Add files using upload-large-folder tool
Browse files- ood-vifactcheck-phobert-topic-domain-06-s43-20260826T103734Z-bdd2c308/trainer/checkpoint-1288/added_tokens.json +3 -0
- ood-vifactcheck-phobert-topic-domain-06-s43-20260826T103734Z-bdd2c308/trainer/checkpoint-1288/bpe.codes +0 -0
- ood-vifactcheck-phobert-topic-domain-06-s43-20260826T103734Z-bdd2c308/trainer/checkpoint-1288/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-topic-domain-06-s43-20260826T103734Z-bdd2c308/trainer/checkpoint-1288/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-topic-domain-06-s43-20260826T103734Z-bdd2c308/trainer/checkpoint-1288/trainer_state.json +477 -0
- ood-vifactcheck-phobert-topic-domain-06-s43-20260826T103734Z-bdd2c308/trainer/checkpoint-1288/vocab.txt +0 -0
- ood-vifactcheck-phobert-topic-domain-06-s43-20260826T103734Z-bdd2c308/trainer/checkpoint-736/added_tokens.json +3 -0
- ood-vifactcheck-phobert-topic-domain-06-s43-20260826T103734Z-bdd2c308/trainer/checkpoint-736/bpe.codes +0 -0
- ood-vifactcheck-phobert-topic-domain-06-s43-20260826T103734Z-bdd2c308/trainer/checkpoint-736/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-topic-domain-06-s43-20260826T103734Z-bdd2c308/trainer/checkpoint-736/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-topic-domain-06-s43-20260826T103734Z-bdd2c308/trainer/checkpoint-736/trainer_state.json +290 -0
- ood-vifactcheck-phobert-topic-domain-06-s43-20260826T103734Z-bdd2c308/trainer/checkpoint-736/vocab.txt +0 -0
- ood-vifactcheck-phobert-topic-domain-07-s43-20260826T104017Z-bdd2c308/trainer/checkpoint-1576/added_tokens.json +3 -0
- ood-vifactcheck-phobert-topic-domain-07-s43-20260826T104017Z-bdd2c308/trainer/checkpoint-1576/bpe.codes +0 -0
- ood-vifactcheck-phobert-topic-domain-07-s43-20260826T104017Z-bdd2c308/trainer/checkpoint-1576/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-topic-domain-07-s43-20260826T104017Z-bdd2c308/trainer/checkpoint-1576/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-topic-domain-07-s43-20260826T104017Z-bdd2c308/trainer/checkpoint-1576/trainer_state.json +572 -0
- ood-vifactcheck-phobert-topic-domain-07-s43-20260826T104017Z-bdd2c308/trainer/checkpoint-1576/vocab.txt +0 -0
- ood-vifactcheck-phobert-topic-domain-07-s43-20260826T104017Z-bdd2c308/trainer/checkpoint-2167/added_tokens.json +3 -0
- ood-vifactcheck-phobert-topic-domain-07-s43-20260826T104017Z-bdd2c308/trainer/checkpoint-2167/bpe.codes +0 -0
- ood-vifactcheck-phobert-topic-domain-07-s43-20260826T104017Z-bdd2c308/trainer/checkpoint-2167/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-topic-domain-07-s43-20260826T104017Z-bdd2c308/trainer/checkpoint-2167/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-topic-domain-07-s43-20260826T104017Z-bdd2c308/trainer/checkpoint-2167/trainer_state.json +766 -0
- ood-vifactcheck-phobert-topic-domain-07-s43-20260826T104017Z-bdd2c308/trainer/checkpoint-2167/vocab.txt +0 -0
- ood-vifactcheck-phobert-topic-domain-08-s43-20260826T104414Z-bdd2c308/trainer/checkpoint-1700/added_tokens.json +3 -0
- ood-vifactcheck-phobert-topic-domain-08-s43-20260826T104414Z-bdd2c308/trainer/checkpoint-1700/bpe.codes +0 -0
- ood-vifactcheck-phobert-topic-domain-08-s43-20260826T104414Z-bdd2c308/trainer/checkpoint-1700/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-topic-domain-08-s43-20260826T104414Z-bdd2c308/trainer/checkpoint-1700/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-topic-domain-08-s43-20260826T104414Z-bdd2c308/trainer/checkpoint-1700/trainer_state.json +629 -0
- ood-vifactcheck-phobert-topic-domain-08-s43-20260826T104414Z-bdd2c308/trainer/checkpoint-1700/vocab.txt +0 -0
- ood-vifactcheck-phobert-topic-domain-08-s43-20260826T104414Z-bdd2c308/trainer/checkpoint-2210/added_tokens.json +3 -0
- ood-vifactcheck-phobert-topic-domain-08-s43-20260826T104414Z-bdd2c308/trainer/checkpoint-2210/bpe.codes +0 -0
- ood-vifactcheck-phobert-topic-domain-08-s43-20260826T104414Z-bdd2c308/trainer/checkpoint-2210/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-topic-domain-08-s43-20260826T104414Z-bdd2c308/trainer/checkpoint-2210/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-topic-domain-08-s43-20260826T104414Z-bdd2c308/trainer/checkpoint-2210/trainer_state.json +802 -0
- ood-vifactcheck-phobert-topic-domain-08-s43-20260826T104414Z-bdd2c308/trainer/checkpoint-2210/vocab.txt +0 -0
- ood-vifactcheck-phobert-topic-domain-09-s43-20260826T104827Z-bdd2c308/trainer/checkpoint-1309/added_tokens.json +3 -0
- ood-vifactcheck-phobert-topic-domain-09-s43-20260826T104827Z-bdd2c308/trainer/checkpoint-1309/bpe.codes +0 -0
- ood-vifactcheck-phobert-topic-domain-09-s43-20260826T104827Z-bdd2c308/trainer/checkpoint-1309/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-topic-domain-09-s43-20260826T104827Z-bdd2c308/trainer/checkpoint-1309/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-topic-domain-09-s43-20260826T104827Z-bdd2c308/trainer/checkpoint-1309/trainer_state.json +484 -0
- ood-vifactcheck-phobert-topic-domain-09-s43-20260826T104827Z-bdd2c308/trainer/checkpoint-1309/vocab.txt +0 -0
- ood-vifactcheck-phobert-topic-domain-09-s43-20260826T104827Z-bdd2c308/trainer/checkpoint-748/added_tokens.json +3 -0
- ood-vifactcheck-phobert-topic-domain-09-s43-20260826T104827Z-bdd2c308/trainer/checkpoint-748/bpe.codes +0 -0
- ood-vifactcheck-phobert-topic-domain-09-s43-20260826T104827Z-bdd2c308/trainer/checkpoint-748/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-topic-domain-09-s43-20260826T104827Z-bdd2c308/trainer/checkpoint-748/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-topic-domain-09-s43-20260826T104827Z-bdd2c308/trainer/checkpoint-748/trainer_state.json +290 -0
- ood-vifactcheck-phobert-topic-domain-09-s43-20260826T104827Z-bdd2c308/trainer/checkpoint-748/vocab.txt +0 -0
- ood-vifactcheck-phobert-topic-domain-10-s43-20260826T105057Z-bdd2c308/trainer/checkpoint-1158/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-topic-domain-10-s43-20260826T105057Z-bdd2c308/trainer/checkpoint-1158/tokenizer_config.json +55 -0
ood-vifactcheck-phobert-topic-domain-06-s43-20260826T103734Z-bdd2c308/trainer/checkpoint-1288/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-topic-domain-06-s43-20260826T103734Z-bdd2c308/trainer/checkpoint-1288/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-06-s43-20260826T103734Z-bdd2c308/trainer/checkpoint-1288/special_tokens_map.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": "<mask>",
|
| 6 |
+
"pad_token": "<pad>",
|
| 7 |
+
"sep_token": "</s>",
|
| 8 |
+
"unk_token": "<unk>"
|
| 9 |
+
}
|
ood-vifactcheck-phobert-topic-domain-06-s43-20260826T103734Z-bdd2c308/trainer/checkpoint-1288/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"64000": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 256,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "PhobertTokenizer",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-phobert-topic-domain-06-s43-20260826T103734Z-bdd2c308/trainer/checkpoint-1288/trainer_state.json
ADDED
|
@@ -0,0 +1,477 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 736,
|
| 3 |
+
"best_metric": 0.8503029635167291,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-topic-domain-06-s43-20260826T103734Z-bdd2c308/trainer/checkpoint-736",
|
| 5 |
+
"epoch": 7.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1288,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.1364256480218281,
|
| 14 |
+
"grad_norm": 10.707963943481445,
|
| 15 |
+
"learning_rate": 1.3043478260869566e-06,
|
| 16 |
+
"loss": 4.477,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.2728512960436562,
|
| 21 |
+
"grad_norm": 7.000658988952637,
|
| 22 |
+
"learning_rate": 2.6630434782608698e-06,
|
| 23 |
+
"loss": 4.4246,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.4092769440654843,
|
| 28 |
+
"grad_norm": 7.4163923263549805,
|
| 29 |
+
"learning_rate": 4.021739130434783e-06,
|
| 30 |
+
"loss": 4.3889,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5457025920873124,
|
| 35 |
+
"grad_norm": 9.124894142150879,
|
| 36 |
+
"learning_rate": 5.380434782608695e-06,
|
| 37 |
+
"loss": 4.3538,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6821282401091405,
|
| 42 |
+
"grad_norm": 11.21931266784668,
|
| 43 |
+
"learning_rate": 6.739130434782609e-06,
|
| 44 |
+
"loss": 4.3169,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8185538881309686,
|
| 49 |
+
"grad_norm": 10.78498363494873,
|
| 50 |
+
"learning_rate": 8.097826086956523e-06,
|
| 51 |
+
"loss": 4.272,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9549795361527967,
|
| 56 |
+
"grad_norm": 17.10068130493164,
|
| 57 |
+
"learning_rate": 9.456521739130436e-06,
|
| 58 |
+
"loss": 4.1387,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.5842245989304813,
|
| 64 |
+
"eval_loss": 0.9436517953872681,
|
| 65 |
+
"eval_macro_f1": 0.578389871607064,
|
| 66 |
+
"eval_runtime": 0.4863,
|
| 67 |
+
"eval_samples_per_second": 1538.077,
|
| 68 |
+
"eval_steps_per_second": 96.644,
|
| 69 |
+
"eval_weighted_f1": 0.580405545899353,
|
| 70 |
+
"step": 184
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.08731241473397,
|
| 74 |
+
"grad_norm": 34.56974411010742,
|
| 75 |
+
"learning_rate": 1.0815217391304349e-05,
|
| 76 |
+
"loss": 3.72,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.223738062755798,
|
| 81 |
+
"grad_norm": 25.000581741333008,
|
| 82 |
+
"learning_rate": 1.2173913043478263e-05,
|
| 83 |
+
"loss": 3.6713,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.360163710777626,
|
| 88 |
+
"grad_norm": 23.774730682373047,
|
| 89 |
+
"learning_rate": 1.3532608695652174e-05,
|
| 90 |
+
"loss": 3.2532,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.4965893587994543,
|
| 95 |
+
"grad_norm": 22.013011932373047,
|
| 96 |
+
"learning_rate": 1.4891304347826087e-05,
|
| 97 |
+
"loss": 2.8507,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.6330150068212825,
|
| 102 |
+
"grad_norm": 65.38465881347656,
|
| 103 |
+
"learning_rate": 1.6250000000000002e-05,
|
| 104 |
+
"loss": 2.5617,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7694406548431105,
|
| 109 |
+
"grad_norm": 36.88359069824219,
|
| 110 |
+
"learning_rate": 1.7608695652173915e-05,
|
| 111 |
+
"loss": 2.5845,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.9058663028649385,
|
| 116 |
+
"grad_norm": 53.36519241333008,
|
| 117 |
+
"learning_rate": 1.8967391304347828e-05,
|
| 118 |
+
"loss": 2.1653,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.7901069518716578,
|
| 124 |
+
"eval_loss": 0.5848486423492432,
|
| 125 |
+
"eval_macro_f1": 0.7882365349077537,
|
| 126 |
+
"eval_runtime": 0.5002,
|
| 127 |
+
"eval_samples_per_second": 1495.473,
|
| 128 |
+
"eval_steps_per_second": 93.967,
|
| 129 |
+
"eval_weighted_f1": 0.7886694705625077,
|
| 130 |
+
"step": 368
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.038199181446112,
|
| 134 |
+
"grad_norm": 40.243167877197266,
|
| 135 |
+
"learning_rate": 1.9963768115942032e-05,
|
| 136 |
+
"loss": 2.0896,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.17462482946794,
|
| 141 |
+
"grad_norm": 31.98969841003418,
|
| 142 |
+
"learning_rate": 1.9812801932367153e-05,
|
| 143 |
+
"loss": 1.7439,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.311050477489768,
|
| 148 |
+
"grad_norm": 25.488317489624023,
|
| 149 |
+
"learning_rate": 1.966183574879227e-05,
|
| 150 |
+
"loss": 1.8658,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.447476125511596,
|
| 155 |
+
"grad_norm": 40.44835662841797,
|
| 156 |
+
"learning_rate": 1.9510869565217393e-05,
|
| 157 |
+
"loss": 1.6862,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.5839017735334244,
|
| 162 |
+
"grad_norm": 21.379457473754883,
|
| 163 |
+
"learning_rate": 1.9359903381642515e-05,
|
| 164 |
+
"loss": 1.6167,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.720327421555252,
|
| 169 |
+
"grad_norm": 43.95497512817383,
|
| 170 |
+
"learning_rate": 1.9208937198067633e-05,
|
| 171 |
+
"loss": 1.932,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.8567530695770804,
|
| 176 |
+
"grad_norm": 34.28624725341797,
|
| 177 |
+
"learning_rate": 1.9057971014492754e-05,
|
| 178 |
+
"loss": 1.7187,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 2.9931787175989086,
|
| 183 |
+
"grad_norm": 53.723106384277344,
|
| 184 |
+
"learning_rate": 1.8907004830917876e-05,
|
| 185 |
+
"loss": 1.5052,
|
| 186 |
+
"step": 550
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 3.0,
|
| 190 |
+
"eval_accuracy": 0.8409090909090909,
|
| 191 |
+
"eval_loss": 0.46382901072502136,
|
| 192 |
+
"eval_macro_f1": 0.8396040728891307,
|
| 193 |
+
"eval_runtime": 0.4934,
|
| 194 |
+
"eval_samples_per_second": 1515.993,
|
| 195 |
+
"eval_steps_per_second": 95.256,
|
| 196 |
+
"eval_weighted_f1": 0.8400092576092671,
|
| 197 |
+
"step": 552
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.1255115961800817,
|
| 201 |
+
"grad_norm": 49.558616638183594,
|
| 202 |
+
"learning_rate": 1.8756038647342997e-05,
|
| 203 |
+
"loss": 1.071,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.26193724420191,
|
| 208 |
+
"grad_norm": 12.786978721618652,
|
| 209 |
+
"learning_rate": 1.8605072463768115e-05,
|
| 210 |
+
"loss": 1.2295,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.398362892223738,
|
| 215 |
+
"grad_norm": 20.064556121826172,
|
| 216 |
+
"learning_rate": 1.8454106280193237e-05,
|
| 217 |
+
"loss": 1.0179,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.5347885402455663,
|
| 222 |
+
"grad_norm": 22.385499954223633,
|
| 223 |
+
"learning_rate": 1.830314009661836e-05,
|
| 224 |
+
"loss": 1.048,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.6712141882673945,
|
| 229 |
+
"grad_norm": 89.25599670410156,
|
| 230 |
+
"learning_rate": 1.815217391304348e-05,
|
| 231 |
+
"loss": 1.1422,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.8076398362892223,
|
| 236 |
+
"grad_norm": 42.8995361328125,
|
| 237 |
+
"learning_rate": 1.8001207729468598e-05,
|
| 238 |
+
"loss": 1.1751,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.9440654843110505,
|
| 243 |
+
"grad_norm": 42.35470199584961,
|
| 244 |
+
"learning_rate": 1.785024154589372e-05,
|
| 245 |
+
"loss": 1.413,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 4.0,
|
| 250 |
+
"eval_accuracy": 0.8502673796791443,
|
| 251 |
+
"eval_loss": 0.45021700859069824,
|
| 252 |
+
"eval_macro_f1": 0.8503029635167291,
|
| 253 |
+
"eval_runtime": 0.5171,
|
| 254 |
+
"eval_samples_per_second": 1446.657,
|
| 255 |
+
"eval_steps_per_second": 90.9,
|
| 256 |
+
"eval_weighted_f1": 0.8504358554523623,
|
| 257 |
+
"step": 736
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.076398362892224,
|
| 261 |
+
"grad_norm": 36.52999496459961,
|
| 262 |
+
"learning_rate": 1.769927536231884e-05,
|
| 263 |
+
"loss": 0.8835,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.212824010914052,
|
| 268 |
+
"grad_norm": 39.73047637939453,
|
| 269 |
+
"learning_rate": 1.7548309178743963e-05,
|
| 270 |
+
"loss": 0.8655,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.34924965893588,
|
| 275 |
+
"grad_norm": 11.442521095275879,
|
| 276 |
+
"learning_rate": 1.7397342995169084e-05,
|
| 277 |
+
"loss": 0.8858,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.485675306957708,
|
| 282 |
+
"grad_norm": 19.212554931640625,
|
| 283 |
+
"learning_rate": 1.7246376811594206e-05,
|
| 284 |
+
"loss": 0.6952,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.622100954979536,
|
| 289 |
+
"grad_norm": 5.703792095184326,
|
| 290 |
+
"learning_rate": 1.7095410628019324e-05,
|
| 291 |
+
"loss": 0.641,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.758526603001364,
|
| 296 |
+
"grad_norm": 39.71404266357422,
|
| 297 |
+
"learning_rate": 1.6944444444444446e-05,
|
| 298 |
+
"loss": 0.7758,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.894952251023192,
|
| 303 |
+
"grad_norm": 2.4770050048828125,
|
| 304 |
+
"learning_rate": 1.6793478260869567e-05,
|
| 305 |
+
"loss": 0.6994,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 5.0,
|
| 310 |
+
"eval_accuracy": 0.8435828877005348,
|
| 311 |
+
"eval_loss": 0.5535889267921448,
|
| 312 |
+
"eval_macro_f1": 0.8430263347963173,
|
| 313 |
+
"eval_runtime": 0.5027,
|
| 314 |
+
"eval_samples_per_second": 1488.053,
|
| 315 |
+
"eval_steps_per_second": 93.501,
|
| 316 |
+
"eval_weighted_f1": 0.8433197619097942,
|
| 317 |
+
"step": 920
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.0272851296043655,
|
| 321 |
+
"grad_norm": 3.105264902114868,
|
| 322 |
+
"learning_rate": 1.664251207729469e-05,
|
| 323 |
+
"loss": 0.7367,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.163710777626194,
|
| 328 |
+
"grad_norm": 58.483734130859375,
|
| 329 |
+
"learning_rate": 1.649154589371981e-05,
|
| 330 |
+
"loss": 0.5008,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.300136425648022,
|
| 335 |
+
"grad_norm": 11.83707046508789,
|
| 336 |
+
"learning_rate": 1.634057971014493e-05,
|
| 337 |
+
"loss": 0.4637,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.43656207366985,
|
| 342 |
+
"grad_norm": 14.147911071777344,
|
| 343 |
+
"learning_rate": 1.618961352657005e-05,
|
| 344 |
+
"loss": 0.6002,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.572987721691678,
|
| 349 |
+
"grad_norm": 13.883936882019043,
|
| 350 |
+
"learning_rate": 1.603864734299517e-05,
|
| 351 |
+
"loss": 0.4841,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.709413369713506,
|
| 356 |
+
"grad_norm": 44.2625846862793,
|
| 357 |
+
"learning_rate": 1.5887681159420293e-05,
|
| 358 |
+
"loss": 0.5551,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.845839017735334,
|
| 363 |
+
"grad_norm": 70.52825927734375,
|
| 364 |
+
"learning_rate": 1.5736714975845415e-05,
|
| 365 |
+
"loss": 0.5785,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 5.982264665757162,
|
| 370 |
+
"grad_norm": 97.37578582763672,
|
| 371 |
+
"learning_rate": 1.5585748792270533e-05,
|
| 372 |
+
"loss": 0.5314,
|
| 373 |
+
"step": 1100
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 6.0,
|
| 377 |
+
"eval_accuracy": 0.8235294117647058,
|
| 378 |
+
"eval_loss": 0.704562246799469,
|
| 379 |
+
"eval_macro_f1": 0.8248513306939934,
|
| 380 |
+
"eval_runtime": 0.5046,
|
| 381 |
+
"eval_samples_per_second": 1482.294,
|
| 382 |
+
"eval_steps_per_second": 93.139,
|
| 383 |
+
"eval_weighted_f1": 0.8252133604269288,
|
| 384 |
+
"step": 1104
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 6.114597544338336,
|
| 388 |
+
"grad_norm": 18.764827728271484,
|
| 389 |
+
"learning_rate": 1.5434782608695654e-05,
|
| 390 |
+
"loss": 0.5193,
|
| 391 |
+
"step": 1125
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 6.251023192360163,
|
| 395 |
+
"grad_norm": 80.26335906982422,
|
| 396 |
+
"learning_rate": 1.5283816425120776e-05,
|
| 397 |
+
"loss": 0.5252,
|
| 398 |
+
"step": 1150
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.387448840381992,
|
| 402 |
+
"grad_norm": 32.1051139831543,
|
| 403 |
+
"learning_rate": 1.5132850241545896e-05,
|
| 404 |
+
"loss": 0.3743,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.52387448840382,
|
| 409 |
+
"grad_norm": 40.98142623901367,
|
| 410 |
+
"learning_rate": 1.4981884057971017e-05,
|
| 411 |
+
"loss": 0.4355,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.660300136425648,
|
| 416 |
+
"grad_norm": 2.946929693222046,
|
| 417 |
+
"learning_rate": 1.4830917874396135e-05,
|
| 418 |
+
"loss": 0.4702,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.796725784447476,
|
| 423 |
+
"grad_norm": 34.52414321899414,
|
| 424 |
+
"learning_rate": 1.4679951690821257e-05,
|
| 425 |
+
"loss": 0.5157,
|
| 426 |
+
"step": 1250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 6.933151432469304,
|
| 430 |
+
"grad_norm": 84.21067810058594,
|
| 431 |
+
"learning_rate": 1.4528985507246378e-05,
|
| 432 |
+
"loss": 0.397,
|
| 433 |
+
"step": 1275
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 7.0,
|
| 437 |
+
"eval_accuracy": 0.8382352941176471,
|
| 438 |
+
"eval_loss": 0.710689127445221,
|
| 439 |
+
"eval_macro_f1": 0.8376373389124337,
|
| 440 |
+
"eval_runtime": 0.5162,
|
| 441 |
+
"eval_samples_per_second": 1449.117,
|
| 442 |
+
"eval_steps_per_second": 91.054,
|
| 443 |
+
"eval_weighted_f1": 0.8379628056493621,
|
| 444 |
+
"step": 1288
|
| 445 |
+
}
|
| 446 |
+
],
|
| 447 |
+
"logging_steps": 25,
|
| 448 |
+
"max_steps": 3680,
|
| 449 |
+
"num_input_tokens_seen": 0,
|
| 450 |
+
"num_train_epochs": 20,
|
| 451 |
+
"save_steps": 500,
|
| 452 |
+
"stateful_callbacks": {
|
| 453 |
+
"EarlyStoppingCallback": {
|
| 454 |
+
"args": {
|
| 455 |
+
"early_stopping_patience": 3,
|
| 456 |
+
"early_stopping_threshold": 0.0
|
| 457 |
+
},
|
| 458 |
+
"attributes": {
|
| 459 |
+
"early_stopping_patience_counter": 3
|
| 460 |
+
}
|
| 461 |
+
},
|
| 462 |
+
"TrainerControl": {
|
| 463 |
+
"args": {
|
| 464 |
+
"should_epoch_stop": false,
|
| 465 |
+
"should_evaluate": false,
|
| 466 |
+
"should_log": false,
|
| 467 |
+
"should_save": true,
|
| 468 |
+
"should_training_stop": true
|
| 469 |
+
},
|
| 470 |
+
"attributes": {}
|
| 471 |
+
}
|
| 472 |
+
},
|
| 473 |
+
"total_flos": 0.0,
|
| 474 |
+
"train_batch_size": 8,
|
| 475 |
+
"trial_name": null,
|
| 476 |
+
"trial_params": null
|
| 477 |
+
}
|
ood-vifactcheck-phobert-topic-domain-06-s43-20260826T103734Z-bdd2c308/trainer/checkpoint-1288/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-06-s43-20260826T103734Z-bdd2c308/trainer/checkpoint-736/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-topic-domain-06-s43-20260826T103734Z-bdd2c308/trainer/checkpoint-736/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-06-s43-20260826T103734Z-bdd2c308/trainer/checkpoint-736/special_tokens_map.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": "<mask>",
|
| 6 |
+
"pad_token": "<pad>",
|
| 7 |
+
"sep_token": "</s>",
|
| 8 |
+
"unk_token": "<unk>"
|
| 9 |
+
}
|
ood-vifactcheck-phobert-topic-domain-06-s43-20260826T103734Z-bdd2c308/trainer/checkpoint-736/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"64000": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 256,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "PhobertTokenizer",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-phobert-topic-domain-06-s43-20260826T103734Z-bdd2c308/trainer/checkpoint-736/trainer_state.json
ADDED
|
@@ -0,0 +1,290 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 736,
|
| 3 |
+
"best_metric": 0.8503029635167291,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-topic-domain-06-s43-20260826T103734Z-bdd2c308/trainer/checkpoint-736",
|
| 5 |
+
"epoch": 4.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 736,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.1364256480218281,
|
| 14 |
+
"grad_norm": 10.707963943481445,
|
| 15 |
+
"learning_rate": 1.3043478260869566e-06,
|
| 16 |
+
"loss": 4.477,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.2728512960436562,
|
| 21 |
+
"grad_norm": 7.000658988952637,
|
| 22 |
+
"learning_rate": 2.6630434782608698e-06,
|
| 23 |
+
"loss": 4.4246,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.4092769440654843,
|
| 28 |
+
"grad_norm": 7.4163923263549805,
|
| 29 |
+
"learning_rate": 4.021739130434783e-06,
|
| 30 |
+
"loss": 4.3889,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5457025920873124,
|
| 35 |
+
"grad_norm": 9.124894142150879,
|
| 36 |
+
"learning_rate": 5.380434782608695e-06,
|
| 37 |
+
"loss": 4.3538,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6821282401091405,
|
| 42 |
+
"grad_norm": 11.21931266784668,
|
| 43 |
+
"learning_rate": 6.739130434782609e-06,
|
| 44 |
+
"loss": 4.3169,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8185538881309686,
|
| 49 |
+
"grad_norm": 10.78498363494873,
|
| 50 |
+
"learning_rate": 8.097826086956523e-06,
|
| 51 |
+
"loss": 4.272,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9549795361527967,
|
| 56 |
+
"grad_norm": 17.10068130493164,
|
| 57 |
+
"learning_rate": 9.456521739130436e-06,
|
| 58 |
+
"loss": 4.1387,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.5842245989304813,
|
| 64 |
+
"eval_loss": 0.9436517953872681,
|
| 65 |
+
"eval_macro_f1": 0.578389871607064,
|
| 66 |
+
"eval_runtime": 0.4863,
|
| 67 |
+
"eval_samples_per_second": 1538.077,
|
| 68 |
+
"eval_steps_per_second": 96.644,
|
| 69 |
+
"eval_weighted_f1": 0.580405545899353,
|
| 70 |
+
"step": 184
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.08731241473397,
|
| 74 |
+
"grad_norm": 34.56974411010742,
|
| 75 |
+
"learning_rate": 1.0815217391304349e-05,
|
| 76 |
+
"loss": 3.72,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.223738062755798,
|
| 81 |
+
"grad_norm": 25.000581741333008,
|
| 82 |
+
"learning_rate": 1.2173913043478263e-05,
|
| 83 |
+
"loss": 3.6713,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.360163710777626,
|
| 88 |
+
"grad_norm": 23.774730682373047,
|
| 89 |
+
"learning_rate": 1.3532608695652174e-05,
|
| 90 |
+
"loss": 3.2532,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.4965893587994543,
|
| 95 |
+
"grad_norm": 22.013011932373047,
|
| 96 |
+
"learning_rate": 1.4891304347826087e-05,
|
| 97 |
+
"loss": 2.8507,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.6330150068212825,
|
| 102 |
+
"grad_norm": 65.38465881347656,
|
| 103 |
+
"learning_rate": 1.6250000000000002e-05,
|
| 104 |
+
"loss": 2.5617,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7694406548431105,
|
| 109 |
+
"grad_norm": 36.88359069824219,
|
| 110 |
+
"learning_rate": 1.7608695652173915e-05,
|
| 111 |
+
"loss": 2.5845,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.9058663028649385,
|
| 116 |
+
"grad_norm": 53.36519241333008,
|
| 117 |
+
"learning_rate": 1.8967391304347828e-05,
|
| 118 |
+
"loss": 2.1653,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.7901069518716578,
|
| 124 |
+
"eval_loss": 0.5848486423492432,
|
| 125 |
+
"eval_macro_f1": 0.7882365349077537,
|
| 126 |
+
"eval_runtime": 0.5002,
|
| 127 |
+
"eval_samples_per_second": 1495.473,
|
| 128 |
+
"eval_steps_per_second": 93.967,
|
| 129 |
+
"eval_weighted_f1": 0.7886694705625077,
|
| 130 |
+
"step": 368
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.038199181446112,
|
| 134 |
+
"grad_norm": 40.243167877197266,
|
| 135 |
+
"learning_rate": 1.9963768115942032e-05,
|
| 136 |
+
"loss": 2.0896,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.17462482946794,
|
| 141 |
+
"grad_norm": 31.98969841003418,
|
| 142 |
+
"learning_rate": 1.9812801932367153e-05,
|
| 143 |
+
"loss": 1.7439,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.311050477489768,
|
| 148 |
+
"grad_norm": 25.488317489624023,
|
| 149 |
+
"learning_rate": 1.966183574879227e-05,
|
| 150 |
+
"loss": 1.8658,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.447476125511596,
|
| 155 |
+
"grad_norm": 40.44835662841797,
|
| 156 |
+
"learning_rate": 1.9510869565217393e-05,
|
| 157 |
+
"loss": 1.6862,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.5839017735334244,
|
| 162 |
+
"grad_norm": 21.379457473754883,
|
| 163 |
+
"learning_rate": 1.9359903381642515e-05,
|
| 164 |
+
"loss": 1.6167,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.720327421555252,
|
| 169 |
+
"grad_norm": 43.95497512817383,
|
| 170 |
+
"learning_rate": 1.9208937198067633e-05,
|
| 171 |
+
"loss": 1.932,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.8567530695770804,
|
| 176 |
+
"grad_norm": 34.28624725341797,
|
| 177 |
+
"learning_rate": 1.9057971014492754e-05,
|
| 178 |
+
"loss": 1.7187,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 2.9931787175989086,
|
| 183 |
+
"grad_norm": 53.723106384277344,
|
| 184 |
+
"learning_rate": 1.8907004830917876e-05,
|
| 185 |
+
"loss": 1.5052,
|
| 186 |
+
"step": 550
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 3.0,
|
| 190 |
+
"eval_accuracy": 0.8409090909090909,
|
| 191 |
+
"eval_loss": 0.46382901072502136,
|
| 192 |
+
"eval_macro_f1": 0.8396040728891307,
|
| 193 |
+
"eval_runtime": 0.4934,
|
| 194 |
+
"eval_samples_per_second": 1515.993,
|
| 195 |
+
"eval_steps_per_second": 95.256,
|
| 196 |
+
"eval_weighted_f1": 0.8400092576092671,
|
| 197 |
+
"step": 552
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.1255115961800817,
|
| 201 |
+
"grad_norm": 49.558616638183594,
|
| 202 |
+
"learning_rate": 1.8756038647342997e-05,
|
| 203 |
+
"loss": 1.071,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.26193724420191,
|
| 208 |
+
"grad_norm": 12.786978721618652,
|
| 209 |
+
"learning_rate": 1.8605072463768115e-05,
|
| 210 |
+
"loss": 1.2295,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.398362892223738,
|
| 215 |
+
"grad_norm": 20.064556121826172,
|
| 216 |
+
"learning_rate": 1.8454106280193237e-05,
|
| 217 |
+
"loss": 1.0179,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.5347885402455663,
|
| 222 |
+
"grad_norm": 22.385499954223633,
|
| 223 |
+
"learning_rate": 1.830314009661836e-05,
|
| 224 |
+
"loss": 1.048,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.6712141882673945,
|
| 229 |
+
"grad_norm": 89.25599670410156,
|
| 230 |
+
"learning_rate": 1.815217391304348e-05,
|
| 231 |
+
"loss": 1.1422,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.8076398362892223,
|
| 236 |
+
"grad_norm": 42.8995361328125,
|
| 237 |
+
"learning_rate": 1.8001207729468598e-05,
|
| 238 |
+
"loss": 1.1751,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.9440654843110505,
|
| 243 |
+
"grad_norm": 42.35470199584961,
|
| 244 |
+
"learning_rate": 1.785024154589372e-05,
|
| 245 |
+
"loss": 1.413,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 4.0,
|
| 250 |
+
"eval_accuracy": 0.8502673796791443,
|
| 251 |
+
"eval_loss": 0.45021700859069824,
|
| 252 |
+
"eval_macro_f1": 0.8503029635167291,
|
| 253 |
+
"eval_runtime": 0.5171,
|
| 254 |
+
"eval_samples_per_second": 1446.657,
|
| 255 |
+
"eval_steps_per_second": 90.9,
|
| 256 |
+
"eval_weighted_f1": 0.8504358554523623,
|
| 257 |
+
"step": 736
|
| 258 |
+
}
|
| 259 |
+
],
|
| 260 |
+
"logging_steps": 25,
|
| 261 |
+
"max_steps": 3680,
|
| 262 |
+
"num_input_tokens_seen": 0,
|
| 263 |
+
"num_train_epochs": 20,
|
| 264 |
+
"save_steps": 500,
|
| 265 |
+
"stateful_callbacks": {
|
| 266 |
+
"EarlyStoppingCallback": {
|
| 267 |
+
"args": {
|
| 268 |
+
"early_stopping_patience": 3,
|
| 269 |
+
"early_stopping_threshold": 0.0
|
| 270 |
+
},
|
| 271 |
+
"attributes": {
|
| 272 |
+
"early_stopping_patience_counter": 0
|
| 273 |
+
}
|
| 274 |
+
},
|
| 275 |
+
"TrainerControl": {
|
| 276 |
+
"args": {
|
| 277 |
+
"should_epoch_stop": false,
|
| 278 |
+
"should_evaluate": false,
|
| 279 |
+
"should_log": false,
|
| 280 |
+
"should_save": true,
|
| 281 |
+
"should_training_stop": false
|
| 282 |
+
},
|
| 283 |
+
"attributes": {}
|
| 284 |
+
}
|
| 285 |
+
},
|
| 286 |
+
"total_flos": 0.0,
|
| 287 |
+
"train_batch_size": 8,
|
| 288 |
+
"trial_name": null,
|
| 289 |
+
"trial_params": null
|
| 290 |
+
}
|
ood-vifactcheck-phobert-topic-domain-06-s43-20260826T103734Z-bdd2c308/trainer/checkpoint-736/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-07-s43-20260826T104017Z-bdd2c308/trainer/checkpoint-1576/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-topic-domain-07-s43-20260826T104017Z-bdd2c308/trainer/checkpoint-1576/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-07-s43-20260826T104017Z-bdd2c308/trainer/checkpoint-1576/special_tokens_map.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": "<mask>",
|
| 6 |
+
"pad_token": "<pad>",
|
| 7 |
+
"sep_token": "</s>",
|
| 8 |
+
"unk_token": "<unk>"
|
| 9 |
+
}
|
ood-vifactcheck-phobert-topic-domain-07-s43-20260826T104017Z-bdd2c308/trainer/checkpoint-1576/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"64000": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 256,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "PhobertTokenizer",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-phobert-topic-domain-07-s43-20260826T104017Z-bdd2c308/trainer/checkpoint-1576/trainer_state.json
ADDED
|
@@ -0,0 +1,572 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1576,
|
| 3 |
+
"best_metric": 0.8697670712510894,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-topic-domain-07-s43-20260826T104017Z-bdd2c308/trainer/checkpoint-1576",
|
| 5 |
+
"epoch": 8.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1576,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.12706480304955528,
|
| 14 |
+
"grad_norm": 7.840640068054199,
|
| 15 |
+
"learning_rate": 1.218274111675127e-06,
|
| 16 |
+
"loss": 4.4391,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.25412960609911056,
|
| 21 |
+
"grad_norm": 9.192355155944824,
|
| 22 |
+
"learning_rate": 2.487309644670051e-06,
|
| 23 |
+
"loss": 4.4269,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.3811944091486658,
|
| 28 |
+
"grad_norm": 6.094736099243164,
|
| 29 |
+
"learning_rate": 3.756345177664975e-06,
|
| 30 |
+
"loss": 4.4098,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5082592121982211,
|
| 35 |
+
"grad_norm": 6.114234447479248,
|
| 36 |
+
"learning_rate": 5.025380710659899e-06,
|
| 37 |
+
"loss": 4.3733,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6353240152477764,
|
| 42 |
+
"grad_norm": 7.231252670288086,
|
| 43 |
+
"learning_rate": 6.2944162436548225e-06,
|
| 44 |
+
"loss": 4.3516,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.7623888182973316,
|
| 49 |
+
"grad_norm": 7.103808879852295,
|
| 50 |
+
"learning_rate": 7.563451776649747e-06,
|
| 51 |
+
"loss": 4.308,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.8894536213468869,
|
| 56 |
+
"grad_norm": 9.182961463928223,
|
| 57 |
+
"learning_rate": 8.832487309644671e-06,
|
| 58 |
+
"loss": 4.1646,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.5201793721973094,
|
| 64 |
+
"eval_loss": 0.9736663699150085,
|
| 65 |
+
"eval_macro_f1": 0.5035790059404246,
|
| 66 |
+
"eval_runtime": 0.4876,
|
| 67 |
+
"eval_samples_per_second": 1372.164,
|
| 68 |
+
"eval_steps_per_second": 86.145,
|
| 69 |
+
"eval_weighted_f1": 0.5050535446503744,
|
| 70 |
+
"step": 197
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0152477763659467,
|
| 74 |
+
"grad_norm": 14.830670356750488,
|
| 75 |
+
"learning_rate": 1.0101522842639596e-05,
|
| 76 |
+
"loss": 3.9881,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.1423125794155018,
|
| 81 |
+
"grad_norm": 30.27711296081543,
|
| 82 |
+
"learning_rate": 1.1370558375634519e-05,
|
| 83 |
+
"loss": 3.6781,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.2693773824650572,
|
| 88 |
+
"grad_norm": 25.726818084716797,
|
| 89 |
+
"learning_rate": 1.2639593908629442e-05,
|
| 90 |
+
"loss": 3.5262,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.3964421855146125,
|
| 95 |
+
"grad_norm": 31.014514923095703,
|
| 96 |
+
"learning_rate": 1.3908629441624366e-05,
|
| 97 |
+
"loss": 3.0567,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.5235069885641677,
|
| 102 |
+
"grad_norm": 55.37285232543945,
|
| 103 |
+
"learning_rate": 1.5177664974619291e-05,
|
| 104 |
+
"loss": 2.7728,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.650571791613723,
|
| 109 |
+
"grad_norm": 30.21661376953125,
|
| 110 |
+
"learning_rate": 1.6446700507614214e-05,
|
| 111 |
+
"loss": 2.2535,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.7776365946632784,
|
| 116 |
+
"grad_norm": 31.826101303100586,
|
| 117 |
+
"learning_rate": 1.771573604060914e-05,
|
| 118 |
+
"loss": 2.3132,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 1.9047013977128335,
|
| 123 |
+
"grad_norm": 32.36261749267578,
|
| 124 |
+
"learning_rate": 1.8984771573604063e-05,
|
| 125 |
+
"loss": 2.328,
|
| 126 |
+
"step": 375
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"epoch": 2.0,
|
| 130 |
+
"eval_accuracy": 0.828101644245142,
|
| 131 |
+
"eval_loss": 0.43558773398399353,
|
| 132 |
+
"eval_macro_f1": 0.827129523433047,
|
| 133 |
+
"eval_runtime": 0.4341,
|
| 134 |
+
"eval_samples_per_second": 1541.073,
|
| 135 |
+
"eval_steps_per_second": 96.749,
|
| 136 |
+
"eval_weighted_f1": 0.8274341669244806,
|
| 137 |
+
"step": 394
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.0304955527318933,
|
| 141 |
+
"grad_norm": 26.921672821044922,
|
| 142 |
+
"learning_rate": 1.997179921037789e-05,
|
| 143 |
+
"loss": 2.0015,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.1575603557814484,
|
| 148 |
+
"grad_norm": 49.792816162109375,
|
| 149 |
+
"learning_rate": 1.9830795262267346e-05,
|
| 150 |
+
"loss": 1.7676,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.2846251588310036,
|
| 155 |
+
"grad_norm": 15.953386306762695,
|
| 156 |
+
"learning_rate": 1.96897913141568e-05,
|
| 157 |
+
"loss": 1.6312,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.411689961880559,
|
| 162 |
+
"grad_norm": 39.05951690673828,
|
| 163 |
+
"learning_rate": 1.954878736604625e-05,
|
| 164 |
+
"loss": 1.5709,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.5387547649301143,
|
| 169 |
+
"grad_norm": 20.817716598510742,
|
| 170 |
+
"learning_rate": 1.9407783417935703e-05,
|
| 171 |
+
"loss": 1.479,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.6658195679796695,
|
| 176 |
+
"grad_norm": 44.90846633911133,
|
| 177 |
+
"learning_rate": 1.9266779469825156e-05,
|
| 178 |
+
"loss": 1.5412,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 2.792884371029225,
|
| 183 |
+
"grad_norm": 27.34243392944336,
|
| 184 |
+
"learning_rate": 1.912577552171461e-05,
|
| 185 |
+
"loss": 1.6144,
|
| 186 |
+
"step": 550
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 2.91994917407878,
|
| 190 |
+
"grad_norm": 36.2662467956543,
|
| 191 |
+
"learning_rate": 1.8984771573604063e-05,
|
| 192 |
+
"loss": 1.5522,
|
| 193 |
+
"step": 575
|
| 194 |
+
},
|
| 195 |
+
{
|
| 196 |
+
"epoch": 3.0,
|
| 197 |
+
"eval_accuracy": 0.8565022421524664,
|
| 198 |
+
"eval_loss": 0.4360232949256897,
|
| 199 |
+
"eval_macro_f1": 0.8571425408167445,
|
| 200 |
+
"eval_runtime": 0.4388,
|
| 201 |
+
"eval_samples_per_second": 1524.756,
|
| 202 |
+
"eval_steps_per_second": 95.725,
|
| 203 |
+
"eval_weighted_f1": 0.8572613460958276,
|
| 204 |
+
"step": 591
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.04574332909784,
|
| 208 |
+
"grad_norm": 93.00135040283203,
|
| 209 |
+
"learning_rate": 1.8843767625493517e-05,
|
| 210 |
+
"loss": 1.3073,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.172808132147395,
|
| 215 |
+
"grad_norm": 30.055824279785156,
|
| 216 |
+
"learning_rate": 1.870276367738297e-05,
|
| 217 |
+
"loss": 1.0737,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.2998729351969507,
|
| 222 |
+
"grad_norm": 42.56252670288086,
|
| 223 |
+
"learning_rate": 1.856175972927242e-05,
|
| 224 |
+
"loss": 1.0309,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.426937738246506,
|
| 229 |
+
"grad_norm": 33.56625747680664,
|
| 230 |
+
"learning_rate": 1.8420755781161873e-05,
|
| 231 |
+
"loss": 1.0478,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.554002541296061,
|
| 236 |
+
"grad_norm": 32.38334274291992,
|
| 237 |
+
"learning_rate": 1.8279751833051327e-05,
|
| 238 |
+
"loss": 1.0917,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.681067344345616,
|
| 243 |
+
"grad_norm": 51.50391387939453,
|
| 244 |
+
"learning_rate": 1.813874788494078e-05,
|
| 245 |
+
"loss": 0.9555,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 3.8081321473951717,
|
| 250 |
+
"grad_norm": 21.810012817382812,
|
| 251 |
+
"learning_rate": 1.7997743936830234e-05,
|
| 252 |
+
"loss": 1.0255,
|
| 253 |
+
"step": 750
|
| 254 |
+
},
|
| 255 |
+
{
|
| 256 |
+
"epoch": 3.935196950444727,
|
| 257 |
+
"grad_norm": 71.86371612548828,
|
| 258 |
+
"learning_rate": 1.7856739988719684e-05,
|
| 259 |
+
"loss": 1.0445,
|
| 260 |
+
"step": 775
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 4.0,
|
| 264 |
+
"eval_accuracy": 0.8490284005979073,
|
| 265 |
+
"eval_loss": 0.46472203731536865,
|
| 266 |
+
"eval_macro_f1": 0.847540962102534,
|
| 267 |
+
"eval_runtime": 0.4337,
|
| 268 |
+
"eval_samples_per_second": 1542.48,
|
| 269 |
+
"eval_steps_per_second": 96.837,
|
| 270 |
+
"eval_weighted_f1": 0.8478064610134746,
|
| 271 |
+
"step": 788
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.060991105463787,
|
| 275 |
+
"grad_norm": 60.584476470947266,
|
| 276 |
+
"learning_rate": 1.771573604060914e-05,
|
| 277 |
+
"loss": 0.9837,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.188055908513342,
|
| 282 |
+
"grad_norm": 57.55632781982422,
|
| 283 |
+
"learning_rate": 1.7574732092498594e-05,
|
| 284 |
+
"loss": 0.6709,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.315120711562897,
|
| 289 |
+
"grad_norm": 14.654173851013184,
|
| 290 |
+
"learning_rate": 1.7433728144388044e-05,
|
| 291 |
+
"loss": 0.7421,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.4421855146124525,
|
| 296 |
+
"grad_norm": 17.030200958251953,
|
| 297 |
+
"learning_rate": 1.7292724196277497e-05,
|
| 298 |
+
"loss": 0.6059,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.569250317662007,
|
| 303 |
+
"grad_norm": 32.25056076049805,
|
| 304 |
+
"learning_rate": 1.715172024816695e-05,
|
| 305 |
+
"loss": 0.7991,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 4.696315120711563,
|
| 310 |
+
"grad_norm": 75.96150970458984,
|
| 311 |
+
"learning_rate": 1.7010716300056404e-05,
|
| 312 |
+
"loss": 0.564,
|
| 313 |
+
"step": 925
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 4.823379923761118,
|
| 317 |
+
"grad_norm": 29.156265258789062,
|
| 318 |
+
"learning_rate": 1.6869712351945854e-05,
|
| 319 |
+
"loss": 0.9521,
|
| 320 |
+
"step": 950
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 4.950444726810673,
|
| 324 |
+
"grad_norm": 6.4156718254089355,
|
| 325 |
+
"learning_rate": 1.672870840383531e-05,
|
| 326 |
+
"loss": 0.7987,
|
| 327 |
+
"step": 975
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 5.0,
|
| 331 |
+
"eval_accuracy": 0.8430493273542601,
|
| 332 |
+
"eval_loss": 0.5255759358406067,
|
| 333 |
+
"eval_macro_f1": 0.8407267445174892,
|
| 334 |
+
"eval_runtime": 0.4664,
|
| 335 |
+
"eval_samples_per_second": 1434.496,
|
| 336 |
+
"eval_steps_per_second": 90.058,
|
| 337 |
+
"eval_weighted_f1": 0.8410549653638321,
|
| 338 |
+
"step": 985
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.076238881829733,
|
| 342 |
+
"grad_norm": 46.026939392089844,
|
| 343 |
+
"learning_rate": 1.6587704455724764e-05,
|
| 344 |
+
"loss": 0.7381,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.203303684879288,
|
| 349 |
+
"grad_norm": 76.59004974365234,
|
| 350 |
+
"learning_rate": 1.6446700507614214e-05,
|
| 351 |
+
"loss": 0.3576,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.330368487928844,
|
| 356 |
+
"grad_norm": 37.609310150146484,
|
| 357 |
+
"learning_rate": 1.6305696559503667e-05,
|
| 358 |
+
"loss": 0.6675,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.457433290978399,
|
| 363 |
+
"grad_norm": 60.973873138427734,
|
| 364 |
+
"learning_rate": 1.616469261139312e-05,
|
| 365 |
+
"loss": 0.59,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 5.584498094027954,
|
| 370 |
+
"grad_norm": 50.5499267578125,
|
| 371 |
+
"learning_rate": 1.6023688663282574e-05,
|
| 372 |
+
"loss": 0.5045,
|
| 373 |
+
"step": 1100
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 5.71156289707751,
|
| 377 |
+
"grad_norm": 3.758920431137085,
|
| 378 |
+
"learning_rate": 1.5882684715172024e-05,
|
| 379 |
+
"loss": 0.5074,
|
| 380 |
+
"step": 1125
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 5.838627700127065,
|
| 384 |
+
"grad_norm": 0.5275550484657288,
|
| 385 |
+
"learning_rate": 1.5741680767061477e-05,
|
| 386 |
+
"loss": 0.6101,
|
| 387 |
+
"step": 1150
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 5.96569250317662,
|
| 391 |
+
"grad_norm": 3.2209837436676025,
|
| 392 |
+
"learning_rate": 1.5600676818950934e-05,
|
| 393 |
+
"loss": 0.6231,
|
| 394 |
+
"step": 1175
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 6.0,
|
| 398 |
+
"eval_accuracy": 0.8579970104633782,
|
| 399 |
+
"eval_loss": 0.5370838046073914,
|
| 400 |
+
"eval_macro_f1": 0.8575227094588144,
|
| 401 |
+
"eval_runtime": 0.4297,
|
| 402 |
+
"eval_samples_per_second": 1556.978,
|
| 403 |
+
"eval_steps_per_second": 97.748,
|
| 404 |
+
"eval_weighted_f1": 0.857809035539433,
|
| 405 |
+
"step": 1182
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.09148665819568,
|
| 409 |
+
"grad_norm": 96.30651092529297,
|
| 410 |
+
"learning_rate": 1.5459672870840384e-05,
|
| 411 |
+
"loss": 0.394,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.2185514612452355,
|
| 416 |
+
"grad_norm": 0.37381041049957275,
|
| 417 |
+
"learning_rate": 1.5318668922729838e-05,
|
| 418 |
+
"loss": 0.3822,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.34561626429479,
|
| 423 |
+
"grad_norm": 3.3887436389923096,
|
| 424 |
+
"learning_rate": 1.5177664974619291e-05,
|
| 425 |
+
"loss": 0.2897,
|
| 426 |
+
"step": 1250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 6.472681067344346,
|
| 430 |
+
"grad_norm": 13.318480491638184,
|
| 431 |
+
"learning_rate": 1.5036661026508743e-05,
|
| 432 |
+
"loss": 0.3784,
|
| 433 |
+
"step": 1275
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 6.599745870393901,
|
| 437 |
+
"grad_norm": 7.0607380867004395,
|
| 438 |
+
"learning_rate": 1.4895657078398196e-05,
|
| 439 |
+
"loss": 0.6321,
|
| 440 |
+
"step": 1300
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 6.726810673443456,
|
| 444 |
+
"grad_norm": 29.00809097290039,
|
| 445 |
+
"learning_rate": 1.4754653130287648e-05,
|
| 446 |
+
"loss": 0.5246,
|
| 447 |
+
"step": 1325
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 6.853875476493012,
|
| 451 |
+
"grad_norm": 0.35460567474365234,
|
| 452 |
+
"learning_rate": 1.4613649182177103e-05,
|
| 453 |
+
"loss": 0.5236,
|
| 454 |
+
"step": 1350
|
| 455 |
+
},
|
| 456 |
+
{
|
| 457 |
+
"epoch": 6.980940279542566,
|
| 458 |
+
"grad_norm": 77.04814147949219,
|
| 459 |
+
"learning_rate": 1.4472645234066556e-05,
|
| 460 |
+
"loss": 0.4391,
|
| 461 |
+
"step": 1375
|
| 462 |
+
},
|
| 463 |
+
{
|
| 464 |
+
"epoch": 7.0,
|
| 465 |
+
"eval_accuracy": 0.8699551569506726,
|
| 466 |
+
"eval_loss": 0.5865257382392883,
|
| 467 |
+
"eval_macro_f1": 0.8695884179961292,
|
| 468 |
+
"eval_runtime": 0.4319,
|
| 469 |
+
"eval_samples_per_second": 1549.141,
|
| 470 |
+
"eval_steps_per_second": 97.256,
|
| 471 |
+
"eval_weighted_f1": 0.8697169006165271,
|
| 472 |
+
"step": 1379
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 7.106734434561626,
|
| 476 |
+
"grad_norm": 119.25450134277344,
|
| 477 |
+
"learning_rate": 1.4331641285956008e-05,
|
| 478 |
+
"loss": 0.2171,
|
| 479 |
+
"step": 1400
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 7.233799237611182,
|
| 483 |
+
"grad_norm": 0.3869434595108032,
|
| 484 |
+
"learning_rate": 1.4190637337845461e-05,
|
| 485 |
+
"loss": 0.3106,
|
| 486 |
+
"step": 1425
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 7.360864040660737,
|
| 490 |
+
"grad_norm": 0.2600739300251007,
|
| 491 |
+
"learning_rate": 1.4049633389734913e-05,
|
| 492 |
+
"loss": 0.2511,
|
| 493 |
+
"step": 1450
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 7.487928843710292,
|
| 497 |
+
"grad_norm": 3.6103007793426514,
|
| 498 |
+
"learning_rate": 1.3908629441624366e-05,
|
| 499 |
+
"loss": 0.3216,
|
| 500 |
+
"step": 1475
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 7.614993646759848,
|
| 504 |
+
"grad_norm": 88.88577270507812,
|
| 505 |
+
"learning_rate": 1.3767625493513818e-05,
|
| 506 |
+
"loss": 0.2862,
|
| 507 |
+
"step": 1500
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 7.742058449809403,
|
| 511 |
+
"grad_norm": 0.307084321975708,
|
| 512 |
+
"learning_rate": 1.3626621545403271e-05,
|
| 513 |
+
"loss": 0.1929,
|
| 514 |
+
"step": 1525
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 7.869123252858958,
|
| 518 |
+
"grad_norm": 194.25897216796875,
|
| 519 |
+
"learning_rate": 1.3485617597292726e-05,
|
| 520 |
+
"loss": 0.3958,
|
| 521 |
+
"step": 1550
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 7.9961880559085134,
|
| 525 |
+
"grad_norm": 39.095245361328125,
|
| 526 |
+
"learning_rate": 1.3344613649182178e-05,
|
| 527 |
+
"loss": 0.3975,
|
| 528 |
+
"step": 1575
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 8.0,
|
| 532 |
+
"eval_accuracy": 0.8699551569506726,
|
| 533 |
+
"eval_loss": 0.6406896114349365,
|
| 534 |
+
"eval_macro_f1": 0.8697670712510894,
|
| 535 |
+
"eval_runtime": 0.4679,
|
| 536 |
+
"eval_samples_per_second": 1429.935,
|
| 537 |
+
"eval_steps_per_second": 89.772,
|
| 538 |
+
"eval_weighted_f1": 0.8699278771227295,
|
| 539 |
+
"step": 1576
|
| 540 |
+
}
|
| 541 |
+
],
|
| 542 |
+
"logging_steps": 25,
|
| 543 |
+
"max_steps": 3940,
|
| 544 |
+
"num_input_tokens_seen": 0,
|
| 545 |
+
"num_train_epochs": 20,
|
| 546 |
+
"save_steps": 500,
|
| 547 |
+
"stateful_callbacks": {
|
| 548 |
+
"EarlyStoppingCallback": {
|
| 549 |
+
"args": {
|
| 550 |
+
"early_stopping_patience": 3,
|
| 551 |
+
"early_stopping_threshold": 0.0
|
| 552 |
+
},
|
| 553 |
+
"attributes": {
|
| 554 |
+
"early_stopping_patience_counter": 0
|
| 555 |
+
}
|
| 556 |
+
},
|
| 557 |
+
"TrainerControl": {
|
| 558 |
+
"args": {
|
| 559 |
+
"should_epoch_stop": false,
|
| 560 |
+
"should_evaluate": false,
|
| 561 |
+
"should_log": false,
|
| 562 |
+
"should_save": true,
|
| 563 |
+
"should_training_stop": false
|
| 564 |
+
},
|
| 565 |
+
"attributes": {}
|
| 566 |
+
}
|
| 567 |
+
},
|
| 568 |
+
"total_flos": 0.0,
|
| 569 |
+
"train_batch_size": 8,
|
| 570 |
+
"trial_name": null,
|
| 571 |
+
"trial_params": null
|
| 572 |
+
}
|
ood-vifactcheck-phobert-topic-domain-07-s43-20260826T104017Z-bdd2c308/trainer/checkpoint-1576/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-07-s43-20260826T104017Z-bdd2c308/trainer/checkpoint-2167/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-topic-domain-07-s43-20260826T104017Z-bdd2c308/trainer/checkpoint-2167/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-07-s43-20260826T104017Z-bdd2c308/trainer/checkpoint-2167/special_tokens_map.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": "<mask>",
|
| 6 |
+
"pad_token": "<pad>",
|
| 7 |
+
"sep_token": "</s>",
|
| 8 |
+
"unk_token": "<unk>"
|
| 9 |
+
}
|
ood-vifactcheck-phobert-topic-domain-07-s43-20260826T104017Z-bdd2c308/trainer/checkpoint-2167/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"64000": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 256,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "PhobertTokenizer",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-phobert-topic-domain-07-s43-20260826T104017Z-bdd2c308/trainer/checkpoint-2167/trainer_state.json
ADDED
|
@@ -0,0 +1,766 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1576,
|
| 3 |
+
"best_metric": 0.8697670712510894,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-topic-domain-07-s43-20260826T104017Z-bdd2c308/trainer/checkpoint-1576",
|
| 5 |
+
"epoch": 11.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 2167,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.12706480304955528,
|
| 14 |
+
"grad_norm": 7.840640068054199,
|
| 15 |
+
"learning_rate": 1.218274111675127e-06,
|
| 16 |
+
"loss": 4.4391,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.25412960609911056,
|
| 21 |
+
"grad_norm": 9.192355155944824,
|
| 22 |
+
"learning_rate": 2.487309644670051e-06,
|
| 23 |
+
"loss": 4.4269,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.3811944091486658,
|
| 28 |
+
"grad_norm": 6.094736099243164,
|
| 29 |
+
"learning_rate": 3.756345177664975e-06,
|
| 30 |
+
"loss": 4.4098,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5082592121982211,
|
| 35 |
+
"grad_norm": 6.114234447479248,
|
| 36 |
+
"learning_rate": 5.025380710659899e-06,
|
| 37 |
+
"loss": 4.3733,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6353240152477764,
|
| 42 |
+
"grad_norm": 7.231252670288086,
|
| 43 |
+
"learning_rate": 6.2944162436548225e-06,
|
| 44 |
+
"loss": 4.3516,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.7623888182973316,
|
| 49 |
+
"grad_norm": 7.103808879852295,
|
| 50 |
+
"learning_rate": 7.563451776649747e-06,
|
| 51 |
+
"loss": 4.308,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.8894536213468869,
|
| 56 |
+
"grad_norm": 9.182961463928223,
|
| 57 |
+
"learning_rate": 8.832487309644671e-06,
|
| 58 |
+
"loss": 4.1646,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.5201793721973094,
|
| 64 |
+
"eval_loss": 0.9736663699150085,
|
| 65 |
+
"eval_macro_f1": 0.5035790059404246,
|
| 66 |
+
"eval_runtime": 0.4876,
|
| 67 |
+
"eval_samples_per_second": 1372.164,
|
| 68 |
+
"eval_steps_per_second": 86.145,
|
| 69 |
+
"eval_weighted_f1": 0.5050535446503744,
|
| 70 |
+
"step": 197
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0152477763659467,
|
| 74 |
+
"grad_norm": 14.830670356750488,
|
| 75 |
+
"learning_rate": 1.0101522842639596e-05,
|
| 76 |
+
"loss": 3.9881,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.1423125794155018,
|
| 81 |
+
"grad_norm": 30.27711296081543,
|
| 82 |
+
"learning_rate": 1.1370558375634519e-05,
|
| 83 |
+
"loss": 3.6781,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.2693773824650572,
|
| 88 |
+
"grad_norm": 25.726818084716797,
|
| 89 |
+
"learning_rate": 1.2639593908629442e-05,
|
| 90 |
+
"loss": 3.5262,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.3964421855146125,
|
| 95 |
+
"grad_norm": 31.014514923095703,
|
| 96 |
+
"learning_rate": 1.3908629441624366e-05,
|
| 97 |
+
"loss": 3.0567,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.5235069885641677,
|
| 102 |
+
"grad_norm": 55.37285232543945,
|
| 103 |
+
"learning_rate": 1.5177664974619291e-05,
|
| 104 |
+
"loss": 2.7728,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.650571791613723,
|
| 109 |
+
"grad_norm": 30.21661376953125,
|
| 110 |
+
"learning_rate": 1.6446700507614214e-05,
|
| 111 |
+
"loss": 2.2535,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.7776365946632784,
|
| 116 |
+
"grad_norm": 31.826101303100586,
|
| 117 |
+
"learning_rate": 1.771573604060914e-05,
|
| 118 |
+
"loss": 2.3132,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 1.9047013977128335,
|
| 123 |
+
"grad_norm": 32.36261749267578,
|
| 124 |
+
"learning_rate": 1.8984771573604063e-05,
|
| 125 |
+
"loss": 2.328,
|
| 126 |
+
"step": 375
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"epoch": 2.0,
|
| 130 |
+
"eval_accuracy": 0.828101644245142,
|
| 131 |
+
"eval_loss": 0.43558773398399353,
|
| 132 |
+
"eval_macro_f1": 0.827129523433047,
|
| 133 |
+
"eval_runtime": 0.4341,
|
| 134 |
+
"eval_samples_per_second": 1541.073,
|
| 135 |
+
"eval_steps_per_second": 96.749,
|
| 136 |
+
"eval_weighted_f1": 0.8274341669244806,
|
| 137 |
+
"step": 394
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.0304955527318933,
|
| 141 |
+
"grad_norm": 26.921672821044922,
|
| 142 |
+
"learning_rate": 1.997179921037789e-05,
|
| 143 |
+
"loss": 2.0015,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.1575603557814484,
|
| 148 |
+
"grad_norm": 49.792816162109375,
|
| 149 |
+
"learning_rate": 1.9830795262267346e-05,
|
| 150 |
+
"loss": 1.7676,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.2846251588310036,
|
| 155 |
+
"grad_norm": 15.953386306762695,
|
| 156 |
+
"learning_rate": 1.96897913141568e-05,
|
| 157 |
+
"loss": 1.6312,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.411689961880559,
|
| 162 |
+
"grad_norm": 39.05951690673828,
|
| 163 |
+
"learning_rate": 1.954878736604625e-05,
|
| 164 |
+
"loss": 1.5709,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.5387547649301143,
|
| 169 |
+
"grad_norm": 20.817716598510742,
|
| 170 |
+
"learning_rate": 1.9407783417935703e-05,
|
| 171 |
+
"loss": 1.479,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.6658195679796695,
|
| 176 |
+
"grad_norm": 44.90846633911133,
|
| 177 |
+
"learning_rate": 1.9266779469825156e-05,
|
| 178 |
+
"loss": 1.5412,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 2.792884371029225,
|
| 183 |
+
"grad_norm": 27.34243392944336,
|
| 184 |
+
"learning_rate": 1.912577552171461e-05,
|
| 185 |
+
"loss": 1.6144,
|
| 186 |
+
"step": 550
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 2.91994917407878,
|
| 190 |
+
"grad_norm": 36.2662467956543,
|
| 191 |
+
"learning_rate": 1.8984771573604063e-05,
|
| 192 |
+
"loss": 1.5522,
|
| 193 |
+
"step": 575
|
| 194 |
+
},
|
| 195 |
+
{
|
| 196 |
+
"epoch": 3.0,
|
| 197 |
+
"eval_accuracy": 0.8565022421524664,
|
| 198 |
+
"eval_loss": 0.4360232949256897,
|
| 199 |
+
"eval_macro_f1": 0.8571425408167445,
|
| 200 |
+
"eval_runtime": 0.4388,
|
| 201 |
+
"eval_samples_per_second": 1524.756,
|
| 202 |
+
"eval_steps_per_second": 95.725,
|
| 203 |
+
"eval_weighted_f1": 0.8572613460958276,
|
| 204 |
+
"step": 591
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.04574332909784,
|
| 208 |
+
"grad_norm": 93.00135040283203,
|
| 209 |
+
"learning_rate": 1.8843767625493517e-05,
|
| 210 |
+
"loss": 1.3073,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.172808132147395,
|
| 215 |
+
"grad_norm": 30.055824279785156,
|
| 216 |
+
"learning_rate": 1.870276367738297e-05,
|
| 217 |
+
"loss": 1.0737,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.2998729351969507,
|
| 222 |
+
"grad_norm": 42.56252670288086,
|
| 223 |
+
"learning_rate": 1.856175972927242e-05,
|
| 224 |
+
"loss": 1.0309,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.426937738246506,
|
| 229 |
+
"grad_norm": 33.56625747680664,
|
| 230 |
+
"learning_rate": 1.8420755781161873e-05,
|
| 231 |
+
"loss": 1.0478,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.554002541296061,
|
| 236 |
+
"grad_norm": 32.38334274291992,
|
| 237 |
+
"learning_rate": 1.8279751833051327e-05,
|
| 238 |
+
"loss": 1.0917,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.681067344345616,
|
| 243 |
+
"grad_norm": 51.50391387939453,
|
| 244 |
+
"learning_rate": 1.813874788494078e-05,
|
| 245 |
+
"loss": 0.9555,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 3.8081321473951717,
|
| 250 |
+
"grad_norm": 21.810012817382812,
|
| 251 |
+
"learning_rate": 1.7997743936830234e-05,
|
| 252 |
+
"loss": 1.0255,
|
| 253 |
+
"step": 750
|
| 254 |
+
},
|
| 255 |
+
{
|
| 256 |
+
"epoch": 3.935196950444727,
|
| 257 |
+
"grad_norm": 71.86371612548828,
|
| 258 |
+
"learning_rate": 1.7856739988719684e-05,
|
| 259 |
+
"loss": 1.0445,
|
| 260 |
+
"step": 775
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 4.0,
|
| 264 |
+
"eval_accuracy": 0.8490284005979073,
|
| 265 |
+
"eval_loss": 0.46472203731536865,
|
| 266 |
+
"eval_macro_f1": 0.847540962102534,
|
| 267 |
+
"eval_runtime": 0.4337,
|
| 268 |
+
"eval_samples_per_second": 1542.48,
|
| 269 |
+
"eval_steps_per_second": 96.837,
|
| 270 |
+
"eval_weighted_f1": 0.8478064610134746,
|
| 271 |
+
"step": 788
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.060991105463787,
|
| 275 |
+
"grad_norm": 60.584476470947266,
|
| 276 |
+
"learning_rate": 1.771573604060914e-05,
|
| 277 |
+
"loss": 0.9837,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.188055908513342,
|
| 282 |
+
"grad_norm": 57.55632781982422,
|
| 283 |
+
"learning_rate": 1.7574732092498594e-05,
|
| 284 |
+
"loss": 0.6709,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.315120711562897,
|
| 289 |
+
"grad_norm": 14.654173851013184,
|
| 290 |
+
"learning_rate": 1.7433728144388044e-05,
|
| 291 |
+
"loss": 0.7421,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.4421855146124525,
|
| 296 |
+
"grad_norm": 17.030200958251953,
|
| 297 |
+
"learning_rate": 1.7292724196277497e-05,
|
| 298 |
+
"loss": 0.6059,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.569250317662007,
|
| 303 |
+
"grad_norm": 32.25056076049805,
|
| 304 |
+
"learning_rate": 1.715172024816695e-05,
|
| 305 |
+
"loss": 0.7991,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 4.696315120711563,
|
| 310 |
+
"grad_norm": 75.96150970458984,
|
| 311 |
+
"learning_rate": 1.7010716300056404e-05,
|
| 312 |
+
"loss": 0.564,
|
| 313 |
+
"step": 925
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 4.823379923761118,
|
| 317 |
+
"grad_norm": 29.156265258789062,
|
| 318 |
+
"learning_rate": 1.6869712351945854e-05,
|
| 319 |
+
"loss": 0.9521,
|
| 320 |
+
"step": 950
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 4.950444726810673,
|
| 324 |
+
"grad_norm": 6.4156718254089355,
|
| 325 |
+
"learning_rate": 1.672870840383531e-05,
|
| 326 |
+
"loss": 0.7987,
|
| 327 |
+
"step": 975
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 5.0,
|
| 331 |
+
"eval_accuracy": 0.8430493273542601,
|
| 332 |
+
"eval_loss": 0.5255759358406067,
|
| 333 |
+
"eval_macro_f1": 0.8407267445174892,
|
| 334 |
+
"eval_runtime": 0.4664,
|
| 335 |
+
"eval_samples_per_second": 1434.496,
|
| 336 |
+
"eval_steps_per_second": 90.058,
|
| 337 |
+
"eval_weighted_f1": 0.8410549653638321,
|
| 338 |
+
"step": 985
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.076238881829733,
|
| 342 |
+
"grad_norm": 46.026939392089844,
|
| 343 |
+
"learning_rate": 1.6587704455724764e-05,
|
| 344 |
+
"loss": 0.7381,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.203303684879288,
|
| 349 |
+
"grad_norm": 76.59004974365234,
|
| 350 |
+
"learning_rate": 1.6446700507614214e-05,
|
| 351 |
+
"loss": 0.3576,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.330368487928844,
|
| 356 |
+
"grad_norm": 37.609310150146484,
|
| 357 |
+
"learning_rate": 1.6305696559503667e-05,
|
| 358 |
+
"loss": 0.6675,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.457433290978399,
|
| 363 |
+
"grad_norm": 60.973873138427734,
|
| 364 |
+
"learning_rate": 1.616469261139312e-05,
|
| 365 |
+
"loss": 0.59,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 5.584498094027954,
|
| 370 |
+
"grad_norm": 50.5499267578125,
|
| 371 |
+
"learning_rate": 1.6023688663282574e-05,
|
| 372 |
+
"loss": 0.5045,
|
| 373 |
+
"step": 1100
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 5.71156289707751,
|
| 377 |
+
"grad_norm": 3.758920431137085,
|
| 378 |
+
"learning_rate": 1.5882684715172024e-05,
|
| 379 |
+
"loss": 0.5074,
|
| 380 |
+
"step": 1125
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 5.838627700127065,
|
| 384 |
+
"grad_norm": 0.5275550484657288,
|
| 385 |
+
"learning_rate": 1.5741680767061477e-05,
|
| 386 |
+
"loss": 0.6101,
|
| 387 |
+
"step": 1150
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 5.96569250317662,
|
| 391 |
+
"grad_norm": 3.2209837436676025,
|
| 392 |
+
"learning_rate": 1.5600676818950934e-05,
|
| 393 |
+
"loss": 0.6231,
|
| 394 |
+
"step": 1175
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 6.0,
|
| 398 |
+
"eval_accuracy": 0.8579970104633782,
|
| 399 |
+
"eval_loss": 0.5370838046073914,
|
| 400 |
+
"eval_macro_f1": 0.8575227094588144,
|
| 401 |
+
"eval_runtime": 0.4297,
|
| 402 |
+
"eval_samples_per_second": 1556.978,
|
| 403 |
+
"eval_steps_per_second": 97.748,
|
| 404 |
+
"eval_weighted_f1": 0.857809035539433,
|
| 405 |
+
"step": 1182
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.09148665819568,
|
| 409 |
+
"grad_norm": 96.30651092529297,
|
| 410 |
+
"learning_rate": 1.5459672870840384e-05,
|
| 411 |
+
"loss": 0.394,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.2185514612452355,
|
| 416 |
+
"grad_norm": 0.37381041049957275,
|
| 417 |
+
"learning_rate": 1.5318668922729838e-05,
|
| 418 |
+
"loss": 0.3822,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.34561626429479,
|
| 423 |
+
"grad_norm": 3.3887436389923096,
|
| 424 |
+
"learning_rate": 1.5177664974619291e-05,
|
| 425 |
+
"loss": 0.2897,
|
| 426 |
+
"step": 1250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 6.472681067344346,
|
| 430 |
+
"grad_norm": 13.318480491638184,
|
| 431 |
+
"learning_rate": 1.5036661026508743e-05,
|
| 432 |
+
"loss": 0.3784,
|
| 433 |
+
"step": 1275
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 6.599745870393901,
|
| 437 |
+
"grad_norm": 7.0607380867004395,
|
| 438 |
+
"learning_rate": 1.4895657078398196e-05,
|
| 439 |
+
"loss": 0.6321,
|
| 440 |
+
"step": 1300
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 6.726810673443456,
|
| 444 |
+
"grad_norm": 29.00809097290039,
|
| 445 |
+
"learning_rate": 1.4754653130287648e-05,
|
| 446 |
+
"loss": 0.5246,
|
| 447 |
+
"step": 1325
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 6.853875476493012,
|
| 451 |
+
"grad_norm": 0.35460567474365234,
|
| 452 |
+
"learning_rate": 1.4613649182177103e-05,
|
| 453 |
+
"loss": 0.5236,
|
| 454 |
+
"step": 1350
|
| 455 |
+
},
|
| 456 |
+
{
|
| 457 |
+
"epoch": 6.980940279542566,
|
| 458 |
+
"grad_norm": 77.04814147949219,
|
| 459 |
+
"learning_rate": 1.4472645234066556e-05,
|
| 460 |
+
"loss": 0.4391,
|
| 461 |
+
"step": 1375
|
| 462 |
+
},
|
| 463 |
+
{
|
| 464 |
+
"epoch": 7.0,
|
| 465 |
+
"eval_accuracy": 0.8699551569506726,
|
| 466 |
+
"eval_loss": 0.5865257382392883,
|
| 467 |
+
"eval_macro_f1": 0.8695884179961292,
|
| 468 |
+
"eval_runtime": 0.4319,
|
| 469 |
+
"eval_samples_per_second": 1549.141,
|
| 470 |
+
"eval_steps_per_second": 97.256,
|
| 471 |
+
"eval_weighted_f1": 0.8697169006165271,
|
| 472 |
+
"step": 1379
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 7.106734434561626,
|
| 476 |
+
"grad_norm": 119.25450134277344,
|
| 477 |
+
"learning_rate": 1.4331641285956008e-05,
|
| 478 |
+
"loss": 0.2171,
|
| 479 |
+
"step": 1400
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 7.233799237611182,
|
| 483 |
+
"grad_norm": 0.3869434595108032,
|
| 484 |
+
"learning_rate": 1.4190637337845461e-05,
|
| 485 |
+
"loss": 0.3106,
|
| 486 |
+
"step": 1425
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 7.360864040660737,
|
| 490 |
+
"grad_norm": 0.2600739300251007,
|
| 491 |
+
"learning_rate": 1.4049633389734913e-05,
|
| 492 |
+
"loss": 0.2511,
|
| 493 |
+
"step": 1450
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 7.487928843710292,
|
| 497 |
+
"grad_norm": 3.6103007793426514,
|
| 498 |
+
"learning_rate": 1.3908629441624366e-05,
|
| 499 |
+
"loss": 0.3216,
|
| 500 |
+
"step": 1475
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 7.614993646759848,
|
| 504 |
+
"grad_norm": 88.88577270507812,
|
| 505 |
+
"learning_rate": 1.3767625493513818e-05,
|
| 506 |
+
"loss": 0.2862,
|
| 507 |
+
"step": 1500
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 7.742058449809403,
|
| 511 |
+
"grad_norm": 0.307084321975708,
|
| 512 |
+
"learning_rate": 1.3626621545403271e-05,
|
| 513 |
+
"loss": 0.1929,
|
| 514 |
+
"step": 1525
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 7.869123252858958,
|
| 518 |
+
"grad_norm": 194.25897216796875,
|
| 519 |
+
"learning_rate": 1.3485617597292726e-05,
|
| 520 |
+
"loss": 0.3958,
|
| 521 |
+
"step": 1550
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 7.9961880559085134,
|
| 525 |
+
"grad_norm": 39.095245361328125,
|
| 526 |
+
"learning_rate": 1.3344613649182178e-05,
|
| 527 |
+
"loss": 0.3975,
|
| 528 |
+
"step": 1575
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 8.0,
|
| 532 |
+
"eval_accuracy": 0.8699551569506726,
|
| 533 |
+
"eval_loss": 0.6406896114349365,
|
| 534 |
+
"eval_macro_f1": 0.8697670712510894,
|
| 535 |
+
"eval_runtime": 0.4679,
|
| 536 |
+
"eval_samples_per_second": 1429.935,
|
| 537 |
+
"eval_steps_per_second": 89.772,
|
| 538 |
+
"eval_weighted_f1": 0.8699278771227295,
|
| 539 |
+
"step": 1576
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 8.121982210927573,
|
| 543 |
+
"grad_norm": 170.27928161621094,
|
| 544 |
+
"learning_rate": 1.3203609701071631e-05,
|
| 545 |
+
"loss": 0.1785,
|
| 546 |
+
"step": 1600
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 8.249047013977128,
|
| 550 |
+
"grad_norm": 0.37956324219703674,
|
| 551 |
+
"learning_rate": 1.3062605752961083e-05,
|
| 552 |
+
"loss": 0.1519,
|
| 553 |
+
"step": 1625
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 8.376111817026684,
|
| 557 |
+
"grad_norm": 5.647228240966797,
|
| 558 |
+
"learning_rate": 1.2921601804850537e-05,
|
| 559 |
+
"loss": 0.2054,
|
| 560 |
+
"step": 1650
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 8.503176620076239,
|
| 564 |
+
"grad_norm": 47.2303466796875,
|
| 565 |
+
"learning_rate": 1.278059785673999e-05,
|
| 566 |
+
"loss": 0.2222,
|
| 567 |
+
"step": 1675
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 8.630241423125794,
|
| 571 |
+
"grad_norm": 3.3748857975006104,
|
| 572 |
+
"learning_rate": 1.2639593908629442e-05,
|
| 573 |
+
"loss": 0.4165,
|
| 574 |
+
"step": 1700
|
| 575 |
+
},
|
| 576 |
+
{
|
| 577 |
+
"epoch": 8.757306226175348,
|
| 578 |
+
"grad_norm": 0.22848127782344818,
|
| 579 |
+
"learning_rate": 1.2498589960518897e-05,
|
| 580 |
+
"loss": 0.3987,
|
| 581 |
+
"step": 1725
|
| 582 |
+
},
|
| 583 |
+
{
|
| 584 |
+
"epoch": 8.884371029224905,
|
| 585 |
+
"grad_norm": 0.18938516080379486,
|
| 586 |
+
"learning_rate": 1.2357586012408348e-05,
|
| 587 |
+
"loss": 0.2883,
|
| 588 |
+
"step": 1750
|
| 589 |
+
},
|
| 590 |
+
{
|
| 591 |
+
"epoch": 9.0,
|
| 592 |
+
"eval_accuracy": 0.8654708520179372,
|
| 593 |
+
"eval_loss": 0.6702370643615723,
|
| 594 |
+
"eval_macro_f1": 0.8652422284834063,
|
| 595 |
+
"eval_runtime": 0.4665,
|
| 596 |
+
"eval_samples_per_second": 1433.967,
|
| 597 |
+
"eval_steps_per_second": 90.025,
|
| 598 |
+
"eval_weighted_f1": 0.8653970880583551,
|
| 599 |
+
"step": 1773
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 9.010165184243965,
|
| 603 |
+
"grad_norm": 40.12786865234375,
|
| 604 |
+
"learning_rate": 1.2216582064297802e-05,
|
| 605 |
+
"loss": 0.1492,
|
| 606 |
+
"step": 1775
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"epoch": 9.13722998729352,
|
| 610 |
+
"grad_norm": 0.2552495300769806,
|
| 611 |
+
"learning_rate": 1.2075578116187253e-05,
|
| 612 |
+
"loss": 0.1198,
|
| 613 |
+
"step": 1800
|
| 614 |
+
},
|
| 615 |
+
{
|
| 616 |
+
"epoch": 9.264294790343074,
|
| 617 |
+
"grad_norm": 0.16413116455078125,
|
| 618 |
+
"learning_rate": 1.1934574168076707e-05,
|
| 619 |
+
"loss": 0.1545,
|
| 620 |
+
"step": 1825
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 9.39135959339263,
|
| 624 |
+
"grad_norm": 87.56383514404297,
|
| 625 |
+
"learning_rate": 1.179357021996616e-05,
|
| 626 |
+
"loss": 0.1214,
|
| 627 |
+
"step": 1850
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 9.518424396442185,
|
| 631 |
+
"grad_norm": 39.00825119018555,
|
| 632 |
+
"learning_rate": 1.1652566271855612e-05,
|
| 633 |
+
"loss": 0.255,
|
| 634 |
+
"step": 1875
|
| 635 |
+
},
|
| 636 |
+
{
|
| 637 |
+
"epoch": 9.64548919949174,
|
| 638 |
+
"grad_norm": 19.036500930786133,
|
| 639 |
+
"learning_rate": 1.1511562323745067e-05,
|
| 640 |
+
"loss": 0.1561,
|
| 641 |
+
"step": 1900
|
| 642 |
+
},
|
| 643 |
+
{
|
| 644 |
+
"epoch": 9.772554002541296,
|
| 645 |
+
"grad_norm": 0.156470388174057,
|
| 646 |
+
"learning_rate": 1.1370558375634519e-05,
|
| 647 |
+
"loss": 0.1564,
|
| 648 |
+
"step": 1925
|
| 649 |
+
},
|
| 650 |
+
{
|
| 651 |
+
"epoch": 9.899618805590851,
|
| 652 |
+
"grad_norm": 0.15398135781288147,
|
| 653 |
+
"learning_rate": 1.1229554427523972e-05,
|
| 654 |
+
"loss": 0.3,
|
| 655 |
+
"step": 1950
|
| 656 |
+
},
|
| 657 |
+
{
|
| 658 |
+
"epoch": 10.0,
|
| 659 |
+
"eval_accuracy": 0.8654708520179372,
|
| 660 |
+
"eval_loss": 0.7278981804847717,
|
| 661 |
+
"eval_macro_f1": 0.8650225486987954,
|
| 662 |
+
"eval_runtime": 0.4307,
|
| 663 |
+
"eval_samples_per_second": 1553.375,
|
| 664 |
+
"eval_steps_per_second": 97.521,
|
| 665 |
+
"eval_weighted_f1": 0.8651810204853743,
|
| 666 |
+
"step": 1970
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"epoch": 10.025412960609911,
|
| 670 |
+
"grad_norm": 2.287661075592041,
|
| 671 |
+
"learning_rate": 1.1088550479413425e-05,
|
| 672 |
+
"loss": 0.1987,
|
| 673 |
+
"step": 1975
|
| 674 |
+
},
|
| 675 |
+
{
|
| 676 |
+
"epoch": 10.152477763659466,
|
| 677 |
+
"grad_norm": 256.2874755859375,
|
| 678 |
+
"learning_rate": 1.0947546531302877e-05,
|
| 679 |
+
"loss": 0.1249,
|
| 680 |
+
"step": 2000
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 10.279542566709022,
|
| 684 |
+
"grad_norm": 0.16010211408138275,
|
| 685 |
+
"learning_rate": 1.080654258319233e-05,
|
| 686 |
+
"loss": 0.0344,
|
| 687 |
+
"step": 2025
|
| 688 |
+
},
|
| 689 |
+
{
|
| 690 |
+
"epoch": 10.406607369758577,
|
| 691 |
+
"grad_norm": 0.28062617778778076,
|
| 692 |
+
"learning_rate": 1.0665538635081782e-05,
|
| 693 |
+
"loss": 0.1282,
|
| 694 |
+
"step": 2050
|
| 695 |
+
},
|
| 696 |
+
{
|
| 697 |
+
"epoch": 10.533672172808132,
|
| 698 |
+
"grad_norm": 0.13897547125816345,
|
| 699 |
+
"learning_rate": 1.0524534686971236e-05,
|
| 700 |
+
"loss": 0.2207,
|
| 701 |
+
"step": 2075
|
| 702 |
+
},
|
| 703 |
+
{
|
| 704 |
+
"epoch": 10.660736975857688,
|
| 705 |
+
"grad_norm": 15.94092082977295,
|
| 706 |
+
"learning_rate": 1.038353073886069e-05,
|
| 707 |
+
"loss": 0.2183,
|
| 708 |
+
"step": 2100
|
| 709 |
+
},
|
| 710 |
+
{
|
| 711 |
+
"epoch": 10.787801778907243,
|
| 712 |
+
"grad_norm": 0.12359653413295746,
|
| 713 |
+
"learning_rate": 1.0242526790750142e-05,
|
| 714 |
+
"loss": 0.1388,
|
| 715 |
+
"step": 2125
|
| 716 |
+
},
|
| 717 |
+
{
|
| 718 |
+
"epoch": 10.914866581956797,
|
| 719 |
+
"grad_norm": 353.05291748046875,
|
| 720 |
+
"learning_rate": 1.0101522842639596e-05,
|
| 721 |
+
"loss": 0.1265,
|
| 722 |
+
"step": 2150
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 11.0,
|
| 726 |
+
"eval_accuracy": 0.8624813153961136,
|
| 727 |
+
"eval_loss": 0.7706709504127502,
|
| 728 |
+
"eval_macro_f1": 0.8615987424508366,
|
| 729 |
+
"eval_runtime": 0.4817,
|
| 730 |
+
"eval_samples_per_second": 1388.955,
|
| 731 |
+
"eval_steps_per_second": 87.199,
|
| 732 |
+
"eval_weighted_f1": 0.861828714679428,
|
| 733 |
+
"step": 2167
|
| 734 |
+
}
|
| 735 |
+
],
|
| 736 |
+
"logging_steps": 25,
|
| 737 |
+
"max_steps": 3940,
|
| 738 |
+
"num_input_tokens_seen": 0,
|
| 739 |
+
"num_train_epochs": 20,
|
| 740 |
+
"save_steps": 500,
|
| 741 |
+
"stateful_callbacks": {
|
| 742 |
+
"EarlyStoppingCallback": {
|
| 743 |
+
"args": {
|
| 744 |
+
"early_stopping_patience": 3,
|
| 745 |
+
"early_stopping_threshold": 0.0
|
| 746 |
+
},
|
| 747 |
+
"attributes": {
|
| 748 |
+
"early_stopping_patience_counter": 3
|
| 749 |
+
}
|
| 750 |
+
},
|
| 751 |
+
"TrainerControl": {
|
| 752 |
+
"args": {
|
| 753 |
+
"should_epoch_stop": false,
|
| 754 |
+
"should_evaluate": false,
|
| 755 |
+
"should_log": false,
|
| 756 |
+
"should_save": true,
|
| 757 |
+
"should_training_stop": true
|
| 758 |
+
},
|
| 759 |
+
"attributes": {}
|
| 760 |
+
}
|
| 761 |
+
},
|
| 762 |
+
"total_flos": 0.0,
|
| 763 |
+
"train_batch_size": 8,
|
| 764 |
+
"trial_name": null,
|
| 765 |
+
"trial_params": null
|
| 766 |
+
}
|
ood-vifactcheck-phobert-topic-domain-07-s43-20260826T104017Z-bdd2c308/trainer/checkpoint-2167/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-08-s43-20260826T104414Z-bdd2c308/trainer/checkpoint-1700/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-topic-domain-08-s43-20260826T104414Z-bdd2c308/trainer/checkpoint-1700/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-08-s43-20260826T104414Z-bdd2c308/trainer/checkpoint-1700/special_tokens_map.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": "<mask>",
|
| 6 |
+
"pad_token": "<pad>",
|
| 7 |
+
"sep_token": "</s>",
|
| 8 |
+
"unk_token": "<unk>"
|
| 9 |
+
}
|
ood-vifactcheck-phobert-topic-domain-08-s43-20260826T104414Z-bdd2c308/trainer/checkpoint-1700/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"64000": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 256,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "PhobertTokenizer",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-phobert-topic-domain-08-s43-20260826T104414Z-bdd2c308/trainer/checkpoint-1700/trainer_state.json
ADDED
|
@@ -0,0 +1,629 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1700,
|
| 3 |
+
"best_metric": 0.8555290120507512,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-topic-domain-08-s43-20260826T104414Z-bdd2c308/trainer/checkpoint-1700",
|
| 5 |
+
"epoch": 10.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1700,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.14727540500736377,
|
| 14 |
+
"grad_norm": 6.486724376678467,
|
| 15 |
+
"learning_rate": 1.4117647058823531e-06,
|
| 16 |
+
"loss": 4.4315,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.29455081001472755,
|
| 21 |
+
"grad_norm": 6.825076580047607,
|
| 22 |
+
"learning_rate": 2.8823529411764704e-06,
|
| 23 |
+
"loss": 4.4254,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.4418262150220913,
|
| 28 |
+
"grad_norm": 10.510482788085938,
|
| 29 |
+
"learning_rate": 4.352941176470588e-06,
|
| 30 |
+
"loss": 4.4231,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5891016200294551,
|
| 35 |
+
"grad_norm": 7.851330280303955,
|
| 36 |
+
"learning_rate": 5.823529411764707e-06,
|
| 37 |
+
"loss": 4.3527,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.7363770250368189,
|
| 42 |
+
"grad_norm": 7.356525421142578,
|
| 43 |
+
"learning_rate": 7.294117647058823e-06,
|
| 44 |
+
"loss": 4.3409,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8836524300441826,
|
| 49 |
+
"grad_norm": 11.370673179626465,
|
| 50 |
+
"learning_rate": 8.764705882352942e-06,
|
| 51 |
+
"loss": 4.2806,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 1.0,
|
| 56 |
+
"eval_accuracy": 0.5060658578856152,
|
| 57 |
+
"eval_loss": 1.0168977975845337,
|
| 58 |
+
"eval_macro_f1": 0.4620515757934062,
|
| 59 |
+
"eval_runtime": 0.4062,
|
| 60 |
+
"eval_samples_per_second": 1420.62,
|
| 61 |
+
"eval_steps_per_second": 91.097,
|
| 62 |
+
"eval_weighted_f1": 0.46388317555088354,
|
| 63 |
+
"step": 170
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 1.0294550810014726,
|
| 67 |
+
"grad_norm": 13.594656944274902,
|
| 68 |
+
"learning_rate": 1.023529411764706e-05,
|
| 69 |
+
"loss": 4.0534,
|
| 70 |
+
"step": 175
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.1767304860088366,
|
| 74 |
+
"grad_norm": 18.098419189453125,
|
| 75 |
+
"learning_rate": 1.1705882352941178e-05,
|
| 76 |
+
"loss": 3.9659,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.3240058910162003,
|
| 81 |
+
"grad_norm": 40.77829360961914,
|
| 82 |
+
"learning_rate": 1.3176470588235294e-05,
|
| 83 |
+
"loss": 3.5762,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.4712812960235642,
|
| 88 |
+
"grad_norm": 23.517812728881836,
|
| 89 |
+
"learning_rate": 1.4647058823529412e-05,
|
| 90 |
+
"loss": 3.2345,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.6185567010309279,
|
| 95 |
+
"grad_norm": 42.20502853393555,
|
| 96 |
+
"learning_rate": 1.6117647058823533e-05,
|
| 97 |
+
"loss": 2.6742,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.7658321060382915,
|
| 102 |
+
"grad_norm": 34.6644287109375,
|
| 103 |
+
"learning_rate": 1.7588235294117647e-05,
|
| 104 |
+
"loss": 2.4416,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.9131075110456552,
|
| 109 |
+
"grad_norm": 42.68540573120117,
|
| 110 |
+
"learning_rate": 1.9058823529411764e-05,
|
| 111 |
+
"loss": 2.3888,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 2.0,
|
| 116 |
+
"eval_accuracy": 0.7781629116117851,
|
| 117 |
+
"eval_loss": 0.5841764807701111,
|
| 118 |
+
"eval_macro_f1": 0.7723877166984728,
|
| 119 |
+
"eval_runtime": 0.4313,
|
| 120 |
+
"eval_samples_per_second": 1337.716,
|
| 121 |
+
"eval_steps_per_second": 85.781,
|
| 122 |
+
"eval_weighted_f1": 0.7727571503414384,
|
| 123 |
+
"step": 340
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 2.0589101620029453,
|
| 127 |
+
"grad_norm": 42.46329116821289,
|
| 128 |
+
"learning_rate": 1.9941176470588238e-05,
|
| 129 |
+
"loss": 2.156,
|
| 130 |
+
"step": 350
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.2061855670103094,
|
| 134 |
+
"grad_norm": 36.36359405517578,
|
| 135 |
+
"learning_rate": 1.977777777777778e-05,
|
| 136 |
+
"loss": 1.8867,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.353460972017673,
|
| 141 |
+
"grad_norm": 24.46672248840332,
|
| 142 |
+
"learning_rate": 1.961437908496732e-05,
|
| 143 |
+
"loss": 1.8585,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.500736377025037,
|
| 148 |
+
"grad_norm": 32.51788330078125,
|
| 149 |
+
"learning_rate": 1.9450980392156864e-05,
|
| 150 |
+
"loss": 1.7426,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.6480117820324005,
|
| 155 |
+
"grad_norm": 35.60721969604492,
|
| 156 |
+
"learning_rate": 1.9287581699346406e-05,
|
| 157 |
+
"loss": 1.5971,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.795287187039764,
|
| 162 |
+
"grad_norm": 62.60719299316406,
|
| 163 |
+
"learning_rate": 1.912418300653595e-05,
|
| 164 |
+
"loss": 1.5509,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.9425625920471283,
|
| 169 |
+
"grad_norm": 42.51921844482422,
|
| 170 |
+
"learning_rate": 1.8960784313725492e-05,
|
| 171 |
+
"loss": 1.625,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 3.0,
|
| 176 |
+
"eval_accuracy": 0.7989601386481803,
|
| 177 |
+
"eval_loss": 0.5089037418365479,
|
| 178 |
+
"eval_macro_f1": 0.7978597976926946,
|
| 179 |
+
"eval_runtime": 0.4247,
|
| 180 |
+
"eval_samples_per_second": 1358.617,
|
| 181 |
+
"eval_steps_per_second": 87.121,
|
| 182 |
+
"eval_weighted_f1": 0.7981043008967045,
|
| 183 |
+
"step": 510
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 3.0883652430044184,
|
| 187 |
+
"grad_norm": 11.980600357055664,
|
| 188 |
+
"learning_rate": 1.8797385620915035e-05,
|
| 189 |
+
"loss": 1.4646,
|
| 190 |
+
"step": 525
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.235640648011782,
|
| 194 |
+
"grad_norm": 45.5418586730957,
|
| 195 |
+
"learning_rate": 1.8633986928104578e-05,
|
| 196 |
+
"loss": 1.2165,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.3829160530191458,
|
| 201 |
+
"grad_norm": 29.921146392822266,
|
| 202 |
+
"learning_rate": 1.847058823529412e-05,
|
| 203 |
+
"loss": 1.0933,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.5301914580265095,
|
| 208 |
+
"grad_norm": 39.36044692993164,
|
| 209 |
+
"learning_rate": 1.830718954248366e-05,
|
| 210 |
+
"loss": 1.0977,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.6774668630338736,
|
| 215 |
+
"grad_norm": 54.666873931884766,
|
| 216 |
+
"learning_rate": 1.8143790849673203e-05,
|
| 217 |
+
"loss": 1.0885,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.824742268041237,
|
| 222 |
+
"grad_norm": 18.631126403808594,
|
| 223 |
+
"learning_rate": 1.7980392156862746e-05,
|
| 224 |
+
"loss": 1.2287,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.972017673048601,
|
| 229 |
+
"grad_norm": 30.073881149291992,
|
| 230 |
+
"learning_rate": 1.781699346405229e-05,
|
| 231 |
+
"loss": 1.1017,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 4.0,
|
| 236 |
+
"eval_accuracy": 0.8214904679376083,
|
| 237 |
+
"eval_loss": 0.5388094782829285,
|
| 238 |
+
"eval_macro_f1": 0.8202496696341766,
|
| 239 |
+
"eval_runtime": 0.4989,
|
| 240 |
+
"eval_samples_per_second": 1156.495,
|
| 241 |
+
"eval_steps_per_second": 74.16,
|
| 242 |
+
"eval_weighted_f1": 0.8202470374342306,
|
| 243 |
+
"step": 680
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 4.117820324005891,
|
| 247 |
+
"grad_norm": 15.715169906616211,
|
| 248 |
+
"learning_rate": 1.7653594771241832e-05,
|
| 249 |
+
"loss": 0.6694,
|
| 250 |
+
"step": 700
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 4.265095729013255,
|
| 254 |
+
"grad_norm": 11.938085556030273,
|
| 255 |
+
"learning_rate": 1.7490196078431375e-05,
|
| 256 |
+
"loss": 0.8803,
|
| 257 |
+
"step": 725
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.412371134020619,
|
| 261 |
+
"grad_norm": 20.64615821838379,
|
| 262 |
+
"learning_rate": 1.7326797385620918e-05,
|
| 263 |
+
"loss": 1.0727,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.559646539027982,
|
| 268 |
+
"grad_norm": 73.2588119506836,
|
| 269 |
+
"learning_rate": 1.7163398692810457e-05,
|
| 270 |
+
"loss": 0.7987,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.706921944035346,
|
| 275 |
+
"grad_norm": 31.859439849853516,
|
| 276 |
+
"learning_rate": 1.7e-05,
|
| 277 |
+
"loss": 0.8896,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.8541973490427095,
|
| 282 |
+
"grad_norm": 78.91663360595703,
|
| 283 |
+
"learning_rate": 1.6836601307189543e-05,
|
| 284 |
+
"loss": 0.8773,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 5.0,
|
| 289 |
+
"grad_norm": 2.9804372787475586,
|
| 290 |
+
"learning_rate": 1.6673202614379086e-05,
|
| 291 |
+
"loss": 0.6706,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 5.0,
|
| 296 |
+
"eval_accuracy": 0.8318890814558059,
|
| 297 |
+
"eval_loss": 0.5437874794006348,
|
| 298 |
+
"eval_macro_f1": 0.8319716014570372,
|
| 299 |
+
"eval_runtime": 0.4053,
|
| 300 |
+
"eval_samples_per_second": 1423.555,
|
| 301 |
+
"eval_steps_per_second": 91.285,
|
| 302 |
+
"eval_weighted_f1": 0.8320963268638675,
|
| 303 |
+
"step": 850
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 5.147275405007364,
|
| 307 |
+
"grad_norm": 1.5131419897079468,
|
| 308 |
+
"learning_rate": 1.650980392156863e-05,
|
| 309 |
+
"loss": 0.4754,
|
| 310 |
+
"step": 875
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 5.294550810014727,
|
| 314 |
+
"grad_norm": 1.3377717733383179,
|
| 315 |
+
"learning_rate": 1.634640522875817e-05,
|
| 316 |
+
"loss": 0.5492,
|
| 317 |
+
"step": 900
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.4418262150220915,
|
| 321 |
+
"grad_norm": 6.428271293640137,
|
| 322 |
+
"learning_rate": 1.6183006535947714e-05,
|
| 323 |
+
"loss": 0.6381,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.589101620029455,
|
| 328 |
+
"grad_norm": 48.77164840698242,
|
| 329 |
+
"learning_rate": 1.6019607843137257e-05,
|
| 330 |
+
"loss": 0.7021,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.736377025036819,
|
| 335 |
+
"grad_norm": 143.2543182373047,
|
| 336 |
+
"learning_rate": 1.5856209150326797e-05,
|
| 337 |
+
"loss": 0.6093,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.883652430044183,
|
| 342 |
+
"grad_norm": 78.71709442138672,
|
| 343 |
+
"learning_rate": 1.569281045751634e-05,
|
| 344 |
+
"loss": 0.5926,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 6.0,
|
| 349 |
+
"eval_accuracy": 0.8388214904679376,
|
| 350 |
+
"eval_loss": 0.6194702386856079,
|
| 351 |
+
"eval_macro_f1": 0.8395072600779185,
|
| 352 |
+
"eval_runtime": 0.4227,
|
| 353 |
+
"eval_samples_per_second": 1365.162,
|
| 354 |
+
"eval_steps_per_second": 87.541,
|
| 355 |
+
"eval_weighted_f1": 0.8395318110676193,
|
| 356 |
+
"step": 1020
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 6.029455081001473,
|
| 360 |
+
"grad_norm": 39.40867614746094,
|
| 361 |
+
"learning_rate": 1.5529411764705882e-05,
|
| 362 |
+
"loss": 0.4849,
|
| 363 |
+
"step": 1025
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 6.176730486008837,
|
| 367 |
+
"grad_norm": 63.56937026977539,
|
| 368 |
+
"learning_rate": 1.536601307189543e-05,
|
| 369 |
+
"loss": 0.5081,
|
| 370 |
+
"step": 1050
|
| 371 |
+
},
|
| 372 |
+
{
|
| 373 |
+
"epoch": 6.3240058910162,
|
| 374 |
+
"grad_norm": 73.87124633789062,
|
| 375 |
+
"learning_rate": 1.5202614379084968e-05,
|
| 376 |
+
"loss": 0.3455,
|
| 377 |
+
"step": 1075
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 6.471281296023564,
|
| 381 |
+
"grad_norm": 187.9053192138672,
|
| 382 |
+
"learning_rate": 1.5039215686274511e-05,
|
| 383 |
+
"loss": 0.5318,
|
| 384 |
+
"step": 1100
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 6.618556701030927,
|
| 388 |
+
"grad_norm": 47.76657485961914,
|
| 389 |
+
"learning_rate": 1.4875816993464054e-05,
|
| 390 |
+
"loss": 0.4722,
|
| 391 |
+
"step": 1125
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 6.7658321060382915,
|
| 395 |
+
"grad_norm": 109.37071228027344,
|
| 396 |
+
"learning_rate": 1.4712418300653597e-05,
|
| 397 |
+
"loss": 0.4377,
|
| 398 |
+
"step": 1150
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.913107511045656,
|
| 402 |
+
"grad_norm": 33.112640380859375,
|
| 403 |
+
"learning_rate": 1.4549019607843138e-05,
|
| 404 |
+
"loss": 0.4936,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 7.0,
|
| 409 |
+
"eval_accuracy": 0.8509532062391681,
|
| 410 |
+
"eval_loss": 0.6717767119407654,
|
| 411 |
+
"eval_macro_f1": 0.8496673248475018,
|
| 412 |
+
"eval_runtime": 0.4137,
|
| 413 |
+
"eval_samples_per_second": 1394.868,
|
| 414 |
+
"eval_steps_per_second": 89.446,
|
| 415 |
+
"eval_weighted_f1": 0.8496751745583694,
|
| 416 |
+
"step": 1190
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 7.058910162002945,
|
| 420 |
+
"grad_norm": 10.898920059204102,
|
| 421 |
+
"learning_rate": 1.438562091503268e-05,
|
| 422 |
+
"loss": 0.3105,
|
| 423 |
+
"step": 1200
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 7.206185567010309,
|
| 427 |
+
"grad_norm": 213.54376220703125,
|
| 428 |
+
"learning_rate": 1.4222222222222224e-05,
|
| 429 |
+
"loss": 0.2539,
|
| 430 |
+
"step": 1225
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 7.353460972017673,
|
| 434 |
+
"grad_norm": 15.636307716369629,
|
| 435 |
+
"learning_rate": 1.4058823529411765e-05,
|
| 436 |
+
"loss": 0.3239,
|
| 437 |
+
"step": 1250
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 7.500736377025037,
|
| 441 |
+
"grad_norm": 0.3922487497329712,
|
| 442 |
+
"learning_rate": 1.3895424836601308e-05,
|
| 443 |
+
"loss": 0.3853,
|
| 444 |
+
"step": 1275
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 7.648011782032401,
|
| 448 |
+
"grad_norm": 121.3844223022461,
|
| 449 |
+
"learning_rate": 1.373202614379085e-05,
|
| 450 |
+
"loss": 0.2645,
|
| 451 |
+
"step": 1300
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 7.795287187039764,
|
| 455 |
+
"grad_norm": 2.173407554626465,
|
| 456 |
+
"learning_rate": 1.3568627450980393e-05,
|
| 457 |
+
"loss": 0.3379,
|
| 458 |
+
"step": 1325
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 7.942562592047128,
|
| 462 |
+
"grad_norm": 0.45863407850265503,
|
| 463 |
+
"learning_rate": 1.3405228758169935e-05,
|
| 464 |
+
"loss": 0.3742,
|
| 465 |
+
"step": 1350
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 8.0,
|
| 469 |
+
"eval_accuracy": 0.8336221837088388,
|
| 470 |
+
"eval_loss": 0.8225600719451904,
|
| 471 |
+
"eval_macro_f1": 0.8325196196915833,
|
| 472 |
+
"eval_runtime": 0.4382,
|
| 473 |
+
"eval_samples_per_second": 1316.88,
|
| 474 |
+
"eval_steps_per_second": 84.445,
|
| 475 |
+
"eval_weighted_f1": 0.8324089163924374,
|
| 476 |
+
"step": 1360
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 8.088365243004418,
|
| 480 |
+
"grad_norm": 44.608524322509766,
|
| 481 |
+
"learning_rate": 1.3241830065359478e-05,
|
| 482 |
+
"loss": 0.325,
|
| 483 |
+
"step": 1375
|
| 484 |
+
},
|
| 485 |
+
{
|
| 486 |
+
"epoch": 8.235640648011781,
|
| 487 |
+
"grad_norm": 0.33685338497161865,
|
| 488 |
+
"learning_rate": 1.307843137254902e-05,
|
| 489 |
+
"loss": 0.2776,
|
| 490 |
+
"step": 1400
|
| 491 |
+
},
|
| 492 |
+
{
|
| 493 |
+
"epoch": 8.382916053019146,
|
| 494 |
+
"grad_norm": 4.775301456451416,
|
| 495 |
+
"learning_rate": 1.2915032679738563e-05,
|
| 496 |
+
"loss": 0.2569,
|
| 497 |
+
"step": 1425
|
| 498 |
+
},
|
| 499 |
+
{
|
| 500 |
+
"epoch": 8.53019145802651,
|
| 501 |
+
"grad_norm": 23.33795738220215,
|
| 502 |
+
"learning_rate": 1.2751633986928104e-05,
|
| 503 |
+
"loss": 0.3486,
|
| 504 |
+
"step": 1450
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 8.677466863033873,
|
| 508 |
+
"grad_norm": 7.967680931091309,
|
| 509 |
+
"learning_rate": 1.2588235294117647e-05,
|
| 510 |
+
"loss": 0.3043,
|
| 511 |
+
"step": 1475
|
| 512 |
+
},
|
| 513 |
+
{
|
| 514 |
+
"epoch": 8.824742268041238,
|
| 515 |
+
"grad_norm": 1.7140271663665771,
|
| 516 |
+
"learning_rate": 1.242483660130719e-05,
|
| 517 |
+
"loss": 0.323,
|
| 518 |
+
"step": 1500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 8.972017673048601,
|
| 522 |
+
"grad_norm": 22.50177764892578,
|
| 523 |
+
"learning_rate": 1.2261437908496733e-05,
|
| 524 |
+
"loss": 0.2039,
|
| 525 |
+
"step": 1525
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 9.0,
|
| 529 |
+
"eval_accuracy": 0.8353552859618717,
|
| 530 |
+
"eval_loss": 0.7928684949874878,
|
| 531 |
+
"eval_macro_f1": 0.8351727747753147,
|
| 532 |
+
"eval_runtime": 0.4423,
|
| 533 |
+
"eval_samples_per_second": 1304.458,
|
| 534 |
+
"eval_steps_per_second": 83.648,
|
| 535 |
+
"eval_weighted_f1": 0.8351557296485701,
|
| 536 |
+
"step": 1530
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 9.11782032400589,
|
| 540 |
+
"grad_norm": 1.6093811988830566,
|
| 541 |
+
"learning_rate": 1.2098039215686274e-05,
|
| 542 |
+
"loss": 0.16,
|
| 543 |
+
"step": 1550
|
| 544 |
+
},
|
| 545 |
+
{
|
| 546 |
+
"epoch": 9.265095729013256,
|
| 547 |
+
"grad_norm": 61.401283264160156,
|
| 548 |
+
"learning_rate": 1.1934640522875817e-05,
|
| 549 |
+
"loss": 0.2844,
|
| 550 |
+
"step": 1575
|
| 551 |
+
},
|
| 552 |
+
{
|
| 553 |
+
"epoch": 9.412371134020619,
|
| 554 |
+
"grad_norm": 5.321173191070557,
|
| 555 |
+
"learning_rate": 1.177124183006536e-05,
|
| 556 |
+
"loss": 0.2376,
|
| 557 |
+
"step": 1600
|
| 558 |
+
},
|
| 559 |
+
{
|
| 560 |
+
"epoch": 9.559646539027982,
|
| 561 |
+
"grad_norm": 0.19536180794239044,
|
| 562 |
+
"learning_rate": 1.1607843137254905e-05,
|
| 563 |
+
"loss": 0.1557,
|
| 564 |
+
"step": 1625
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"epoch": 9.706921944035345,
|
| 568 |
+
"grad_norm": 3.606111526489258,
|
| 569 |
+
"learning_rate": 1.1444444444444444e-05,
|
| 570 |
+
"loss": 0.2872,
|
| 571 |
+
"step": 1650
|
| 572 |
+
},
|
| 573 |
+
{
|
| 574 |
+
"epoch": 9.85419734904271,
|
| 575 |
+
"grad_norm": 125.77471160888672,
|
| 576 |
+
"learning_rate": 1.1281045751633987e-05,
|
| 577 |
+
"loss": 0.1534,
|
| 578 |
+
"step": 1675
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 10.0,
|
| 582 |
+
"grad_norm": 0.41708627343177795,
|
| 583 |
+
"learning_rate": 1.1117647058823531e-05,
|
| 584 |
+
"loss": 0.1838,
|
| 585 |
+
"step": 1700
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 10.0,
|
| 589 |
+
"eval_accuracy": 0.8561525129982669,
|
| 590 |
+
"eval_loss": 0.7656728029251099,
|
| 591 |
+
"eval_macro_f1": 0.8555290120507512,
|
| 592 |
+
"eval_runtime": 0.5289,
|
| 593 |
+
"eval_samples_per_second": 1090.958,
|
| 594 |
+
"eval_steps_per_second": 69.957,
|
| 595 |
+
"eval_weighted_f1": 0.8553684629489016,
|
| 596 |
+
"step": 1700
|
| 597 |
+
}
|
| 598 |
+
],
|
| 599 |
+
"logging_steps": 25,
|
| 600 |
+
"max_steps": 3400,
|
| 601 |
+
"num_input_tokens_seen": 0,
|
| 602 |
+
"num_train_epochs": 20,
|
| 603 |
+
"save_steps": 500,
|
| 604 |
+
"stateful_callbacks": {
|
| 605 |
+
"EarlyStoppingCallback": {
|
| 606 |
+
"args": {
|
| 607 |
+
"early_stopping_patience": 3,
|
| 608 |
+
"early_stopping_threshold": 0.0
|
| 609 |
+
},
|
| 610 |
+
"attributes": {
|
| 611 |
+
"early_stopping_patience_counter": 0
|
| 612 |
+
}
|
| 613 |
+
},
|
| 614 |
+
"TrainerControl": {
|
| 615 |
+
"args": {
|
| 616 |
+
"should_epoch_stop": false,
|
| 617 |
+
"should_evaluate": false,
|
| 618 |
+
"should_log": false,
|
| 619 |
+
"should_save": true,
|
| 620 |
+
"should_training_stop": false
|
| 621 |
+
},
|
| 622 |
+
"attributes": {}
|
| 623 |
+
}
|
| 624 |
+
},
|
| 625 |
+
"total_flos": 0.0,
|
| 626 |
+
"train_batch_size": 8,
|
| 627 |
+
"trial_name": null,
|
| 628 |
+
"trial_params": null
|
| 629 |
+
}
|
ood-vifactcheck-phobert-topic-domain-08-s43-20260826T104414Z-bdd2c308/trainer/checkpoint-1700/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-08-s43-20260826T104414Z-bdd2c308/trainer/checkpoint-2210/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-topic-domain-08-s43-20260826T104414Z-bdd2c308/trainer/checkpoint-2210/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-08-s43-20260826T104414Z-bdd2c308/trainer/checkpoint-2210/special_tokens_map.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": "<mask>",
|
| 6 |
+
"pad_token": "<pad>",
|
| 7 |
+
"sep_token": "</s>",
|
| 8 |
+
"unk_token": "<unk>"
|
| 9 |
+
}
|
ood-vifactcheck-phobert-topic-domain-08-s43-20260826T104414Z-bdd2c308/trainer/checkpoint-2210/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"64000": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 256,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "PhobertTokenizer",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-phobert-topic-domain-08-s43-20260826T104414Z-bdd2c308/trainer/checkpoint-2210/trainer_state.json
ADDED
|
@@ -0,0 +1,802 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1700,
|
| 3 |
+
"best_metric": 0.8555290120507512,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-topic-domain-08-s43-20260826T104414Z-bdd2c308/trainer/checkpoint-1700",
|
| 5 |
+
"epoch": 13.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 2210,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.14727540500736377,
|
| 14 |
+
"grad_norm": 6.486724376678467,
|
| 15 |
+
"learning_rate": 1.4117647058823531e-06,
|
| 16 |
+
"loss": 4.4315,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.29455081001472755,
|
| 21 |
+
"grad_norm": 6.825076580047607,
|
| 22 |
+
"learning_rate": 2.8823529411764704e-06,
|
| 23 |
+
"loss": 4.4254,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.4418262150220913,
|
| 28 |
+
"grad_norm": 10.510482788085938,
|
| 29 |
+
"learning_rate": 4.352941176470588e-06,
|
| 30 |
+
"loss": 4.4231,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5891016200294551,
|
| 35 |
+
"grad_norm": 7.851330280303955,
|
| 36 |
+
"learning_rate": 5.823529411764707e-06,
|
| 37 |
+
"loss": 4.3527,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.7363770250368189,
|
| 42 |
+
"grad_norm": 7.356525421142578,
|
| 43 |
+
"learning_rate": 7.294117647058823e-06,
|
| 44 |
+
"loss": 4.3409,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8836524300441826,
|
| 49 |
+
"grad_norm": 11.370673179626465,
|
| 50 |
+
"learning_rate": 8.764705882352942e-06,
|
| 51 |
+
"loss": 4.2806,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 1.0,
|
| 56 |
+
"eval_accuracy": 0.5060658578856152,
|
| 57 |
+
"eval_loss": 1.0168977975845337,
|
| 58 |
+
"eval_macro_f1": 0.4620515757934062,
|
| 59 |
+
"eval_runtime": 0.4062,
|
| 60 |
+
"eval_samples_per_second": 1420.62,
|
| 61 |
+
"eval_steps_per_second": 91.097,
|
| 62 |
+
"eval_weighted_f1": 0.46388317555088354,
|
| 63 |
+
"step": 170
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 1.0294550810014726,
|
| 67 |
+
"grad_norm": 13.594656944274902,
|
| 68 |
+
"learning_rate": 1.023529411764706e-05,
|
| 69 |
+
"loss": 4.0534,
|
| 70 |
+
"step": 175
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.1767304860088366,
|
| 74 |
+
"grad_norm": 18.098419189453125,
|
| 75 |
+
"learning_rate": 1.1705882352941178e-05,
|
| 76 |
+
"loss": 3.9659,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.3240058910162003,
|
| 81 |
+
"grad_norm": 40.77829360961914,
|
| 82 |
+
"learning_rate": 1.3176470588235294e-05,
|
| 83 |
+
"loss": 3.5762,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.4712812960235642,
|
| 88 |
+
"grad_norm": 23.517812728881836,
|
| 89 |
+
"learning_rate": 1.4647058823529412e-05,
|
| 90 |
+
"loss": 3.2345,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.6185567010309279,
|
| 95 |
+
"grad_norm": 42.20502853393555,
|
| 96 |
+
"learning_rate": 1.6117647058823533e-05,
|
| 97 |
+
"loss": 2.6742,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.7658321060382915,
|
| 102 |
+
"grad_norm": 34.6644287109375,
|
| 103 |
+
"learning_rate": 1.7588235294117647e-05,
|
| 104 |
+
"loss": 2.4416,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.9131075110456552,
|
| 109 |
+
"grad_norm": 42.68540573120117,
|
| 110 |
+
"learning_rate": 1.9058823529411764e-05,
|
| 111 |
+
"loss": 2.3888,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 2.0,
|
| 116 |
+
"eval_accuracy": 0.7781629116117851,
|
| 117 |
+
"eval_loss": 0.5841764807701111,
|
| 118 |
+
"eval_macro_f1": 0.7723877166984728,
|
| 119 |
+
"eval_runtime": 0.4313,
|
| 120 |
+
"eval_samples_per_second": 1337.716,
|
| 121 |
+
"eval_steps_per_second": 85.781,
|
| 122 |
+
"eval_weighted_f1": 0.7727571503414384,
|
| 123 |
+
"step": 340
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 2.0589101620029453,
|
| 127 |
+
"grad_norm": 42.46329116821289,
|
| 128 |
+
"learning_rate": 1.9941176470588238e-05,
|
| 129 |
+
"loss": 2.156,
|
| 130 |
+
"step": 350
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.2061855670103094,
|
| 134 |
+
"grad_norm": 36.36359405517578,
|
| 135 |
+
"learning_rate": 1.977777777777778e-05,
|
| 136 |
+
"loss": 1.8867,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.353460972017673,
|
| 141 |
+
"grad_norm": 24.46672248840332,
|
| 142 |
+
"learning_rate": 1.961437908496732e-05,
|
| 143 |
+
"loss": 1.8585,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.500736377025037,
|
| 148 |
+
"grad_norm": 32.51788330078125,
|
| 149 |
+
"learning_rate": 1.9450980392156864e-05,
|
| 150 |
+
"loss": 1.7426,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.6480117820324005,
|
| 155 |
+
"grad_norm": 35.60721969604492,
|
| 156 |
+
"learning_rate": 1.9287581699346406e-05,
|
| 157 |
+
"loss": 1.5971,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.795287187039764,
|
| 162 |
+
"grad_norm": 62.60719299316406,
|
| 163 |
+
"learning_rate": 1.912418300653595e-05,
|
| 164 |
+
"loss": 1.5509,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.9425625920471283,
|
| 169 |
+
"grad_norm": 42.51921844482422,
|
| 170 |
+
"learning_rate": 1.8960784313725492e-05,
|
| 171 |
+
"loss": 1.625,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 3.0,
|
| 176 |
+
"eval_accuracy": 0.7989601386481803,
|
| 177 |
+
"eval_loss": 0.5089037418365479,
|
| 178 |
+
"eval_macro_f1": 0.7978597976926946,
|
| 179 |
+
"eval_runtime": 0.4247,
|
| 180 |
+
"eval_samples_per_second": 1358.617,
|
| 181 |
+
"eval_steps_per_second": 87.121,
|
| 182 |
+
"eval_weighted_f1": 0.7981043008967045,
|
| 183 |
+
"step": 510
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 3.0883652430044184,
|
| 187 |
+
"grad_norm": 11.980600357055664,
|
| 188 |
+
"learning_rate": 1.8797385620915035e-05,
|
| 189 |
+
"loss": 1.4646,
|
| 190 |
+
"step": 525
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.235640648011782,
|
| 194 |
+
"grad_norm": 45.5418586730957,
|
| 195 |
+
"learning_rate": 1.8633986928104578e-05,
|
| 196 |
+
"loss": 1.2165,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.3829160530191458,
|
| 201 |
+
"grad_norm": 29.921146392822266,
|
| 202 |
+
"learning_rate": 1.847058823529412e-05,
|
| 203 |
+
"loss": 1.0933,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.5301914580265095,
|
| 208 |
+
"grad_norm": 39.36044692993164,
|
| 209 |
+
"learning_rate": 1.830718954248366e-05,
|
| 210 |
+
"loss": 1.0977,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.6774668630338736,
|
| 215 |
+
"grad_norm": 54.666873931884766,
|
| 216 |
+
"learning_rate": 1.8143790849673203e-05,
|
| 217 |
+
"loss": 1.0885,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.824742268041237,
|
| 222 |
+
"grad_norm": 18.631126403808594,
|
| 223 |
+
"learning_rate": 1.7980392156862746e-05,
|
| 224 |
+
"loss": 1.2287,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.972017673048601,
|
| 229 |
+
"grad_norm": 30.073881149291992,
|
| 230 |
+
"learning_rate": 1.781699346405229e-05,
|
| 231 |
+
"loss": 1.1017,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 4.0,
|
| 236 |
+
"eval_accuracy": 0.8214904679376083,
|
| 237 |
+
"eval_loss": 0.5388094782829285,
|
| 238 |
+
"eval_macro_f1": 0.8202496696341766,
|
| 239 |
+
"eval_runtime": 0.4989,
|
| 240 |
+
"eval_samples_per_second": 1156.495,
|
| 241 |
+
"eval_steps_per_second": 74.16,
|
| 242 |
+
"eval_weighted_f1": 0.8202470374342306,
|
| 243 |
+
"step": 680
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 4.117820324005891,
|
| 247 |
+
"grad_norm": 15.715169906616211,
|
| 248 |
+
"learning_rate": 1.7653594771241832e-05,
|
| 249 |
+
"loss": 0.6694,
|
| 250 |
+
"step": 700
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 4.265095729013255,
|
| 254 |
+
"grad_norm": 11.938085556030273,
|
| 255 |
+
"learning_rate": 1.7490196078431375e-05,
|
| 256 |
+
"loss": 0.8803,
|
| 257 |
+
"step": 725
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.412371134020619,
|
| 261 |
+
"grad_norm": 20.64615821838379,
|
| 262 |
+
"learning_rate": 1.7326797385620918e-05,
|
| 263 |
+
"loss": 1.0727,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.559646539027982,
|
| 268 |
+
"grad_norm": 73.2588119506836,
|
| 269 |
+
"learning_rate": 1.7163398692810457e-05,
|
| 270 |
+
"loss": 0.7987,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.706921944035346,
|
| 275 |
+
"grad_norm": 31.859439849853516,
|
| 276 |
+
"learning_rate": 1.7e-05,
|
| 277 |
+
"loss": 0.8896,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.8541973490427095,
|
| 282 |
+
"grad_norm": 78.91663360595703,
|
| 283 |
+
"learning_rate": 1.6836601307189543e-05,
|
| 284 |
+
"loss": 0.8773,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 5.0,
|
| 289 |
+
"grad_norm": 2.9804372787475586,
|
| 290 |
+
"learning_rate": 1.6673202614379086e-05,
|
| 291 |
+
"loss": 0.6706,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 5.0,
|
| 296 |
+
"eval_accuracy": 0.8318890814558059,
|
| 297 |
+
"eval_loss": 0.5437874794006348,
|
| 298 |
+
"eval_macro_f1": 0.8319716014570372,
|
| 299 |
+
"eval_runtime": 0.4053,
|
| 300 |
+
"eval_samples_per_second": 1423.555,
|
| 301 |
+
"eval_steps_per_second": 91.285,
|
| 302 |
+
"eval_weighted_f1": 0.8320963268638675,
|
| 303 |
+
"step": 850
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 5.147275405007364,
|
| 307 |
+
"grad_norm": 1.5131419897079468,
|
| 308 |
+
"learning_rate": 1.650980392156863e-05,
|
| 309 |
+
"loss": 0.4754,
|
| 310 |
+
"step": 875
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 5.294550810014727,
|
| 314 |
+
"grad_norm": 1.3377717733383179,
|
| 315 |
+
"learning_rate": 1.634640522875817e-05,
|
| 316 |
+
"loss": 0.5492,
|
| 317 |
+
"step": 900
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.4418262150220915,
|
| 321 |
+
"grad_norm": 6.428271293640137,
|
| 322 |
+
"learning_rate": 1.6183006535947714e-05,
|
| 323 |
+
"loss": 0.6381,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.589101620029455,
|
| 328 |
+
"grad_norm": 48.77164840698242,
|
| 329 |
+
"learning_rate": 1.6019607843137257e-05,
|
| 330 |
+
"loss": 0.7021,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.736377025036819,
|
| 335 |
+
"grad_norm": 143.2543182373047,
|
| 336 |
+
"learning_rate": 1.5856209150326797e-05,
|
| 337 |
+
"loss": 0.6093,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.883652430044183,
|
| 342 |
+
"grad_norm": 78.71709442138672,
|
| 343 |
+
"learning_rate": 1.569281045751634e-05,
|
| 344 |
+
"loss": 0.5926,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 6.0,
|
| 349 |
+
"eval_accuracy": 0.8388214904679376,
|
| 350 |
+
"eval_loss": 0.6194702386856079,
|
| 351 |
+
"eval_macro_f1": 0.8395072600779185,
|
| 352 |
+
"eval_runtime": 0.4227,
|
| 353 |
+
"eval_samples_per_second": 1365.162,
|
| 354 |
+
"eval_steps_per_second": 87.541,
|
| 355 |
+
"eval_weighted_f1": 0.8395318110676193,
|
| 356 |
+
"step": 1020
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 6.029455081001473,
|
| 360 |
+
"grad_norm": 39.40867614746094,
|
| 361 |
+
"learning_rate": 1.5529411764705882e-05,
|
| 362 |
+
"loss": 0.4849,
|
| 363 |
+
"step": 1025
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 6.176730486008837,
|
| 367 |
+
"grad_norm": 63.56937026977539,
|
| 368 |
+
"learning_rate": 1.536601307189543e-05,
|
| 369 |
+
"loss": 0.5081,
|
| 370 |
+
"step": 1050
|
| 371 |
+
},
|
| 372 |
+
{
|
| 373 |
+
"epoch": 6.3240058910162,
|
| 374 |
+
"grad_norm": 73.87124633789062,
|
| 375 |
+
"learning_rate": 1.5202614379084968e-05,
|
| 376 |
+
"loss": 0.3455,
|
| 377 |
+
"step": 1075
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 6.471281296023564,
|
| 381 |
+
"grad_norm": 187.9053192138672,
|
| 382 |
+
"learning_rate": 1.5039215686274511e-05,
|
| 383 |
+
"loss": 0.5318,
|
| 384 |
+
"step": 1100
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 6.618556701030927,
|
| 388 |
+
"grad_norm": 47.76657485961914,
|
| 389 |
+
"learning_rate": 1.4875816993464054e-05,
|
| 390 |
+
"loss": 0.4722,
|
| 391 |
+
"step": 1125
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 6.7658321060382915,
|
| 395 |
+
"grad_norm": 109.37071228027344,
|
| 396 |
+
"learning_rate": 1.4712418300653597e-05,
|
| 397 |
+
"loss": 0.4377,
|
| 398 |
+
"step": 1150
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.913107511045656,
|
| 402 |
+
"grad_norm": 33.112640380859375,
|
| 403 |
+
"learning_rate": 1.4549019607843138e-05,
|
| 404 |
+
"loss": 0.4936,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 7.0,
|
| 409 |
+
"eval_accuracy": 0.8509532062391681,
|
| 410 |
+
"eval_loss": 0.6717767119407654,
|
| 411 |
+
"eval_macro_f1": 0.8496673248475018,
|
| 412 |
+
"eval_runtime": 0.4137,
|
| 413 |
+
"eval_samples_per_second": 1394.868,
|
| 414 |
+
"eval_steps_per_second": 89.446,
|
| 415 |
+
"eval_weighted_f1": 0.8496751745583694,
|
| 416 |
+
"step": 1190
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 7.058910162002945,
|
| 420 |
+
"grad_norm": 10.898920059204102,
|
| 421 |
+
"learning_rate": 1.438562091503268e-05,
|
| 422 |
+
"loss": 0.3105,
|
| 423 |
+
"step": 1200
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 7.206185567010309,
|
| 427 |
+
"grad_norm": 213.54376220703125,
|
| 428 |
+
"learning_rate": 1.4222222222222224e-05,
|
| 429 |
+
"loss": 0.2539,
|
| 430 |
+
"step": 1225
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 7.353460972017673,
|
| 434 |
+
"grad_norm": 15.636307716369629,
|
| 435 |
+
"learning_rate": 1.4058823529411765e-05,
|
| 436 |
+
"loss": 0.3239,
|
| 437 |
+
"step": 1250
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 7.500736377025037,
|
| 441 |
+
"grad_norm": 0.3922487497329712,
|
| 442 |
+
"learning_rate": 1.3895424836601308e-05,
|
| 443 |
+
"loss": 0.3853,
|
| 444 |
+
"step": 1275
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 7.648011782032401,
|
| 448 |
+
"grad_norm": 121.3844223022461,
|
| 449 |
+
"learning_rate": 1.373202614379085e-05,
|
| 450 |
+
"loss": 0.2645,
|
| 451 |
+
"step": 1300
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 7.795287187039764,
|
| 455 |
+
"grad_norm": 2.173407554626465,
|
| 456 |
+
"learning_rate": 1.3568627450980393e-05,
|
| 457 |
+
"loss": 0.3379,
|
| 458 |
+
"step": 1325
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 7.942562592047128,
|
| 462 |
+
"grad_norm": 0.45863407850265503,
|
| 463 |
+
"learning_rate": 1.3405228758169935e-05,
|
| 464 |
+
"loss": 0.3742,
|
| 465 |
+
"step": 1350
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 8.0,
|
| 469 |
+
"eval_accuracy": 0.8336221837088388,
|
| 470 |
+
"eval_loss": 0.8225600719451904,
|
| 471 |
+
"eval_macro_f1": 0.8325196196915833,
|
| 472 |
+
"eval_runtime": 0.4382,
|
| 473 |
+
"eval_samples_per_second": 1316.88,
|
| 474 |
+
"eval_steps_per_second": 84.445,
|
| 475 |
+
"eval_weighted_f1": 0.8324089163924374,
|
| 476 |
+
"step": 1360
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 8.088365243004418,
|
| 480 |
+
"grad_norm": 44.608524322509766,
|
| 481 |
+
"learning_rate": 1.3241830065359478e-05,
|
| 482 |
+
"loss": 0.325,
|
| 483 |
+
"step": 1375
|
| 484 |
+
},
|
| 485 |
+
{
|
| 486 |
+
"epoch": 8.235640648011781,
|
| 487 |
+
"grad_norm": 0.33685338497161865,
|
| 488 |
+
"learning_rate": 1.307843137254902e-05,
|
| 489 |
+
"loss": 0.2776,
|
| 490 |
+
"step": 1400
|
| 491 |
+
},
|
| 492 |
+
{
|
| 493 |
+
"epoch": 8.382916053019146,
|
| 494 |
+
"grad_norm": 4.775301456451416,
|
| 495 |
+
"learning_rate": 1.2915032679738563e-05,
|
| 496 |
+
"loss": 0.2569,
|
| 497 |
+
"step": 1425
|
| 498 |
+
},
|
| 499 |
+
{
|
| 500 |
+
"epoch": 8.53019145802651,
|
| 501 |
+
"grad_norm": 23.33795738220215,
|
| 502 |
+
"learning_rate": 1.2751633986928104e-05,
|
| 503 |
+
"loss": 0.3486,
|
| 504 |
+
"step": 1450
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 8.677466863033873,
|
| 508 |
+
"grad_norm": 7.967680931091309,
|
| 509 |
+
"learning_rate": 1.2588235294117647e-05,
|
| 510 |
+
"loss": 0.3043,
|
| 511 |
+
"step": 1475
|
| 512 |
+
},
|
| 513 |
+
{
|
| 514 |
+
"epoch": 8.824742268041238,
|
| 515 |
+
"grad_norm": 1.7140271663665771,
|
| 516 |
+
"learning_rate": 1.242483660130719e-05,
|
| 517 |
+
"loss": 0.323,
|
| 518 |
+
"step": 1500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 8.972017673048601,
|
| 522 |
+
"grad_norm": 22.50177764892578,
|
| 523 |
+
"learning_rate": 1.2261437908496733e-05,
|
| 524 |
+
"loss": 0.2039,
|
| 525 |
+
"step": 1525
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 9.0,
|
| 529 |
+
"eval_accuracy": 0.8353552859618717,
|
| 530 |
+
"eval_loss": 0.7928684949874878,
|
| 531 |
+
"eval_macro_f1": 0.8351727747753147,
|
| 532 |
+
"eval_runtime": 0.4423,
|
| 533 |
+
"eval_samples_per_second": 1304.458,
|
| 534 |
+
"eval_steps_per_second": 83.648,
|
| 535 |
+
"eval_weighted_f1": 0.8351557296485701,
|
| 536 |
+
"step": 1530
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 9.11782032400589,
|
| 540 |
+
"grad_norm": 1.6093811988830566,
|
| 541 |
+
"learning_rate": 1.2098039215686274e-05,
|
| 542 |
+
"loss": 0.16,
|
| 543 |
+
"step": 1550
|
| 544 |
+
},
|
| 545 |
+
{
|
| 546 |
+
"epoch": 9.265095729013256,
|
| 547 |
+
"grad_norm": 61.401283264160156,
|
| 548 |
+
"learning_rate": 1.1934640522875817e-05,
|
| 549 |
+
"loss": 0.2844,
|
| 550 |
+
"step": 1575
|
| 551 |
+
},
|
| 552 |
+
{
|
| 553 |
+
"epoch": 9.412371134020619,
|
| 554 |
+
"grad_norm": 5.321173191070557,
|
| 555 |
+
"learning_rate": 1.177124183006536e-05,
|
| 556 |
+
"loss": 0.2376,
|
| 557 |
+
"step": 1600
|
| 558 |
+
},
|
| 559 |
+
{
|
| 560 |
+
"epoch": 9.559646539027982,
|
| 561 |
+
"grad_norm": 0.19536180794239044,
|
| 562 |
+
"learning_rate": 1.1607843137254905e-05,
|
| 563 |
+
"loss": 0.1557,
|
| 564 |
+
"step": 1625
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"epoch": 9.706921944035345,
|
| 568 |
+
"grad_norm": 3.606111526489258,
|
| 569 |
+
"learning_rate": 1.1444444444444444e-05,
|
| 570 |
+
"loss": 0.2872,
|
| 571 |
+
"step": 1650
|
| 572 |
+
},
|
| 573 |
+
{
|
| 574 |
+
"epoch": 9.85419734904271,
|
| 575 |
+
"grad_norm": 125.77471160888672,
|
| 576 |
+
"learning_rate": 1.1281045751633987e-05,
|
| 577 |
+
"loss": 0.1534,
|
| 578 |
+
"step": 1675
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 10.0,
|
| 582 |
+
"grad_norm": 0.41708627343177795,
|
| 583 |
+
"learning_rate": 1.1117647058823531e-05,
|
| 584 |
+
"loss": 0.1838,
|
| 585 |
+
"step": 1700
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 10.0,
|
| 589 |
+
"eval_accuracy": 0.8561525129982669,
|
| 590 |
+
"eval_loss": 0.7656728029251099,
|
| 591 |
+
"eval_macro_f1": 0.8555290120507512,
|
| 592 |
+
"eval_runtime": 0.5289,
|
| 593 |
+
"eval_samples_per_second": 1090.958,
|
| 594 |
+
"eval_steps_per_second": 69.957,
|
| 595 |
+
"eval_weighted_f1": 0.8553684629489016,
|
| 596 |
+
"step": 1700
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 10.147275405007363,
|
| 600 |
+
"grad_norm": 1.6638110876083374,
|
| 601 |
+
"learning_rate": 1.0954248366013074e-05,
|
| 602 |
+
"loss": 0.1581,
|
| 603 |
+
"step": 1725
|
| 604 |
+
},
|
| 605 |
+
{
|
| 606 |
+
"epoch": 10.294550810014728,
|
| 607 |
+
"grad_norm": 11.684467315673828,
|
| 608 |
+
"learning_rate": 1.0790849673202614e-05,
|
| 609 |
+
"loss": 0.2023,
|
| 610 |
+
"step": 1750
|
| 611 |
+
},
|
| 612 |
+
{
|
| 613 |
+
"epoch": 10.441826215022092,
|
| 614 |
+
"grad_norm": 0.17502157390117645,
|
| 615 |
+
"learning_rate": 1.0627450980392158e-05,
|
| 616 |
+
"loss": 0.2333,
|
| 617 |
+
"step": 1775
|
| 618 |
+
},
|
| 619 |
+
{
|
| 620 |
+
"epoch": 10.589101620029455,
|
| 621 |
+
"grad_norm": 13.039484977722168,
|
| 622 |
+
"learning_rate": 1.0464052287581701e-05,
|
| 623 |
+
"loss": 0.0847,
|
| 624 |
+
"step": 1800
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"epoch": 10.73637702503682,
|
| 628 |
+
"grad_norm": 31.663728713989258,
|
| 629 |
+
"learning_rate": 1.030065359477124e-05,
|
| 630 |
+
"loss": 0.2268,
|
| 631 |
+
"step": 1825
|
| 632 |
+
},
|
| 633 |
+
{
|
| 634 |
+
"epoch": 10.883652430044183,
|
| 635 |
+
"grad_norm": 0.15277110040187836,
|
| 636 |
+
"learning_rate": 1.0137254901960785e-05,
|
| 637 |
+
"loss": 0.2063,
|
| 638 |
+
"step": 1850
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 11.0,
|
| 642 |
+
"eval_accuracy": 0.8370883882149047,
|
| 643 |
+
"eval_loss": 0.8938233852386475,
|
| 644 |
+
"eval_macro_f1": 0.8383162808989478,
|
| 645 |
+
"eval_runtime": 0.4535,
|
| 646 |
+
"eval_samples_per_second": 1272.354,
|
| 647 |
+
"eval_steps_per_second": 81.589,
|
| 648 |
+
"eval_weighted_f1": 0.8384185272369665,
|
| 649 |
+
"step": 1870
|
| 650 |
+
},
|
| 651 |
+
{
|
| 652 |
+
"epoch": 11.029455081001473,
|
| 653 |
+
"grad_norm": 0.2940146327018738,
|
| 654 |
+
"learning_rate": 9.973856209150328e-06,
|
| 655 |
+
"loss": 0.2933,
|
| 656 |
+
"step": 1875
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 11.176730486008836,
|
| 660 |
+
"grad_norm": 0.8003905415534973,
|
| 661 |
+
"learning_rate": 9.81045751633987e-06,
|
| 662 |
+
"loss": 0.1135,
|
| 663 |
+
"step": 1900
|
| 664 |
+
},
|
| 665 |
+
{
|
| 666 |
+
"epoch": 11.324005891016201,
|
| 667 |
+
"grad_norm": 0.1670132428407669,
|
| 668 |
+
"learning_rate": 9.647058823529412e-06,
|
| 669 |
+
"loss": 0.2165,
|
| 670 |
+
"step": 1925
|
| 671 |
+
},
|
| 672 |
+
{
|
| 673 |
+
"epoch": 11.471281296023564,
|
| 674 |
+
"grad_norm": 0.14817725121974945,
|
| 675 |
+
"learning_rate": 9.483660130718955e-06,
|
| 676 |
+
"loss": 0.1218,
|
| 677 |
+
"step": 1950
|
| 678 |
+
},
|
| 679 |
+
{
|
| 680 |
+
"epoch": 11.618556701030927,
|
| 681 |
+
"grad_norm": 38.14286804199219,
|
| 682 |
+
"learning_rate": 9.320261437908498e-06,
|
| 683 |
+
"loss": 0.0698,
|
| 684 |
+
"step": 1975
|
| 685 |
+
},
|
| 686 |
+
{
|
| 687 |
+
"epoch": 11.765832106038292,
|
| 688 |
+
"grad_norm": 0.13266968727111816,
|
| 689 |
+
"learning_rate": 9.15686274509804e-06,
|
| 690 |
+
"loss": 0.1231,
|
| 691 |
+
"step": 2000
|
| 692 |
+
},
|
| 693 |
+
{
|
| 694 |
+
"epoch": 11.913107511045656,
|
| 695 |
+
"grad_norm": 0.13211187720298767,
|
| 696 |
+
"learning_rate": 8.993464052287582e-06,
|
| 697 |
+
"loss": 0.0607,
|
| 698 |
+
"step": 2025
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 12.0,
|
| 702 |
+
"eval_accuracy": 0.8509532062391681,
|
| 703 |
+
"eval_loss": 0.8488597273826599,
|
| 704 |
+
"eval_macro_f1": 0.8497052359515935,
|
| 705 |
+
"eval_runtime": 0.4287,
|
| 706 |
+
"eval_samples_per_second": 1345.907,
|
| 707 |
+
"eval_steps_per_second": 86.306,
|
| 708 |
+
"eval_weighted_f1": 0.8496494573063796,
|
| 709 |
+
"step": 2040
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 12.058910162002945,
|
| 713 |
+
"grad_norm": 2.0559182167053223,
|
| 714 |
+
"learning_rate": 8.830065359477125e-06,
|
| 715 |
+
"loss": 0.1512,
|
| 716 |
+
"step": 2050
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 12.206185567010309,
|
| 720 |
+
"grad_norm": 0.20694656670093536,
|
| 721 |
+
"learning_rate": 8.666666666666668e-06,
|
| 722 |
+
"loss": 0.1394,
|
| 723 |
+
"step": 2075
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 12.353460972017674,
|
| 727 |
+
"grad_norm": 0.4082277715206146,
|
| 728 |
+
"learning_rate": 8.503267973856209e-06,
|
| 729 |
+
"loss": 0.0929,
|
| 730 |
+
"step": 2100
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 12.500736377025037,
|
| 734 |
+
"grad_norm": 0.25394535064697266,
|
| 735 |
+
"learning_rate": 8.339869281045752e-06,
|
| 736 |
+
"loss": 0.1508,
|
| 737 |
+
"step": 2125
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 12.6480117820324,
|
| 741 |
+
"grad_norm": 0.6221270561218262,
|
| 742 |
+
"learning_rate": 8.176470588235295e-06,
|
| 743 |
+
"loss": 0.0836,
|
| 744 |
+
"step": 2150
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 12.795287187039765,
|
| 748 |
+
"grad_norm": 0.33584606647491455,
|
| 749 |
+
"learning_rate": 8.013071895424838e-06,
|
| 750 |
+
"loss": 0.0801,
|
| 751 |
+
"step": 2175
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 12.942562592047128,
|
| 755 |
+
"grad_norm": 0.11268585920333862,
|
| 756 |
+
"learning_rate": 7.849673202614379e-06,
|
| 757 |
+
"loss": 0.0953,
|
| 758 |
+
"step": 2200
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 13.0,
|
| 762 |
+
"eval_accuracy": 0.8457538994800693,
|
| 763 |
+
"eval_loss": 0.915432870388031,
|
| 764 |
+
"eval_macro_f1": 0.8461174274381698,
|
| 765 |
+
"eval_runtime": 0.4258,
|
| 766 |
+
"eval_samples_per_second": 1355.067,
|
| 767 |
+
"eval_steps_per_second": 86.893,
|
| 768 |
+
"eval_weighted_f1": 0.8460160804566235,
|
| 769 |
+
"step": 2210
|
| 770 |
+
}
|
| 771 |
+
],
|
| 772 |
+
"logging_steps": 25,
|
| 773 |
+
"max_steps": 3400,
|
| 774 |
+
"num_input_tokens_seen": 0,
|
| 775 |
+
"num_train_epochs": 20,
|
| 776 |
+
"save_steps": 500,
|
| 777 |
+
"stateful_callbacks": {
|
| 778 |
+
"EarlyStoppingCallback": {
|
| 779 |
+
"args": {
|
| 780 |
+
"early_stopping_patience": 3,
|
| 781 |
+
"early_stopping_threshold": 0.0
|
| 782 |
+
},
|
| 783 |
+
"attributes": {
|
| 784 |
+
"early_stopping_patience_counter": 3
|
| 785 |
+
}
|
| 786 |
+
},
|
| 787 |
+
"TrainerControl": {
|
| 788 |
+
"args": {
|
| 789 |
+
"should_epoch_stop": false,
|
| 790 |
+
"should_evaluate": false,
|
| 791 |
+
"should_log": false,
|
| 792 |
+
"should_save": true,
|
| 793 |
+
"should_training_stop": true
|
| 794 |
+
},
|
| 795 |
+
"attributes": {}
|
| 796 |
+
}
|
| 797 |
+
},
|
| 798 |
+
"total_flos": 0.0,
|
| 799 |
+
"train_batch_size": 8,
|
| 800 |
+
"trial_name": null,
|
| 801 |
+
"trial_params": null
|
| 802 |
+
}
|
ood-vifactcheck-phobert-topic-domain-08-s43-20260826T104414Z-bdd2c308/trainer/checkpoint-2210/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-09-s43-20260826T104827Z-bdd2c308/trainer/checkpoint-1309/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-topic-domain-09-s43-20260826T104827Z-bdd2c308/trainer/checkpoint-1309/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-09-s43-20260826T104827Z-bdd2c308/trainer/checkpoint-1309/special_tokens_map.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": "<mask>",
|
| 6 |
+
"pad_token": "<pad>",
|
| 7 |
+
"sep_token": "</s>",
|
| 8 |
+
"unk_token": "<unk>"
|
| 9 |
+
}
|
ood-vifactcheck-phobert-topic-domain-09-s43-20260826T104827Z-bdd2c308/trainer/checkpoint-1309/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"64000": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 256,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "PhobertTokenizer",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-phobert-topic-domain-09-s43-20260826T104827Z-bdd2c308/trainer/checkpoint-1309/trainer_state.json
ADDED
|
@@ -0,0 +1,484 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 748,
|
| 3 |
+
"best_metric": 0.8342195001437763,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-topic-domain-09-s43-20260826T104827Z-bdd2c308/trainer/checkpoint-748",
|
| 5 |
+
"epoch": 7.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1309,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.13404825737265416,
|
| 14 |
+
"grad_norm": 11.106708526611328,
|
| 15 |
+
"learning_rate": 1.2834224598930483e-06,
|
| 16 |
+
"loss": 4.4694,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.2680965147453083,
|
| 21 |
+
"grad_norm": 8.281728744506836,
|
| 22 |
+
"learning_rate": 2.6203208556149735e-06,
|
| 23 |
+
"loss": 4.4259,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.40214477211796246,
|
| 28 |
+
"grad_norm": 7.9853010177612305,
|
| 29 |
+
"learning_rate": 3.957219251336899e-06,
|
| 30 |
+
"loss": 4.4321,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5361930294906166,
|
| 35 |
+
"grad_norm": 8.687393188476562,
|
| 36 |
+
"learning_rate": 5.294117647058824e-06,
|
| 37 |
+
"loss": 4.4068,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6702412868632708,
|
| 42 |
+
"grad_norm": 6.5193562507629395,
|
| 43 |
+
"learning_rate": 6.631016042780749e-06,
|
| 44 |
+
"loss": 4.3373,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8042895442359249,
|
| 49 |
+
"grad_norm": 9.821297645568848,
|
| 50 |
+
"learning_rate": 7.967914438502674e-06,
|
| 51 |
+
"loss": 4.3273,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.938337801608579,
|
| 56 |
+
"grad_norm": 15.924212455749512,
|
| 57 |
+
"learning_rate": 9.3048128342246e-06,
|
| 58 |
+
"loss": 4.1514,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.5501432664756447,
|
| 64 |
+
"eval_loss": 0.9608283638954163,
|
| 65 |
+
"eval_macro_f1": 0.5340551988439312,
|
| 66 |
+
"eval_runtime": 0.4864,
|
| 67 |
+
"eval_samples_per_second": 1435.021,
|
| 68 |
+
"eval_steps_per_second": 90.46,
|
| 69 |
+
"eval_weighted_f1": 0.5372554901591585,
|
| 70 |
+
"step": 187
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0697050938337802,
|
| 74 |
+
"grad_norm": 29.98418426513672,
|
| 75 |
+
"learning_rate": 1.0641711229946525e-05,
|
| 76 |
+
"loss": 3.7826,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.2037533512064342,
|
| 81 |
+
"grad_norm": 20.26018524169922,
|
| 82 |
+
"learning_rate": 1.197860962566845e-05,
|
| 83 |
+
"loss": 3.6481,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.3378016085790885,
|
| 88 |
+
"grad_norm": 25.071537017822266,
|
| 89 |
+
"learning_rate": 1.3315508021390375e-05,
|
| 90 |
+
"loss": 3.3248,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.4718498659517425,
|
| 95 |
+
"grad_norm": 43.62504577636719,
|
| 96 |
+
"learning_rate": 1.4652406417112302e-05,
|
| 97 |
+
"loss": 2.9501,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.6058981233243967,
|
| 102 |
+
"grad_norm": 44.7548942565918,
|
| 103 |
+
"learning_rate": 1.5989304812834226e-05,
|
| 104 |
+
"loss": 2.782,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.739946380697051,
|
| 109 |
+
"grad_norm": 30.186731338500977,
|
| 110 |
+
"learning_rate": 1.7326203208556153e-05,
|
| 111 |
+
"loss": 2.379,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.8739946380697052,
|
| 116 |
+
"grad_norm": 57.74986267089844,
|
| 117 |
+
"learning_rate": 1.8663101604278075e-05,
|
| 118 |
+
"loss": 2.3917,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.8008595988538681,
|
| 124 |
+
"eval_loss": 0.5468553900718689,
|
| 125 |
+
"eval_macro_f1": 0.7998930731413584,
|
| 126 |
+
"eval_runtime": 0.4757,
|
| 127 |
+
"eval_samples_per_second": 1467.393,
|
| 128 |
+
"eval_steps_per_second": 92.5,
|
| 129 |
+
"eval_weighted_f1": 0.7996827839704836,
|
| 130 |
+
"step": 374
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.005361930294906,
|
| 134 |
+
"grad_norm": 29.33355712890625,
|
| 135 |
+
"learning_rate": 2e-05,
|
| 136 |
+
"loss": 2.0973,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.1394101876675604,
|
| 141 |
+
"grad_norm": 38.9906120300293,
|
| 142 |
+
"learning_rate": 1.9851455733808676e-05,
|
| 143 |
+
"loss": 1.7304,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.2734584450402147,
|
| 148 |
+
"grad_norm": 22.55340003967285,
|
| 149 |
+
"learning_rate": 1.9702911467617353e-05,
|
| 150 |
+
"loss": 1.7385,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.4075067024128685,
|
| 155 |
+
"grad_norm": 34.23171615600586,
|
| 156 |
+
"learning_rate": 1.9554367201426024e-05,
|
| 157 |
+
"loss": 1.7645,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.5415549597855227,
|
| 162 |
+
"grad_norm": 29.3786678314209,
|
| 163 |
+
"learning_rate": 1.94058229352347e-05,
|
| 164 |
+
"loss": 1.6072,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.675603217158177,
|
| 169 |
+
"grad_norm": 38.421531677246094,
|
| 170 |
+
"learning_rate": 1.9257278669043376e-05,
|
| 171 |
+
"loss": 1.5397,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.809651474530831,
|
| 176 |
+
"grad_norm": 41.767757415771484,
|
| 177 |
+
"learning_rate": 1.9108734402852053e-05,
|
| 178 |
+
"loss": 1.6584,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 2.943699731903485,
|
| 183 |
+
"grad_norm": 84.88451385498047,
|
| 184 |
+
"learning_rate": 1.8960190136660727e-05,
|
| 185 |
+
"loss": 1.6798,
|
| 186 |
+
"step": 550
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 3.0,
|
| 190 |
+
"eval_accuracy": 0.8094555873925502,
|
| 191 |
+
"eval_loss": 0.5263932347297668,
|
| 192 |
+
"eval_macro_f1": 0.8066628544230484,
|
| 193 |
+
"eval_runtime": 0.4881,
|
| 194 |
+
"eval_samples_per_second": 1430.105,
|
| 195 |
+
"eval_steps_per_second": 90.15,
|
| 196 |
+
"eval_weighted_f1": 0.8070424208942077,
|
| 197 |
+
"step": 561
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.0750670241286864,
|
| 201 |
+
"grad_norm": 27.140485763549805,
|
| 202 |
+
"learning_rate": 1.88116458704694e-05,
|
| 203 |
+
"loss": 1.2582,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.2091152815013406,
|
| 208 |
+
"grad_norm": 52.028648376464844,
|
| 209 |
+
"learning_rate": 1.8663101604278075e-05,
|
| 210 |
+
"loss": 1.0739,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.343163538873995,
|
| 215 |
+
"grad_norm": 38.068023681640625,
|
| 216 |
+
"learning_rate": 1.8514557338086753e-05,
|
| 217 |
+
"loss": 1.0662,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.4772117962466487,
|
| 222 |
+
"grad_norm": 36.84400939941406,
|
| 223 |
+
"learning_rate": 1.8366013071895427e-05,
|
| 224 |
+
"loss": 1.3072,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.611260053619303,
|
| 229 |
+
"grad_norm": 33.67017364501953,
|
| 230 |
+
"learning_rate": 1.82174688057041e-05,
|
| 231 |
+
"loss": 1.1568,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.745308310991957,
|
| 236 |
+
"grad_norm": 44.019596099853516,
|
| 237 |
+
"learning_rate": 1.8068924539512775e-05,
|
| 238 |
+
"loss": 1.0038,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.8793565683646114,
|
| 243 |
+
"grad_norm": 37.88417434692383,
|
| 244 |
+
"learning_rate": 1.7920380273321453e-05,
|
| 245 |
+
"loss": 0.9625,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 4.0,
|
| 250 |
+
"eval_accuracy": 0.835243553008596,
|
| 251 |
+
"eval_loss": 0.5001320242881775,
|
| 252 |
+
"eval_macro_f1": 0.8342195001437763,
|
| 253 |
+
"eval_runtime": 0.425,
|
| 254 |
+
"eval_samples_per_second": 1642.196,
|
| 255 |
+
"eval_steps_per_second": 103.519,
|
| 256 |
+
"eval_weighted_f1": 0.8345167191888017,
|
| 257 |
+
"step": 748
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.010723860589812,
|
| 261 |
+
"grad_norm": 58.05527877807617,
|
| 262 |
+
"learning_rate": 1.7771836007130127e-05,
|
| 263 |
+
"loss": 1.0919,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.144772117962466,
|
| 268 |
+
"grad_norm": 50.66020965576172,
|
| 269 |
+
"learning_rate": 1.76232917409388e-05,
|
| 270 |
+
"loss": 0.7012,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.278820375335121,
|
| 275 |
+
"grad_norm": 20.69614028930664,
|
| 276 |
+
"learning_rate": 1.7474747474747475e-05,
|
| 277 |
+
"loss": 0.7779,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.412868632707775,
|
| 282 |
+
"grad_norm": 35.349082946777344,
|
| 283 |
+
"learning_rate": 1.7326203208556153e-05,
|
| 284 |
+
"loss": 0.8779,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.546916890080429,
|
| 289 |
+
"grad_norm": 148.05319213867188,
|
| 290 |
+
"learning_rate": 1.7177658942364827e-05,
|
| 291 |
+
"loss": 0.6976,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.680965147453083,
|
| 296 |
+
"grad_norm": 14.235578536987305,
|
| 297 |
+
"learning_rate": 1.70291146761735e-05,
|
| 298 |
+
"loss": 0.8061,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.815013404825737,
|
| 303 |
+
"grad_norm": 77.76993560791016,
|
| 304 |
+
"learning_rate": 1.6880570409982175e-05,
|
| 305 |
+
"loss": 0.7318,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 4.949061662198392,
|
| 310 |
+
"grad_norm": 8.587902069091797,
|
| 311 |
+
"learning_rate": 1.6732026143790852e-05,
|
| 312 |
+
"loss": 0.7936,
|
| 313 |
+
"step": 925
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 5.0,
|
| 317 |
+
"eval_accuracy": 0.8065902578796562,
|
| 318 |
+
"eval_loss": 0.6719850897789001,
|
| 319 |
+
"eval_macro_f1": 0.8065017174692893,
|
| 320 |
+
"eval_runtime": 0.4795,
|
| 321 |
+
"eval_samples_per_second": 1455.798,
|
| 322 |
+
"eval_steps_per_second": 91.769,
|
| 323 |
+
"eval_weighted_f1": 0.8071232687907494,
|
| 324 |
+
"step": 935
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.080428954423592,
|
| 328 |
+
"grad_norm": 171.80938720703125,
|
| 329 |
+
"learning_rate": 1.6583481877599526e-05,
|
| 330 |
+
"loss": 0.8819,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.214477211796247,
|
| 335 |
+
"grad_norm": 18.913583755493164,
|
| 336 |
+
"learning_rate": 1.64349376114082e-05,
|
| 337 |
+
"loss": 0.5116,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.348525469168901,
|
| 342 |
+
"grad_norm": 11.668533325195312,
|
| 343 |
+
"learning_rate": 1.6286393345216875e-05,
|
| 344 |
+
"loss": 0.5832,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.482573726541555,
|
| 349 |
+
"grad_norm": 31.680864334106445,
|
| 350 |
+
"learning_rate": 1.6137849079025552e-05,
|
| 351 |
+
"loss": 0.749,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.616621983914209,
|
| 356 |
+
"grad_norm": 78.5151138305664,
|
| 357 |
+
"learning_rate": 1.5989304812834226e-05,
|
| 358 |
+
"loss": 0.5804,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.750670241286863,
|
| 363 |
+
"grad_norm": 25.79216766357422,
|
| 364 |
+
"learning_rate": 1.58407605466429e-05,
|
| 365 |
+
"loss": 0.7498,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 5.884718498659518,
|
| 370 |
+
"grad_norm": 1.9176758527755737,
|
| 371 |
+
"learning_rate": 1.5692216280451574e-05,
|
| 372 |
+
"loss": 0.6125,
|
| 373 |
+
"step": 1100
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 6.0,
|
| 377 |
+
"eval_accuracy": 0.833810888252149,
|
| 378 |
+
"eval_loss": 0.6656188368797302,
|
| 379 |
+
"eval_macro_f1": 0.8329946176688939,
|
| 380 |
+
"eval_runtime": 0.4705,
|
| 381 |
+
"eval_samples_per_second": 1483.682,
|
| 382 |
+
"eval_steps_per_second": 93.527,
|
| 383 |
+
"eval_weighted_f1": 0.8332698897901097,
|
| 384 |
+
"step": 1122
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 6.016085790884718,
|
| 388 |
+
"grad_norm": 3.862088918685913,
|
| 389 |
+
"learning_rate": 1.5543672014260252e-05,
|
| 390 |
+
"loss": 0.3832,
|
| 391 |
+
"step": 1125
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 6.150134048257373,
|
| 395 |
+
"grad_norm": 1.2668904066085815,
|
| 396 |
+
"learning_rate": 1.5395127748068926e-05,
|
| 397 |
+
"loss": 0.4329,
|
| 398 |
+
"step": 1150
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.2841823056300266,
|
| 402 |
+
"grad_norm": 36.3499641418457,
|
| 403 |
+
"learning_rate": 1.5246583481877602e-05,
|
| 404 |
+
"loss": 0.5255,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.418230563002681,
|
| 409 |
+
"grad_norm": 29.60820960998535,
|
| 410 |
+
"learning_rate": 1.5098039215686276e-05,
|
| 411 |
+
"loss": 0.376,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.552278820375335,
|
| 416 |
+
"grad_norm": 9.774564743041992,
|
| 417 |
+
"learning_rate": 1.4949494949494952e-05,
|
| 418 |
+
"loss": 0.311,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.68632707774799,
|
| 423 |
+
"grad_norm": 25.40093994140625,
|
| 424 |
+
"learning_rate": 1.4800950683303626e-05,
|
| 425 |
+
"loss": 0.4741,
|
| 426 |
+
"step": 1250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 6.8203753351206435,
|
| 430 |
+
"grad_norm": 4.406991958618164,
|
| 431 |
+
"learning_rate": 1.4652406417112302e-05,
|
| 432 |
+
"loss": 0.3515,
|
| 433 |
+
"step": 1275
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 6.954423592493297,
|
| 437 |
+
"grad_norm": 5.418511867523193,
|
| 438 |
+
"learning_rate": 1.4503862150920976e-05,
|
| 439 |
+
"loss": 0.501,
|
| 440 |
+
"step": 1300
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 7.0,
|
| 444 |
+
"eval_accuracy": 0.8223495702005731,
|
| 445 |
+
"eval_loss": 0.7991241812705994,
|
| 446 |
+
"eval_macro_f1": 0.8201864034045121,
|
| 447 |
+
"eval_runtime": 0.4998,
|
| 448 |
+
"eval_samples_per_second": 1396.576,
|
| 449 |
+
"eval_steps_per_second": 88.036,
|
| 450 |
+
"eval_weighted_f1": 0.820905824271272,
|
| 451 |
+
"step": 1309
|
| 452 |
+
}
|
| 453 |
+
],
|
| 454 |
+
"logging_steps": 25,
|
| 455 |
+
"max_steps": 3740,
|
| 456 |
+
"num_input_tokens_seen": 0,
|
| 457 |
+
"num_train_epochs": 20,
|
| 458 |
+
"save_steps": 500,
|
| 459 |
+
"stateful_callbacks": {
|
| 460 |
+
"EarlyStoppingCallback": {
|
| 461 |
+
"args": {
|
| 462 |
+
"early_stopping_patience": 3,
|
| 463 |
+
"early_stopping_threshold": 0.0
|
| 464 |
+
},
|
| 465 |
+
"attributes": {
|
| 466 |
+
"early_stopping_patience_counter": 3
|
| 467 |
+
}
|
| 468 |
+
},
|
| 469 |
+
"TrainerControl": {
|
| 470 |
+
"args": {
|
| 471 |
+
"should_epoch_stop": false,
|
| 472 |
+
"should_evaluate": false,
|
| 473 |
+
"should_log": false,
|
| 474 |
+
"should_save": true,
|
| 475 |
+
"should_training_stop": true
|
| 476 |
+
},
|
| 477 |
+
"attributes": {}
|
| 478 |
+
}
|
| 479 |
+
},
|
| 480 |
+
"total_flos": 0.0,
|
| 481 |
+
"train_batch_size": 8,
|
| 482 |
+
"trial_name": null,
|
| 483 |
+
"trial_params": null
|
| 484 |
+
}
|
ood-vifactcheck-phobert-topic-domain-09-s43-20260826T104827Z-bdd2c308/trainer/checkpoint-1309/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-09-s43-20260826T104827Z-bdd2c308/trainer/checkpoint-748/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-topic-domain-09-s43-20260826T104827Z-bdd2c308/trainer/checkpoint-748/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-09-s43-20260826T104827Z-bdd2c308/trainer/checkpoint-748/special_tokens_map.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": "<mask>",
|
| 6 |
+
"pad_token": "<pad>",
|
| 7 |
+
"sep_token": "</s>",
|
| 8 |
+
"unk_token": "<unk>"
|
| 9 |
+
}
|
ood-vifactcheck-phobert-topic-domain-09-s43-20260826T104827Z-bdd2c308/trainer/checkpoint-748/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"64000": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 256,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "PhobertTokenizer",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-phobert-topic-domain-09-s43-20260826T104827Z-bdd2c308/trainer/checkpoint-748/trainer_state.json
ADDED
|
@@ -0,0 +1,290 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 748,
|
| 3 |
+
"best_metric": 0.8342195001437763,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-topic-domain-09-s43-20260826T104827Z-bdd2c308/trainer/checkpoint-748",
|
| 5 |
+
"epoch": 4.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 748,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.13404825737265416,
|
| 14 |
+
"grad_norm": 11.106708526611328,
|
| 15 |
+
"learning_rate": 1.2834224598930483e-06,
|
| 16 |
+
"loss": 4.4694,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.2680965147453083,
|
| 21 |
+
"grad_norm": 8.281728744506836,
|
| 22 |
+
"learning_rate": 2.6203208556149735e-06,
|
| 23 |
+
"loss": 4.4259,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.40214477211796246,
|
| 28 |
+
"grad_norm": 7.9853010177612305,
|
| 29 |
+
"learning_rate": 3.957219251336899e-06,
|
| 30 |
+
"loss": 4.4321,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5361930294906166,
|
| 35 |
+
"grad_norm": 8.687393188476562,
|
| 36 |
+
"learning_rate": 5.294117647058824e-06,
|
| 37 |
+
"loss": 4.4068,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6702412868632708,
|
| 42 |
+
"grad_norm": 6.5193562507629395,
|
| 43 |
+
"learning_rate": 6.631016042780749e-06,
|
| 44 |
+
"loss": 4.3373,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8042895442359249,
|
| 49 |
+
"grad_norm": 9.821297645568848,
|
| 50 |
+
"learning_rate": 7.967914438502674e-06,
|
| 51 |
+
"loss": 4.3273,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.938337801608579,
|
| 56 |
+
"grad_norm": 15.924212455749512,
|
| 57 |
+
"learning_rate": 9.3048128342246e-06,
|
| 58 |
+
"loss": 4.1514,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.5501432664756447,
|
| 64 |
+
"eval_loss": 0.9608283638954163,
|
| 65 |
+
"eval_macro_f1": 0.5340551988439312,
|
| 66 |
+
"eval_runtime": 0.4864,
|
| 67 |
+
"eval_samples_per_second": 1435.021,
|
| 68 |
+
"eval_steps_per_second": 90.46,
|
| 69 |
+
"eval_weighted_f1": 0.5372554901591585,
|
| 70 |
+
"step": 187
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0697050938337802,
|
| 74 |
+
"grad_norm": 29.98418426513672,
|
| 75 |
+
"learning_rate": 1.0641711229946525e-05,
|
| 76 |
+
"loss": 3.7826,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.2037533512064342,
|
| 81 |
+
"grad_norm": 20.26018524169922,
|
| 82 |
+
"learning_rate": 1.197860962566845e-05,
|
| 83 |
+
"loss": 3.6481,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.3378016085790885,
|
| 88 |
+
"grad_norm": 25.071537017822266,
|
| 89 |
+
"learning_rate": 1.3315508021390375e-05,
|
| 90 |
+
"loss": 3.3248,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.4718498659517425,
|
| 95 |
+
"grad_norm": 43.62504577636719,
|
| 96 |
+
"learning_rate": 1.4652406417112302e-05,
|
| 97 |
+
"loss": 2.9501,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.6058981233243967,
|
| 102 |
+
"grad_norm": 44.7548942565918,
|
| 103 |
+
"learning_rate": 1.5989304812834226e-05,
|
| 104 |
+
"loss": 2.782,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.739946380697051,
|
| 109 |
+
"grad_norm": 30.186731338500977,
|
| 110 |
+
"learning_rate": 1.7326203208556153e-05,
|
| 111 |
+
"loss": 2.379,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.8739946380697052,
|
| 116 |
+
"grad_norm": 57.74986267089844,
|
| 117 |
+
"learning_rate": 1.8663101604278075e-05,
|
| 118 |
+
"loss": 2.3917,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.8008595988538681,
|
| 124 |
+
"eval_loss": 0.5468553900718689,
|
| 125 |
+
"eval_macro_f1": 0.7998930731413584,
|
| 126 |
+
"eval_runtime": 0.4757,
|
| 127 |
+
"eval_samples_per_second": 1467.393,
|
| 128 |
+
"eval_steps_per_second": 92.5,
|
| 129 |
+
"eval_weighted_f1": 0.7996827839704836,
|
| 130 |
+
"step": 374
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.005361930294906,
|
| 134 |
+
"grad_norm": 29.33355712890625,
|
| 135 |
+
"learning_rate": 2e-05,
|
| 136 |
+
"loss": 2.0973,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.1394101876675604,
|
| 141 |
+
"grad_norm": 38.9906120300293,
|
| 142 |
+
"learning_rate": 1.9851455733808676e-05,
|
| 143 |
+
"loss": 1.7304,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.2734584450402147,
|
| 148 |
+
"grad_norm": 22.55340003967285,
|
| 149 |
+
"learning_rate": 1.9702911467617353e-05,
|
| 150 |
+
"loss": 1.7385,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.4075067024128685,
|
| 155 |
+
"grad_norm": 34.23171615600586,
|
| 156 |
+
"learning_rate": 1.9554367201426024e-05,
|
| 157 |
+
"loss": 1.7645,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.5415549597855227,
|
| 162 |
+
"grad_norm": 29.3786678314209,
|
| 163 |
+
"learning_rate": 1.94058229352347e-05,
|
| 164 |
+
"loss": 1.6072,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.675603217158177,
|
| 169 |
+
"grad_norm": 38.421531677246094,
|
| 170 |
+
"learning_rate": 1.9257278669043376e-05,
|
| 171 |
+
"loss": 1.5397,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.809651474530831,
|
| 176 |
+
"grad_norm": 41.767757415771484,
|
| 177 |
+
"learning_rate": 1.9108734402852053e-05,
|
| 178 |
+
"loss": 1.6584,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 2.943699731903485,
|
| 183 |
+
"grad_norm": 84.88451385498047,
|
| 184 |
+
"learning_rate": 1.8960190136660727e-05,
|
| 185 |
+
"loss": 1.6798,
|
| 186 |
+
"step": 550
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 3.0,
|
| 190 |
+
"eval_accuracy": 0.8094555873925502,
|
| 191 |
+
"eval_loss": 0.5263932347297668,
|
| 192 |
+
"eval_macro_f1": 0.8066628544230484,
|
| 193 |
+
"eval_runtime": 0.4881,
|
| 194 |
+
"eval_samples_per_second": 1430.105,
|
| 195 |
+
"eval_steps_per_second": 90.15,
|
| 196 |
+
"eval_weighted_f1": 0.8070424208942077,
|
| 197 |
+
"step": 561
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.0750670241286864,
|
| 201 |
+
"grad_norm": 27.140485763549805,
|
| 202 |
+
"learning_rate": 1.88116458704694e-05,
|
| 203 |
+
"loss": 1.2582,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.2091152815013406,
|
| 208 |
+
"grad_norm": 52.028648376464844,
|
| 209 |
+
"learning_rate": 1.8663101604278075e-05,
|
| 210 |
+
"loss": 1.0739,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.343163538873995,
|
| 215 |
+
"grad_norm": 38.068023681640625,
|
| 216 |
+
"learning_rate": 1.8514557338086753e-05,
|
| 217 |
+
"loss": 1.0662,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.4772117962466487,
|
| 222 |
+
"grad_norm": 36.84400939941406,
|
| 223 |
+
"learning_rate": 1.8366013071895427e-05,
|
| 224 |
+
"loss": 1.3072,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.611260053619303,
|
| 229 |
+
"grad_norm": 33.67017364501953,
|
| 230 |
+
"learning_rate": 1.82174688057041e-05,
|
| 231 |
+
"loss": 1.1568,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.745308310991957,
|
| 236 |
+
"grad_norm": 44.019596099853516,
|
| 237 |
+
"learning_rate": 1.8068924539512775e-05,
|
| 238 |
+
"loss": 1.0038,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.8793565683646114,
|
| 243 |
+
"grad_norm": 37.88417434692383,
|
| 244 |
+
"learning_rate": 1.7920380273321453e-05,
|
| 245 |
+
"loss": 0.9625,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 4.0,
|
| 250 |
+
"eval_accuracy": 0.835243553008596,
|
| 251 |
+
"eval_loss": 0.5001320242881775,
|
| 252 |
+
"eval_macro_f1": 0.8342195001437763,
|
| 253 |
+
"eval_runtime": 0.425,
|
| 254 |
+
"eval_samples_per_second": 1642.196,
|
| 255 |
+
"eval_steps_per_second": 103.519,
|
| 256 |
+
"eval_weighted_f1": 0.8345167191888017,
|
| 257 |
+
"step": 748
|
| 258 |
+
}
|
| 259 |
+
],
|
| 260 |
+
"logging_steps": 25,
|
| 261 |
+
"max_steps": 3740,
|
| 262 |
+
"num_input_tokens_seen": 0,
|
| 263 |
+
"num_train_epochs": 20,
|
| 264 |
+
"save_steps": 500,
|
| 265 |
+
"stateful_callbacks": {
|
| 266 |
+
"EarlyStoppingCallback": {
|
| 267 |
+
"args": {
|
| 268 |
+
"early_stopping_patience": 3,
|
| 269 |
+
"early_stopping_threshold": 0.0
|
| 270 |
+
},
|
| 271 |
+
"attributes": {
|
| 272 |
+
"early_stopping_patience_counter": 0
|
| 273 |
+
}
|
| 274 |
+
},
|
| 275 |
+
"TrainerControl": {
|
| 276 |
+
"args": {
|
| 277 |
+
"should_epoch_stop": false,
|
| 278 |
+
"should_evaluate": false,
|
| 279 |
+
"should_log": false,
|
| 280 |
+
"should_save": true,
|
| 281 |
+
"should_training_stop": false
|
| 282 |
+
},
|
| 283 |
+
"attributes": {}
|
| 284 |
+
}
|
| 285 |
+
},
|
| 286 |
+
"total_flos": 0.0,
|
| 287 |
+
"train_batch_size": 8,
|
| 288 |
+
"trial_name": null,
|
| 289 |
+
"trial_params": null
|
| 290 |
+
}
|
ood-vifactcheck-phobert-topic-domain-09-s43-20260826T104827Z-bdd2c308/trainer/checkpoint-748/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-10-s43-20260826T105057Z-bdd2c308/trainer/checkpoint-1158/special_tokens_map.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": "<mask>",
|
| 6 |
+
"pad_token": "<pad>",
|
| 7 |
+
"sep_token": "</s>",
|
| 8 |
+
"unk_token": "<unk>"
|
| 9 |
+
}
|
ood-vifactcheck-phobert-topic-domain-10-s43-20260826T105057Z-bdd2c308/trainer/checkpoint-1158/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 |
+
}
|