8433515482695813e-05,
|
| 217 |
+
"loss": 1.0358,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.5541838134430725,
|
| 222 |
+
"grad_norm": 32.907127380371094,
|
| 223 |
+
"learning_rate": 1.828172434729812e-05,
|
| 224 |
+
"loss": 1.0118,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.691358024691358,
|
| 229 |
+
"grad_norm": 59.6409797668457,
|
| 230 |
+
"learning_rate": 1.8129933211900424e-05,
|
| 231 |
+
"loss": 1.2099,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.8285322359396434,
|
| 236 |
+
"grad_norm": 24.82037925720215,
|
| 237 |
+
"learning_rate": 1.7978142076502733e-05,
|
| 238 |
+
"loss": 1.2943,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.9657064471879284,
|
| 243 |
+
"grad_norm": 32.69512939453125,
|
| 244 |
+
"learning_rate": 1.7826350941105042e-05,
|
| 245 |
+
"loss": 1.1378,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 4.0,
|
| 250 |
+
"eval_accuracy": 0.8386023294509152,
|
| 251 |
+
"eval_loss": 0.49895599484443665,
|
| 252 |
+
"eval_macro_f1": 0.8379701041238491,
|
| 253 |
+
"eval_runtime": 0.4027,
|
| 254 |
+
"eval_samples_per_second": 1492.432,
|
| 255 |
+
"eval_steps_per_second": 94.363,
|
| 256 |
+
"eval_weighted_f1": 0.8378791458502926,
|
| 257 |
+
"step": 732
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.098765432098766,
|
| 261 |
+
"grad_norm": 34.856910705566406,
|
| 262 |
+
"learning_rate": 1.7674559805707348e-05,
|
| 263 |
+
"loss": 0.6818,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.23593964334705,
|
| 268 |
+
"grad_norm": 66.67054748535156,
|
| 269 |
+
"learning_rate": 1.7522768670309657e-05,
|
| 270 |
+
"loss": 0.7747,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.373113854595336,
|
| 275 |
+
"grad_norm": 44.203922271728516,
|
| 276 |
+
"learning_rate": 1.7370977534911962e-05,
|
| 277 |
+
"loss": 0.9211,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.510288065843621,
|
| 282 |
+
"grad_norm": 23.535808563232422,
|
| 283 |
+
"learning_rate": 1.7219186399514268e-05,
|
| 284 |
+
"loss": 0.8512,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.647462277091907,
|
| 289 |
+
"grad_norm": 29.59877586364746,
|
| 290 |
+
"learning_rate": 1.7067395264116577e-05,
|
| 291 |
+
"loss": 0.7459,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.784636488340192,
|
| 296 |
+
"grad_norm": 63.10303497314453,
|
| 297 |
+
"learning_rate": 1.6915604128718883e-05,
|
| 298 |
+
"loss": 0.6957,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.921810699588478,
|
| 303 |
+
"grad_norm": 38.458412170410156,
|
| 304 |
+
"learning_rate": 1.6763812993321192e-05,
|
| 305 |
+
"loss": 0.8009,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 5.0,
|
| 310 |
+
"eval_accuracy": 0.8602329450915142,
|
| 311 |
+
"eval_loss": 0.47414273023605347,
|
| 312 |
+
"eval_macro_f1": 0.8601913632204276,
|
| 313 |
+
"eval_runtime": 0.4126,
|
| 314 |
+
"eval_samples_per_second": 1456.594,
|
| 315 |
+
"eval_steps_per_second": 92.097,
|
| 316 |
+
"eval_weighted_f1": 0.8601069671941032,
|
| 317 |
+
"step": 915
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.054869684499314,
|
| 321 |
+
"grad_norm": 48.971641540527344,
|
| 322 |
+
"learning_rate": 1.66120218579235e-05,
|
| 323 |
+
"loss": 0.5371,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.1920438957476,
|
| 328 |
+
"grad_norm": 7.917779922485352,
|
| 329 |
+
"learning_rate": 1.6460230722525807e-05,
|
| 330 |
+
"loss": 0.5058,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.329218106995885,
|
| 335 |
+
"grad_norm": 28.134660720825195,
|
| 336 |
+
"learning_rate": 1.6308439587128112e-05,
|
| 337 |
+
"loss": 0.6749,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.4663923182441705,
|
| 342 |
+
"grad_norm": 39.6581916809082,
|
| 343 |
+
"learning_rate": 1.615664845173042e-05,
|
| 344 |
+
"loss": 0.7579,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.603566529492455,
|
| 349 |
+
"grad_norm": 40.38935089111328,
|
| 350 |
+
"learning_rate": 1.6004857316332727e-05,
|
| 351 |
+
"loss": 0.485,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.7407407407407405,
|
| 356 |
+
"grad_norm": 4.382994174957275,
|
| 357 |
+
"learning_rate": 1.5853066180935036e-05,
|
| 358 |
+
"loss": 0.557,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.877914951989026,
|
| 363 |
+
"grad_norm": 11.995711326599121,
|
| 364 |
+
"learning_rate": 1.570127504553734e-05,
|
| 365 |
+
"loss": 0.5849,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 6.0,
|
| 370 |
+
"eval_accuracy": 0.8668885191347754,
|
| 371 |
+
"eval_loss": 0.5215726494789124,
|
| 372 |
+
"eval_macro_f1": 0.8667731384709021,
|
| 373 |
+
"eval_runtime": 0.4191,
|
| 374 |
+
"eval_samples_per_second": 1434.114,
|
| 375 |
+
"eval_steps_per_second": 90.676,
|
| 376 |
+
"eval_weighted_f1": 0.8666583295599006,
|
| 377 |
+
"step": 1098
|
| 378 |
+
}
|
| 379 |
+
],
|
| 380 |
+
"logging_steps": 25,
|
| 381 |
+
"max_steps": 3660,
|
| 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-03-s44-20260826T161012Z-ab4112e3/trainer/checkpoint-1098/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-source-source-03-s44-20260826T161012Z-ab4112e3/trainer/checkpoint-1647/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-source-source-03-s44-20260826T161012Z-ab4112e3/trainer/checkpoint-1647/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-source-source-03-s44-20260826T161012Z-ab4112e3/trainer/checkpoint-1647/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-s44-20260826T161012Z-ab4112e3/trainer/checkpoint-1647/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-s44-20260826T161012Z-ab4112e3/trainer/checkpoint-1647/trainer_state.json
ADDED
|
@@ -0,0 +1,597 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1098,
|
| 3 |
+
"best_metric": 0.8667731384709021,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-source-source-03-s44-20260826T161012Z-ab4112e3/trainer/checkpoint-1098",
|
| 5 |
+
"epoch": 9.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1647,
|
| 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": 9.14084529876709,
|
| 15 |
+
"learning_rate": 1.3114754098360657e-06,
|
| 16 |
+
"loss": 4.4565,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.27434842249657065,
|
| 21 |
+
"grad_norm": 6.638696670532227,
|
| 22 |
+
"learning_rate": 2.677595628415301e-06,
|
| 23 |
+
"loss": 4.3952,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.411522633744856,
|
| 28 |
+
"grad_norm": 6.47113037109375,
|
| 29 |
+
"learning_rate": 4.043715846994536e-06,
|
| 30 |
+
"loss": 4.3677,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5486968449931413,
|
| 35 |
+
"grad_norm": 7.433169364929199,
|
| 36 |
+
"learning_rate": 5.409836065573772e-06,
|
| 37 |
+
"loss": 4.3446,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6858710562414266,
|
| 42 |
+
"grad_norm": 8.611000061035156,
|
| 43 |
+
"learning_rate": 6.775956284153006e-06,
|
| 44 |
+
"loss": 4.2631,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.823045267489712,
|
| 49 |
+
"grad_norm": 14.969688415527344,
|
| 50 |
+
"learning_rate": 8.14207650273224e-06,
|
| 51 |
+
"loss": 4.2146,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9602194787379973,
|
| 56 |
+
"grad_norm": 15.003164291381836,
|
| 57 |
+
"learning_rate": 9.508196721311476e-06,
|
| 58 |
+
"loss": 3.9667,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.5723793677204659,
|
| 64 |
+
"eval_loss": 0.9121304750442505,
|
| 65 |
+
"eval_macro_f1": 0.5530641979463095,
|
| 66 |
+
"eval_runtime": 0.4551,
|
| 67 |
+
"eval_samples_per_second": 1320.641,
|
| 68 |
+
"eval_steps_per_second": 83.501,
|
| 69 |
+
"eval_weighted_f1": 0.5571217836229532,
|
| 70 |
+
"step": 183
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.093278463648834,
|
| 74 |
+
"grad_norm": 27.919403076171875,
|
| 75 |
+
"learning_rate": 1.0874316939890712e-05,
|
| 76 |
+
"loss": 3.5152,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.2304526748971194,
|
| 81 |
+
"grad_norm": 40.56401443481445,
|
| 82 |
+
"learning_rate": 1.2240437158469946e-05,
|
| 83 |
+
"loss": 3.2635,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.3676268861454046,
|
| 88 |
+
"grad_norm": 36.111392974853516,
|
| 89 |
+
"learning_rate": 1.3606557377049181e-05,
|
| 90 |
+
"loss": 3.0643,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.50480109739369,
|
| 95 |
+
"grad_norm": 37.799720764160156,
|
| 96 |
+
"learning_rate": 1.4972677595628417e-05,
|
| 97 |
+
"loss": 2.7515,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.6419753086419753,
|
| 102 |
+
"grad_norm": 27.418241500854492,
|
| 103 |
+
"learning_rate": 1.633879781420765e-05,
|
| 104 |
+
"loss": 2.5363,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7791495198902605,
|
| 109 |
+
"grad_norm": 38.169376373291016,
|
| 110 |
+
"learning_rate": 1.7704918032786887e-05,
|
| 111 |
+
"loss": 2.4116,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.916323731138546,
|
| 116 |
+
"grad_norm": 50.547508239746094,
|
| 117 |
+
"learning_rate": 1.9071038251366122e-05,
|
| 118 |
+
"loss": 2.364,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.8202995008319468,
|
| 124 |
+
"eval_loss": 0.5121298432350159,
|
| 125 |
+
"eval_macro_f1": 0.819253145951842,
|
| 126 |
+
"eval_runtime": 0.4602,
|
| 127 |
+
"eval_samples_per_second": 1306.067,
|
| 128 |
+
"eval_steps_per_second": 82.58,
|
| 129 |
+
"eval_weighted_f1": 0.8191082804521627,
|
| 130 |
+
"step": 366
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.049382716049383,
|
| 134 |
+
"grad_norm": 31.986751556396484,
|
| 135 |
+
"learning_rate": 1.995142683667274e-05,
|
| 136 |
+
"loss": 1.9526,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.186556927297668,
|
| 141 |
+
"grad_norm": 56.468162536621094,
|
| 142 |
+
"learning_rate": 1.9799635701275045e-05,
|
| 143 |
+
"loss": 1.8796,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.3237311385459534,
|
| 148 |
+
"grad_norm": 40.80320358276367,
|
| 149 |
+
"learning_rate": 1.9647844565877354e-05,
|
| 150 |
+
"loss": 1.6696,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.460905349794239,
|
| 155 |
+
"grad_norm": 26.488534927368164,
|
| 156 |
+
"learning_rate": 1.9496053430479663e-05,
|
| 157 |
+
"loss": 1.7246,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.598079561042524,
|
| 162 |
+
"grad_norm": 41.13813018798828,
|
| 163 |
+
"learning_rate": 1.934426229508197e-05,
|
| 164 |
+
"loss": 1.7269,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.7352537722908092,
|
| 169 |
+
"grad_norm": 43.186031341552734,
|
| 170 |
+
"learning_rate": 1.9192471159684274e-05,
|
| 171 |
+
"loss": 1.7226,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.8724279835390947,
|
| 176 |
+
"grad_norm": 45.34514617919922,
|
| 177 |
+
"learning_rate": 1.9040680024286583e-05,
|
| 178 |
+
"loss": 1.4887,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 3.0,
|
| 183 |
+
"eval_accuracy": 0.8186356073211315,
|
| 184 |
+
"eval_loss": 0.48486408591270447,
|
| 185 |
+
"eval_macro_f1": 0.8209874858482937,
|
| 186 |
+
"eval_runtime": 0.4278,
|
| 187 |
+
"eval_samples_per_second": 1404.899,
|
| 188 |
+
"eval_steps_per_second": 88.829,
|
| 189 |
+
"eval_weighted_f1": 0.8211503490462685,
|
| 190 |
+
"step": 549
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.0054869684499312,
|
| 194 |
+
"grad_norm": 32.80615997314453,
|
| 195 |
+
"learning_rate": 1.888888888888889e-05,
|
| 196 |
+
"loss": 1.4462,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.1426611796982167,
|
| 201 |
+
"grad_norm": 46.36381149291992,
|
| 202 |
+
"learning_rate": 1.8737097753491198e-05,
|
| 203 |
+
"loss": 1.3822,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.279835390946502,
|
| 208 |
+
"grad_norm": 37.848907470703125,
|
| 209 |
+
"learning_rate": 1.8585306618093504e-05,
|
| 210 |
+
"loss": 1.1065,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.4170096021947876,
|
| 215 |
+
"grad_norm": 52.04416275024414,
|
| 216 |
+
"learning_rate": 1.8433515482695813e-05,
|
| 217 |
+
"loss": 1.0358,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.5541838134430725,
|
| 222 |
+
"grad_norm": 32.907127380371094,
|
| 223 |
+
"learning_rate": 1.828172434729812e-05,
|
| 224 |
+
"loss": 1.0118,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.691358024691358,
|
| 229 |
+
"grad_norm": 59.6409797668457,
|
| 230 |
+
"learning_rate": 1.8129933211900424e-05,
|
| 231 |
+
"loss": 1.2099,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.8285322359396434,
|
| 236 |
+
"grad_norm": 24.82037925720215,
|
| 237 |
+
"learning_rate": 1.7978142076502733e-05,
|
| 238 |
+
"loss": 1.2943,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.9657064471879284,
|
| 243 |
+
"grad_norm": 32.69512939453125,
|
| 244 |
+
"learning_rate": 1.7826350941105042e-05,
|
| 245 |
+
"loss": 1.1378,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 4.0,
|
| 250 |
+
"eval_accuracy": 0.8386023294509152,
|
| 251 |
+
"eval_loss": 0.49895599484443665,
|
| 252 |
+
"eval_macro_f1": 0.8379701041238491,
|
| 253 |
+
"eval_runtime": 0.4027,
|
| 254 |
+
"eval_samples_per_second": 1492.432,
|
| 255 |
+
"eval_steps_per_second": 94.363,
|
| 256 |
+
"eval_weighted_f1": 0.8378791458502926,
|
| 257 |
+
"step": 732
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.098765432098766,
|
| 261 |
+
"grad_norm": 34.856910705566406,
|
| 262 |
+
"learning_rate": 1.7674559805707348e-05,
|
| 263 |
+
"loss": 0.6818,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.23593964334705,
|
| 268 |
+
"grad_norm": 66.67054748535156,
|
| 269 |
+
"learning_rate": 1.7522768670309657e-05,
|
| 270 |
+
"loss": 0.7747,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.373113854595336,
|
| 275 |
+
"grad_norm": 44.203922271728516,
|
| 276 |
+
"learning_rate": 1.7370977534911962e-05,
|
| 277 |
+
"loss": 0.9211,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.510288065843621,
|
| 282 |
+
"grad_norm": 23.535808563232422,
|
| 283 |
+
"learning_rate": 1.7219186399514268e-05,
|
| 284 |
+
"loss": 0.8512,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.647462277091907,
|
| 289 |
+
"grad_norm": 29.59877586364746,
|
| 290 |
+
"learning_rate": 1.7067395264116577e-05,
|
| 291 |
+
"loss": 0.7459,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.784636488340192,
|
| 296 |
+
"grad_norm": 63.10303497314453,
|
| 297 |
+
"learning_rate": 1.6915604128718883e-05,
|
| 298 |
+
"loss": 0.6957,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.921810699588478,
|
| 303 |
+
"grad_norm": 38.458412170410156,
|
| 304 |
+
"learning_rate": 1.6763812993321192e-05,
|
| 305 |
+
"loss": 0.8009,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 5.0,
|
| 310 |
+
"eval_accuracy": 0.8602329450915142,
|
| 311 |
+
"eval_loss": 0.47414273023605347,
|
| 312 |
+
"eval_macro_f1": 0.8601913632204276,
|
| 313 |
+
"eval_runtime": 0.4126,
|
| 314 |
+
"eval_samples_per_second": 1456.594,
|
| 315 |
+
"eval_steps_per_second": 92.097,
|
| 316 |
+
"eval_weighted_f1": 0.8601069671941032,
|
| 317 |
+
"step": 915
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.054869684499314,
|
| 321 |
+
"grad_norm": 48.971641540527344,
|
| 322 |
+
"learning_rate": 1.66120218579235e-05,
|
| 323 |
+
"loss": 0.5371,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.1920438957476,
|
| 328 |
+
"grad_norm": 7.917779922485352,
|
| 329 |
+
"learning_rate": 1.6460230722525807e-05,
|
| 330 |
+
"loss": 0.5058,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.329218106995885,
|
| 335 |
+
"grad_norm": 28.134660720825195,
|
| 336 |
+
"learning_rate": 1.6308439587128112e-05,
|
| 337 |
+
"loss": 0.6749,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.4663923182441705,
|
| 342 |
+
"grad_norm": 39.6581916809082,
|
| 343 |
+
"learning_rate": 1.615664845173042e-05,
|
| 344 |
+
"loss": 0.7579,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.603566529492455,
|
| 349 |
+
"grad_norm": 40.38935089111328,
|
| 350 |
+
"learning_rate": 1.6004857316332727e-05,
|
| 351 |
+
"loss": 0.485,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.7407407407407405,
|
| 356 |
+
"grad_norm": 4.382994174957275,
|
| 357 |
+
"learning_rate": 1.5853066180935036e-05,
|
| 358 |
+
"loss": 0.557,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.877914951989026,
|
| 363 |
+
"grad_norm": 11.995711326599121,
|
| 364 |
+
"learning_rate": 1.570127504553734e-05,
|
| 365 |
+
"loss": 0.5849,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 6.0,
|
| 370 |
+
"eval_accuracy": 0.8668885191347754,
|
| 371 |
+
"eval_loss": 0.5215726494789124,
|
| 372 |
+
"eval_macro_f1": 0.8667731384709021,
|
| 373 |
+
"eval_runtime": 0.4191,
|
| 374 |
+
"eval_samples_per_second": 1434.114,
|
| 375 |
+
"eval_steps_per_second": 90.676,
|
| 376 |
+
"eval_weighted_f1": 0.8666583295599006,
|
| 377 |
+
"step": 1098
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 6.0109739368998625,
|
| 381 |
+
"grad_norm": 94.9651107788086,
|
| 382 |
+
"learning_rate": 1.5549483910139647e-05,
|
| 383 |
+
"loss": 0.4684,
|
| 384 |
+
"step": 1100
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 6.148148148148148,
|
| 388 |
+
"grad_norm": 17.134449005126953,
|
| 389 |
+
"learning_rate": 1.5397692774741956e-05,
|
| 390 |
+
"loss": 0.508,
|
| 391 |
+
"step": 1125
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 6.285322359396433,
|
| 395 |
+
"grad_norm": 139.90713500976562,
|
| 396 |
+
"learning_rate": 1.5245901639344264e-05,
|
| 397 |
+
"loss": 0.3602,
|
| 398 |
+
"step": 1150
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.422496570644719,
|
| 402 |
+
"grad_norm": 0.6925407648086548,
|
| 403 |
+
"learning_rate": 1.5094110503946571e-05,
|
| 404 |
+
"loss": 0.3667,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.559670781893004,
|
| 409 |
+
"grad_norm": 97.23280334472656,
|
| 410 |
+
"learning_rate": 1.4942319368548878e-05,
|
| 411 |
+
"loss": 0.4593,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.69684499314129,
|
| 416 |
+
"grad_norm": 88.399658203125,
|
| 417 |
+
"learning_rate": 1.4790528233151184e-05,
|
| 418 |
+
"loss": 0.377,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.834019204389575,
|
| 423 |
+
"grad_norm": 17.268003463745117,
|
| 424 |
+
"learning_rate": 1.4638737097753493e-05,
|
| 425 |
+
"loss": 0.3481,
|
| 426 |
+
"step": 1250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 6.97119341563786,
|
| 430 |
+
"grad_norm": 11.108675956726074,
|
| 431 |
+
"learning_rate": 1.44869459623558e-05,
|
| 432 |
+
"loss": 0.4681,
|
| 433 |
+
"step": 1275
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 7.0,
|
| 437 |
+
"eval_accuracy": 0.8552412645590682,
|
| 438 |
+
"eval_loss": 0.6556198000907898,
|
| 439 |
+
"eval_macro_f1": 0.8542878250984037,
|
| 440 |
+
"eval_runtime": 0.4168,
|
| 441 |
+
"eval_samples_per_second": 1441.826,
|
| 442 |
+
"eval_steps_per_second": 91.164,
|
| 443 |
+
"eval_weighted_f1": 0.8543066496482548,
|
| 444 |
+
"step": 1281
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 7.104252400548697,
|
| 448 |
+
"grad_norm": 124.15209197998047,
|
| 449 |
+
"learning_rate": 1.4335154826958106e-05,
|
| 450 |
+
"loss": 0.3962,
|
| 451 |
+
"step": 1300
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 7.2414266117969825,
|
| 455 |
+
"grad_norm": 88.28169250488281,
|
| 456 |
+
"learning_rate": 1.4183363691560415e-05,
|
| 457 |
+
"loss": 0.2921,
|
| 458 |
+
"step": 1325
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 7.378600823045267,
|
| 462 |
+
"grad_norm": 0.39399653673171997,
|
| 463 |
+
"learning_rate": 1.403157255616272e-05,
|
| 464 |
+
"loss": 0.3619,
|
| 465 |
+
"step": 1350
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 7.5157750342935525,
|
| 469 |
+
"grad_norm": 29.161230087280273,
|
| 470 |
+
"learning_rate": 1.3879781420765028e-05,
|
| 471 |
+
"loss": 0.2951,
|
| 472 |
+
"step": 1375
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 7.652949245541838,
|
| 476 |
+
"grad_norm": 195.25039672851562,
|
| 477 |
+
"learning_rate": 1.3727990285367337e-05,
|
| 478 |
+
"loss": 0.3733,
|
| 479 |
+
"step": 1400
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 7.790123456790123,
|
| 483 |
+
"grad_norm": 3.2258872985839844,
|
| 484 |
+
"learning_rate": 1.3576199149969643e-05,
|
| 485 |
+
"loss": 0.2619,
|
| 486 |
+
"step": 1425
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 7.927297668038409,
|
| 490 |
+
"grad_norm": 267.19146728515625,
|
| 491 |
+
"learning_rate": 1.342440801457195e-05,
|
| 492 |
+
"loss": 0.2622,
|
| 493 |
+
"step": 1450
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 8.0,
|
| 497 |
+
"eval_accuracy": 0.8519134775374376,
|
| 498 |
+
"eval_loss": 0.7531983256340027,
|
| 499 |
+
"eval_macro_f1": 0.8512858438076938,
|
| 500 |
+
"eval_runtime": 0.4259,
|
| 501 |
+
"eval_samples_per_second": 1411.28,
|
| 502 |
+
"eval_steps_per_second": 89.232,
|
| 503 |
+
"eval_weighted_f1": 0.8510017584279584,
|
| 504 |
+
"step": 1464
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 8.060356652949245,
|
| 508 |
+
"grad_norm": 91.1015625,
|
| 509 |
+
"learning_rate": 1.3272616879174259e-05,
|
| 510 |
+
"loss": 0.2169,
|
| 511 |
+
"step": 1475
|
| 512 |
+
},
|
| 513 |
+
{
|
| 514 |
+
"epoch": 8.197530864197532,
|
| 515 |
+
"grad_norm": 2.820073366165161,
|
| 516 |
+
"learning_rate": 1.3120825743776565e-05,
|
| 517 |
+
"loss": 0.4201,
|
| 518 |
+
"step": 1500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 8.334705075445816,
|
| 522 |
+
"grad_norm": 3.838801860809326,
|
| 523 |
+
"learning_rate": 1.2969034608378872e-05,
|
| 524 |
+
"loss": 0.1847,
|
| 525 |
+
"step": 1525
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 8.4718792866941,
|
| 529 |
+
"grad_norm": 0.20370107889175415,
|
| 530 |
+
"learning_rate": 1.2817243472981178e-05,
|
| 531 |
+
"loss": 0.1535,
|
| 532 |
+
"step": 1550
|
| 533 |
+
},
|
| 534 |
+
{
|
| 535 |
+
"epoch": 8.609053497942387,
|
| 536 |
+
"grad_norm": 0.616696834564209,
|
| 537 |
+
"learning_rate": 1.2665452337583487e-05,
|
| 538 |
+
"loss": 0.3144,
|
| 539 |
+
"step": 1575
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 8.746227709190672,
|
| 543 |
+
"grad_norm": 0.3781672418117523,
|
| 544 |
+
"learning_rate": 1.2513661202185794e-05,
|
| 545 |
+
"loss": 0.2821,
|
| 546 |
+
"step": 1600
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 8.883401920438958,
|
| 550 |
+
"grad_norm": 15.68051815032959,
|
| 551 |
+
"learning_rate": 1.23618700667881e-05,
|
| 552 |
+
"loss": 0.3461,
|
| 553 |
+
"step": 1625
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 9.0,
|
| 557 |
+
"eval_accuracy": 0.848585690515807,
|
| 558 |
+
"eval_loss": 0.7961008548736572,
|
| 559 |
+
"eval_macro_f1": 0.8484817640151023,
|
| 560 |
+
"eval_runtime": 0.4578,
|
| 561 |
+
"eval_samples_per_second": 1312.835,
|
| 562 |
+
"eval_steps_per_second": 83.008,
|
| 563 |
+
"eval_weighted_f1": 0.8482893105525164,
|
| 564 |
+
"step": 1647
|
| 565 |
+
}
|
| 566 |
+
],
|
| 567 |
+
"logging_steps": 25,
|
| 568 |
+
"max_steps": 3660,
|
| 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-03-s44-20260826T161012Z-ab4112e3/trainer/checkpoint-1647/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-source-source-04-s44-20260826T161317Z-ab4112e3/trainer/checkpoint-1232/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-source-source-04-s44-20260826T161317Z-ab4112e3/trainer/checkpoint-1232/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|