Add files using upload-large-folder tool
Browse filesThis view is limited to 50 files because it contains too many changes. See raw diff
- ood-vifactcheck-phobert-topic-domain-00-s44-20260826T110330Z-ab4112e3/trainer/checkpoint-1122/added_tokens.json +3 -0
- ood-vifactcheck-phobert-topic-domain-00-s44-20260826T110330Z-ab4112e3/trainer/checkpoint-1122/bpe.codes +0 -0
- ood-vifactcheck-phobert-topic-domain-00-s44-20260826T110330Z-ab4112e3/trainer/checkpoint-1122/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-topic-domain-00-s44-20260826T110330Z-ab4112e3/trainer/checkpoint-1122/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-topic-domain-00-s44-20260826T110330Z-ab4112e3/trainer/checkpoint-1122/trainer_state.json +417 -0
- ood-vifactcheck-phobert-topic-domain-00-s44-20260826T110330Z-ab4112e3/trainer/checkpoint-1122/vocab.txt +0 -0
- ood-vifactcheck-phobert-topic-domain-00-s44-20260826T110330Z-ab4112e3/trainer/checkpoint-561/added_tokens.json +3 -0
- ood-vifactcheck-phobert-topic-domain-00-s44-20260826T110330Z-ab4112e3/trainer/checkpoint-561/bpe.codes +0 -0
- ood-vifactcheck-phobert-topic-domain-00-s44-20260826T110330Z-ab4112e3/trainer/checkpoint-561/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-topic-domain-00-s44-20260826T110330Z-ab4112e3/trainer/checkpoint-561/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-topic-domain-00-s44-20260826T110330Z-ab4112e3/trainer/checkpoint-561/trainer_state.json +230 -0
- ood-vifactcheck-phobert-topic-domain-00-s44-20260826T110330Z-ab4112e3/trainer/checkpoint-561/vocab.txt +0 -0
- ood-vifactcheck-phobert-topic-domain-01-s44-20260826T110548Z-ab4112e3/trainer/checkpoint-1337/added_tokens.json +3 -0
- ood-vifactcheck-phobert-topic-domain-01-s44-20260826T110548Z-ab4112e3/trainer/checkpoint-1337/bpe.codes +0 -0
- ood-vifactcheck-phobert-topic-domain-01-s44-20260826T110548Z-ab4112e3/trainer/checkpoint-1337/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-topic-domain-01-s44-20260826T110548Z-ab4112e3/trainer/checkpoint-1337/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-topic-domain-01-s44-20260826T110548Z-ab4112e3/trainer/checkpoint-1337/trainer_state.json +491 -0
- ood-vifactcheck-phobert-topic-domain-01-s44-20260826T110548Z-ab4112e3/trainer/checkpoint-1337/vocab.txt +0 -0
- ood-vifactcheck-phobert-topic-domain-01-s44-20260826T110548Z-ab4112e3/trainer/checkpoint-1910/added_tokens.json +3 -0
- ood-vifactcheck-phobert-topic-domain-01-s44-20260826T110548Z-ab4112e3/trainer/checkpoint-1910/bpe.codes +0 -0
- ood-vifactcheck-phobert-topic-domain-01-s44-20260826T110548Z-ab4112e3/trainer/checkpoint-1910/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-topic-domain-01-s44-20260826T110548Z-ab4112e3/trainer/checkpoint-1910/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-topic-domain-01-s44-20260826T110548Z-ab4112e3/trainer/checkpoint-1910/trainer_state.json +685 -0
- ood-vifactcheck-phobert-topic-domain-01-s44-20260826T110548Z-ab4112e3/trainer/checkpoint-1910/vocab.txt +0 -0
- ood-vifactcheck-phobert-topic-domain-02-s44-20260826T110927Z-ab4112e3/trainer/checkpoint-1274/added_tokens.json +3 -0
- ood-vifactcheck-phobert-topic-domain-02-s44-20260826T110927Z-ab4112e3/trainer/checkpoint-1274/bpe.codes +0 -0
- ood-vifactcheck-phobert-topic-domain-02-s44-20260826T110927Z-ab4112e3/trainer/checkpoint-1274/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-topic-domain-02-s44-20260826T110927Z-ab4112e3/trainer/checkpoint-1274/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-topic-domain-02-s44-20260826T110927Z-ab4112e3/trainer/checkpoint-1274/trainer_state.json +470 -0
- ood-vifactcheck-phobert-topic-domain-02-s44-20260826T110927Z-ab4112e3/trainer/checkpoint-1274/vocab.txt +0 -0
- ood-vifactcheck-phobert-topic-domain-02-s44-20260826T110927Z-ab4112e3/trainer/checkpoint-728/added_tokens.json +3 -0
- ood-vifactcheck-phobert-topic-domain-02-s44-20260826T110927Z-ab4112e3/trainer/checkpoint-728/bpe.codes +0 -0
- ood-vifactcheck-phobert-topic-domain-02-s44-20260826T110927Z-ab4112e3/trainer/checkpoint-728/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-topic-domain-02-s44-20260826T110927Z-ab4112e3/trainer/checkpoint-728/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-topic-domain-02-s44-20260826T110927Z-ab4112e3/trainer/checkpoint-728/trainer_state.json +290 -0
- ood-vifactcheck-phobert-topic-domain-02-s44-20260826T110927Z-ab4112e3/trainer/checkpoint-728/vocab.txt +0 -0
- ood-vifactcheck-phobert-topic-domain-03-s44-20260826T111158Z-ab4112e3/trainer/checkpoint-1170/added_tokens.json +3 -0
- ood-vifactcheck-phobert-topic-domain-03-s44-20260826T111158Z-ab4112e3/trainer/checkpoint-1170/bpe.codes +0 -0
- ood-vifactcheck-phobert-topic-domain-03-s44-20260826T111158Z-ab4112e3/trainer/checkpoint-1170/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-topic-domain-03-s44-20260826T111158Z-ab4112e3/trainer/checkpoint-1170/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-topic-domain-03-s44-20260826T111158Z-ab4112e3/trainer/checkpoint-1170/trainer_state.json +431 -0
- ood-vifactcheck-phobert-topic-domain-03-s44-20260826T111158Z-ab4112e3/trainer/checkpoint-1170/vocab.txt +0 -0
- ood-vifactcheck-phobert-topic-domain-03-s44-20260826T111158Z-ab4112e3/trainer/checkpoint-1755/added_tokens.json +3 -0
- ood-vifactcheck-phobert-topic-domain-03-s44-20260826T111158Z-ab4112e3/trainer/checkpoint-1755/bpe.codes +0 -0
- ood-vifactcheck-phobert-topic-domain-03-s44-20260826T111158Z-ab4112e3/trainer/checkpoint-1755/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-topic-domain-03-s44-20260826T111158Z-ab4112e3/trainer/checkpoint-1755/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-topic-domain-03-s44-20260826T111158Z-ab4112e3/trainer/checkpoint-1755/trainer_state.json +632 -0
- ood-vifactcheck-phobert-topic-domain-03-s44-20260826T111158Z-ab4112e3/trainer/checkpoint-1755/vocab.txt +0 -0
- ood-vifactcheck-phobert-topic-domain-04-s44-20260826T111516Z-ab4112e3/trainer/checkpoint-1316/added_tokens.json +3 -0
- ood-vifactcheck-phobert-topic-domain-04-s44-20260826T111516Z-ab4112e3/trainer/checkpoint-1316/bpe.codes +0 -0
ood-vifactcheck-phobert-topic-domain-00-s44-20260826T110330Z-ab4112e3/trainer/checkpoint-1122/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-topic-domain-00-s44-20260826T110330Z-ab4112e3/trainer/checkpoint-1122/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-00-s44-20260826T110330Z-ab4112e3/trainer/checkpoint-1122/special_tokens_map.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": "<mask>",
|
| 6 |
+
"pad_token": "<pad>",
|
| 7 |
+
"sep_token": "</s>",
|
| 8 |
+
"unk_token": "<unk>"
|
| 9 |
+
}
|
ood-vifactcheck-phobert-topic-domain-00-s44-20260826T110330Z-ab4112e3/trainer/checkpoint-1122/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"64000": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 256,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "PhobertTokenizer",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-phobert-topic-domain-00-s44-20260826T110330Z-ab4112e3/trainer/checkpoint-1122/trainer_state.json
ADDED
|
@@ -0,0 +1,417 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 561,
|
| 3 |
+
"best_metric": 0.8558487435825608,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-topic-domain-00-s44-20260826T110330Z-ab4112e3/trainer/checkpoint-561",
|
| 5 |
+
"epoch": 6.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1122,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.13368983957219252,
|
| 14 |
+
"grad_norm": 8.93509578704834,
|
| 15 |
+
"learning_rate": 1.2834224598930483e-06,
|
| 16 |
+
"loss": 4.4108,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.26737967914438504,
|
| 21 |
+
"grad_norm": 9.546072959899902,
|
| 22 |
+
"learning_rate": 2.6203208556149735e-06,
|
| 23 |
+
"loss": 4.3762,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.40106951871657753,
|
| 28 |
+
"grad_norm": 10.253655433654785,
|
| 29 |
+
"learning_rate": 3.957219251336899e-06,
|
| 30 |
+
"loss": 4.3644,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5347593582887701,
|
| 35 |
+
"grad_norm": 6.676309108734131,
|
| 36 |
+
"learning_rate": 5.294117647058824e-06,
|
| 37 |
+
"loss": 4.3382,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6684491978609626,
|
| 42 |
+
"grad_norm": 10.954312324523926,
|
| 43 |
+
"learning_rate": 6.631016042780749e-06,
|
| 44 |
+
"loss": 4.296,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8021390374331551,
|
| 49 |
+
"grad_norm": 10.327520370483398,
|
| 50 |
+
"learning_rate": 7.967914438502674e-06,
|
| 51 |
+
"loss": 4.1346,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9358288770053476,
|
| 56 |
+
"grad_norm": 21.59717559814453,
|
| 57 |
+
"learning_rate": 9.3048128342246e-06,
|
| 58 |
+
"loss": 3.9314,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.6258741258741258,
|
| 64 |
+
"eval_loss": 0.8830139636993408,
|
| 65 |
+
"eval_macro_f1": 0.6177260306522513,
|
| 66 |
+
"eval_runtime": 0.3903,
|
| 67 |
+
"eval_samples_per_second": 1465.711,
|
| 68 |
+
"eval_steps_per_second": 92.248,
|
| 69 |
+
"eval_weighted_f1": 0.6201395610448718,
|
| 70 |
+
"step": 187
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0695187165775402,
|
| 74 |
+
"grad_norm": 28.900501251220703,
|
| 75 |
+
"learning_rate": 1.0641711229946525e-05,
|
| 76 |
+
"loss": 3.6519,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.2032085561497325,
|
| 81 |
+
"grad_norm": 41.45104217529297,
|
| 82 |
+
"learning_rate": 1.197860962566845e-05,
|
| 83 |
+
"loss": 3.4209,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.3368983957219251,
|
| 88 |
+
"grad_norm": 24.274688720703125,
|
| 89 |
+
"learning_rate": 1.3315508021390375e-05,
|
| 90 |
+
"loss": 3.0569,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.4705882352941178,
|
| 95 |
+
"grad_norm": 31.610830307006836,
|
| 96 |
+
"learning_rate": 1.4652406417112302e-05,
|
| 97 |
+
"loss": 2.8981,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.6042780748663101,
|
| 102 |
+
"grad_norm": 36.086883544921875,
|
| 103 |
+
"learning_rate": 1.5989304812834226e-05,
|
| 104 |
+
"loss": 2.6154,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7379679144385025,
|
| 109 |
+
"grad_norm": 58.68095016479492,
|
| 110 |
+
"learning_rate": 1.7326203208556153e-05,
|
| 111 |
+
"loss": 2.356,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.8716577540106951,
|
| 116 |
+
"grad_norm": 33.54307556152344,
|
| 117 |
+
"learning_rate": 1.8663101604278075e-05,
|
| 118 |
+
"loss": 2.4054,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.8251748251748252,
|
| 124 |
+
"eval_loss": 0.4848474860191345,
|
| 125 |
+
"eval_macro_f1": 0.825687690742074,
|
| 126 |
+
"eval_runtime": 0.4188,
|
| 127 |
+
"eval_samples_per_second": 1365.72,
|
| 128 |
+
"eval_steps_per_second": 85.954,
|
| 129 |
+
"eval_weighted_f1": 0.8252022958446541,
|
| 130 |
+
"step": 374
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.0053475935828877,
|
| 134 |
+
"grad_norm": 29.90228271484375,
|
| 135 |
+
"learning_rate": 2e-05,
|
| 136 |
+
"loss": 2.1596,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.1390374331550803,
|
| 141 |
+
"grad_norm": 33.18424987792969,
|
| 142 |
+
"learning_rate": 1.9851455733808676e-05,
|
| 143 |
+
"loss": 1.572,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.2727272727272725,
|
| 148 |
+
"grad_norm": 40.61027526855469,
|
| 149 |
+
"learning_rate": 1.9702911467617353e-05,
|
| 150 |
+
"loss": 1.6693,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.406417112299465,
|
| 155 |
+
"grad_norm": 26.62150764465332,
|
| 156 |
+
"learning_rate": 1.9554367201426024e-05,
|
| 157 |
+
"loss": 1.8081,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.5401069518716577,
|
| 162 |
+
"grad_norm": 40.96317672729492,
|
| 163 |
+
"learning_rate": 1.94058229352347e-05,
|
| 164 |
+
"loss": 1.839,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.6737967914438503,
|
| 169 |
+
"grad_norm": 71.85723114013672,
|
| 170 |
+
"learning_rate": 1.9257278669043376e-05,
|
| 171 |
+
"loss": 1.7509,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.807486631016043,
|
| 176 |
+
"grad_norm": 27.3375186920166,
|
| 177 |
+
"learning_rate": 1.9108734402852053e-05,
|
| 178 |
+
"loss": 1.6609,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 2.9411764705882355,
|
| 183 |
+
"grad_norm": 67.37564086914062,
|
| 184 |
+
"learning_rate": 1.8960190136660727e-05,
|
| 185 |
+
"loss": 1.5832,
|
| 186 |
+
"step": 550
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 3.0,
|
| 190 |
+
"eval_accuracy": 0.8548951048951049,
|
| 191 |
+
"eval_loss": 0.4113427996635437,
|
| 192 |
+
"eval_macro_f1": 0.8558487435825608,
|
| 193 |
+
"eval_runtime": 0.4094,
|
| 194 |
+
"eval_samples_per_second": 1397.145,
|
| 195 |
+
"eval_steps_per_second": 87.932,
|
| 196 |
+
"eval_weighted_f1": 0.8548911382666233,
|
| 197 |
+
"step": 561
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.0748663101604277,
|
| 201 |
+
"grad_norm": 45.191673278808594,
|
| 202 |
+
"learning_rate": 1.88116458704694e-05,
|
| 203 |
+
"loss": 1.3509,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.2085561497326203,
|
| 208 |
+
"grad_norm": 25.53241729736328,
|
| 209 |
+
"learning_rate": 1.8663101604278075e-05,
|
| 210 |
+
"loss": 1.1761,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.342245989304813,
|
| 215 |
+
"grad_norm": 41.53510665893555,
|
| 216 |
+
"learning_rate": 1.8514557338086753e-05,
|
| 217 |
+
"loss": 1.0847,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.4759358288770055,
|
| 222 |
+
"grad_norm": 31.06566047668457,
|
| 223 |
+
"learning_rate": 1.8366013071895427e-05,
|
| 224 |
+
"loss": 0.9789,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.6096256684491976,
|
| 229 |
+
"grad_norm": 34.52959442138672,
|
| 230 |
+
"learning_rate": 1.82174688057041e-05,
|
| 231 |
+
"loss": 1.2125,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.7433155080213902,
|
| 236 |
+
"grad_norm": 32.005027770996094,
|
| 237 |
+
"learning_rate": 1.8068924539512775e-05,
|
| 238 |
+
"loss": 1.2272,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.877005347593583,
|
| 243 |
+
"grad_norm": 83.8464584350586,
|
| 244 |
+
"learning_rate": 1.7920380273321453e-05,
|
| 245 |
+
"loss": 1.1388,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 4.0,
|
| 250 |
+
"eval_accuracy": 0.8409090909090909,
|
| 251 |
+
"eval_loss": 0.5294616222381592,
|
| 252 |
+
"eval_macro_f1": 0.839036322957118,
|
| 253 |
+
"eval_runtime": 0.4598,
|
| 254 |
+
"eval_samples_per_second": 1244.101,
|
| 255 |
+
"eval_steps_per_second": 78.3,
|
| 256 |
+
"eval_weighted_f1": 0.8386455667878976,
|
| 257 |
+
"step": 748
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.010695187165775,
|
| 261 |
+
"grad_norm": 48.300106048583984,
|
| 262 |
+
"learning_rate": 1.7771836007130127e-05,
|
| 263 |
+
"loss": 1.1598,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.144385026737968,
|
| 268 |
+
"grad_norm": 5.517875671386719,
|
| 269 |
+
"learning_rate": 1.76232917409388e-05,
|
| 270 |
+
"loss": 0.8199,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.278074866310161,
|
| 275 |
+
"grad_norm": 44.1145133972168,
|
| 276 |
+
"learning_rate": 1.7474747474747475e-05,
|
| 277 |
+
"loss": 0.8252,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.411764705882353,
|
| 282 |
+
"grad_norm": 126.29991149902344,
|
| 283 |
+
"learning_rate": 1.7326203208556153e-05,
|
| 284 |
+
"loss": 0.9085,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.545454545454545,
|
| 289 |
+
"grad_norm": 83.34587860107422,
|
| 290 |
+
"learning_rate": 1.7177658942364827e-05,
|
| 291 |
+
"loss": 0.7028,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.6791443850267385,
|
| 296 |
+
"grad_norm": 98.39545440673828,
|
| 297 |
+
"learning_rate": 1.70291146761735e-05,
|
| 298 |
+
"loss": 0.678,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.81283422459893,
|
| 303 |
+
"grad_norm": 56.110313415527344,
|
| 304 |
+
"learning_rate": 1.6880570409982175e-05,
|
| 305 |
+
"loss": 0.7373,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 4.946524064171123,
|
| 310 |
+
"grad_norm": 136.49549865722656,
|
| 311 |
+
"learning_rate": 1.6732026143790852e-05,
|
| 312 |
+
"loss": 0.9215,
|
| 313 |
+
"step": 925
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 5.0,
|
| 317 |
+
"eval_accuracy": 0.833916083916084,
|
| 318 |
+
"eval_loss": 0.5093075633049011,
|
| 319 |
+
"eval_macro_f1": 0.8352833540182201,
|
| 320 |
+
"eval_runtime": 0.4116,
|
| 321 |
+
"eval_samples_per_second": 1389.838,
|
| 322 |
+
"eval_steps_per_second": 87.472,
|
| 323 |
+
"eval_weighted_f1": 0.834396889941452,
|
| 324 |
+
"step": 935
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.080213903743315,
|
| 328 |
+
"grad_norm": 12.642509460449219,
|
| 329 |
+
"learning_rate": 1.6583481877599526e-05,
|
| 330 |
+
"loss": 0.6204,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.213903743315508,
|
| 335 |
+
"grad_norm": 9.276065826416016,
|
| 336 |
+
"learning_rate": 1.64349376114082e-05,
|
| 337 |
+
"loss": 0.4692,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.347593582887701,
|
| 342 |
+
"grad_norm": 43.87451171875,
|
| 343 |
+
"learning_rate": 1.6286393345216875e-05,
|
| 344 |
+
"loss": 0.5851,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.481283422459893,
|
| 349 |
+
"grad_norm": 7.655317783355713,
|
| 350 |
+
"learning_rate": 1.6137849079025552e-05,
|
| 351 |
+
"loss": 0.6233,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.614973262032086,
|
| 356 |
+
"grad_norm": 111.20254516601562,
|
| 357 |
+
"learning_rate": 1.5989304812834226e-05,
|
| 358 |
+
"loss": 0.5837,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.748663101604278,
|
| 363 |
+
"grad_norm": 92.34879302978516,
|
| 364 |
+
"learning_rate": 1.58407605466429e-05,
|
| 365 |
+
"loss": 0.7491,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 5.882352941176471,
|
| 370 |
+
"grad_norm": 104.95001220703125,
|
| 371 |
+
"learning_rate": 1.5692216280451574e-05,
|
| 372 |
+
"loss": 0.6086,
|
| 373 |
+
"step": 1100
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 6.0,
|
| 377 |
+
"eval_accuracy": 0.8479020979020979,
|
| 378 |
+
"eval_loss": 0.5817632079124451,
|
| 379 |
+
"eval_macro_f1": 0.8483336270509397,
|
| 380 |
+
"eval_runtime": 0.5284,
|
| 381 |
+
"eval_samples_per_second": 1082.472,
|
| 382 |
+
"eval_steps_per_second": 68.128,
|
| 383 |
+
"eval_weighted_f1": 0.8473500791236364,
|
| 384 |
+
"step": 1122
|
| 385 |
+
}
|
| 386 |
+
],
|
| 387 |
+
"logging_steps": 25,
|
| 388 |
+
"max_steps": 3740,
|
| 389 |
+
"num_input_tokens_seen": 0,
|
| 390 |
+
"num_train_epochs": 20,
|
| 391 |
+
"save_steps": 500,
|
| 392 |
+
"stateful_callbacks": {
|
| 393 |
+
"EarlyStoppingCallback": {
|
| 394 |
+
"args": {
|
| 395 |
+
"early_stopping_patience": 3,
|
| 396 |
+
"early_stopping_threshold": 0.0
|
| 397 |
+
},
|
| 398 |
+
"attributes": {
|
| 399 |
+
"early_stopping_patience_counter": 3
|
| 400 |
+
}
|
| 401 |
+
},
|
| 402 |
+
"TrainerControl": {
|
| 403 |
+
"args": {
|
| 404 |
+
"should_epoch_stop": false,
|
| 405 |
+
"should_evaluate": false,
|
| 406 |
+
"should_log": false,
|
| 407 |
+
"should_save": true,
|
| 408 |
+
"should_training_stop": true
|
| 409 |
+
},
|
| 410 |
+
"attributes": {}
|
| 411 |
+
}
|
| 412 |
+
},
|
| 413 |
+
"total_flos": 0.0,
|
| 414 |
+
"train_batch_size": 8,
|
| 415 |
+
"trial_name": null,
|
| 416 |
+
"trial_params": null
|
| 417 |
+
}
|
ood-vifactcheck-phobert-topic-domain-00-s44-20260826T110330Z-ab4112e3/trainer/checkpoint-1122/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-00-s44-20260826T110330Z-ab4112e3/trainer/checkpoint-561/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-topic-domain-00-s44-20260826T110330Z-ab4112e3/trainer/checkpoint-561/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-00-s44-20260826T110330Z-ab4112e3/trainer/checkpoint-561/special_tokens_map.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": "<mask>",
|
| 6 |
+
"pad_token": "<pad>",
|
| 7 |
+
"sep_token": "</s>",
|
| 8 |
+
"unk_token": "<unk>"
|
| 9 |
+
}
|
ood-vifactcheck-phobert-topic-domain-00-s44-20260826T110330Z-ab4112e3/trainer/checkpoint-561/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"64000": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 256,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "PhobertTokenizer",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-phobert-topic-domain-00-s44-20260826T110330Z-ab4112e3/trainer/checkpoint-561/trainer_state.json
ADDED
|
@@ -0,0 +1,230 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 561,
|
| 3 |
+
"best_metric": 0.8558487435825608,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-topic-domain-00-s44-20260826T110330Z-ab4112e3/trainer/checkpoint-561",
|
| 5 |
+
"epoch": 3.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 561,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.13368983957219252,
|
| 14 |
+
"grad_norm": 8.93509578704834,
|
| 15 |
+
"learning_rate": 1.2834224598930483e-06,
|
| 16 |
+
"loss": 4.4108,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.26737967914438504,
|
| 21 |
+
"grad_norm": 9.546072959899902,
|
| 22 |
+
"learning_rate": 2.6203208556149735e-06,
|
| 23 |
+
"loss": 4.3762,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.40106951871657753,
|
| 28 |
+
"grad_norm": 10.253655433654785,
|
| 29 |
+
"learning_rate": 3.957219251336899e-06,
|
| 30 |
+
"loss": 4.3644,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5347593582887701,
|
| 35 |
+
"grad_norm": 6.676309108734131,
|
| 36 |
+
"learning_rate": 5.294117647058824e-06,
|
| 37 |
+
"loss": 4.3382,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6684491978609626,
|
| 42 |
+
"grad_norm": 10.954312324523926,
|
| 43 |
+
"learning_rate": 6.631016042780749e-06,
|
| 44 |
+
"loss": 4.296,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8021390374331551,
|
| 49 |
+
"grad_norm": 10.327520370483398,
|
| 50 |
+
"learning_rate": 7.967914438502674e-06,
|
| 51 |
+
"loss": 4.1346,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9358288770053476,
|
| 56 |
+
"grad_norm": 21.59717559814453,
|
| 57 |
+
"learning_rate": 9.3048128342246e-06,
|
| 58 |
+
"loss": 3.9314,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.6258741258741258,
|
| 64 |
+
"eval_loss": 0.8830139636993408,
|
| 65 |
+
"eval_macro_f1": 0.6177260306522513,
|
| 66 |
+
"eval_runtime": 0.3903,
|
| 67 |
+
"eval_samples_per_second": 1465.711,
|
| 68 |
+
"eval_steps_per_second": 92.248,
|
| 69 |
+
"eval_weighted_f1": 0.6201395610448718,
|
| 70 |
+
"step": 187
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0695187165775402,
|
| 74 |
+
"grad_norm": 28.900501251220703,
|
| 75 |
+
"learning_rate": 1.0641711229946525e-05,
|
| 76 |
+
"loss": 3.6519,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.2032085561497325,
|
| 81 |
+
"grad_norm": 41.45104217529297,
|
| 82 |
+
"learning_rate": 1.197860962566845e-05,
|
| 83 |
+
"loss": 3.4209,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.3368983957219251,
|
| 88 |
+
"grad_norm": 24.274688720703125,
|
| 89 |
+
"learning_rate": 1.3315508021390375e-05,
|
| 90 |
+
"loss": 3.0569,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.4705882352941178,
|
| 95 |
+
"grad_norm": 31.610830307006836,
|
| 96 |
+
"learning_rate": 1.4652406417112302e-05,
|
| 97 |
+
"loss": 2.8981,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.6042780748663101,
|
| 102 |
+
"grad_norm": 36.086883544921875,
|
| 103 |
+
"learning_rate": 1.5989304812834226e-05,
|
| 104 |
+
"loss": 2.6154,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7379679144385025,
|
| 109 |
+
"grad_norm": 58.68095016479492,
|
| 110 |
+
"learning_rate": 1.7326203208556153e-05,
|
| 111 |
+
"loss": 2.356,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.8716577540106951,
|
| 116 |
+
"grad_norm": 33.54307556152344,
|
| 117 |
+
"learning_rate": 1.8663101604278075e-05,
|
| 118 |
+
"loss": 2.4054,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.8251748251748252,
|
| 124 |
+
"eval_loss": 0.4848474860191345,
|
| 125 |
+
"eval_macro_f1": 0.825687690742074,
|
| 126 |
+
"eval_runtime": 0.4188,
|
| 127 |
+
"eval_samples_per_second": 1365.72,
|
| 128 |
+
"eval_steps_per_second": 85.954,
|
| 129 |
+
"eval_weighted_f1": 0.8252022958446541,
|
| 130 |
+
"step": 374
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.0053475935828877,
|
| 134 |
+
"grad_norm": 29.90228271484375,
|
| 135 |
+
"learning_rate": 2e-05,
|
| 136 |
+
"loss": 2.1596,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.1390374331550803,
|
| 141 |
+
"grad_norm": 33.18424987792969,
|
| 142 |
+
"learning_rate": 1.9851455733808676e-05,
|
| 143 |
+
"loss": 1.572,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.2727272727272725,
|
| 148 |
+
"grad_norm": 40.61027526855469,
|
| 149 |
+
"learning_rate": 1.9702911467617353e-05,
|
| 150 |
+
"loss": 1.6693,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.406417112299465,
|
| 155 |
+
"grad_norm": 26.62150764465332,
|
| 156 |
+
"learning_rate": 1.9554367201426024e-05,
|
| 157 |
+
"loss": 1.8081,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.5401069518716577,
|
| 162 |
+
"grad_norm": 40.96317672729492,
|
| 163 |
+
"learning_rate": 1.94058229352347e-05,
|
| 164 |
+
"loss": 1.839,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.6737967914438503,
|
| 169 |
+
"grad_norm": 71.85723114013672,
|
| 170 |
+
"learning_rate": 1.9257278669043376e-05,
|
| 171 |
+
"loss": 1.7509,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.807486631016043,
|
| 176 |
+
"grad_norm": 27.3375186920166,
|
| 177 |
+
"learning_rate": 1.9108734402852053e-05,
|
| 178 |
+
"loss": 1.6609,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 2.9411764705882355,
|
| 183 |
+
"grad_norm": 67.37564086914062,
|
| 184 |
+
"learning_rate": 1.8960190136660727e-05,
|
| 185 |
+
"loss": 1.5832,
|
| 186 |
+
"step": 550
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 3.0,
|
| 190 |
+
"eval_accuracy": 0.8548951048951049,
|
| 191 |
+
"eval_loss": 0.4113427996635437,
|
| 192 |
+
"eval_macro_f1": 0.8558487435825608,
|
| 193 |
+
"eval_runtime": 0.4094,
|
| 194 |
+
"eval_samples_per_second": 1397.145,
|
| 195 |
+
"eval_steps_per_second": 87.932,
|
| 196 |
+
"eval_weighted_f1": 0.8548911382666233,
|
| 197 |
+
"step": 561
|
| 198 |
+
}
|
| 199 |
+
],
|
| 200 |
+
"logging_steps": 25,
|
| 201 |
+
"max_steps": 3740,
|
| 202 |
+
"num_input_tokens_seen": 0,
|
| 203 |
+
"num_train_epochs": 20,
|
| 204 |
+
"save_steps": 500,
|
| 205 |
+
"stateful_callbacks": {
|
| 206 |
+
"EarlyStoppingCallback": {
|
| 207 |
+
"args": {
|
| 208 |
+
"early_stopping_patience": 3,
|
| 209 |
+
"early_stopping_threshold": 0.0
|
| 210 |
+
},
|
| 211 |
+
"attributes": {
|
| 212 |
+
"early_stopping_patience_counter": 0
|
| 213 |
+
}
|
| 214 |
+
},
|
| 215 |
+
"TrainerControl": {
|
| 216 |
+
"args": {
|
| 217 |
+
"should_epoch_stop": false,
|
| 218 |
+
"should_evaluate": false,
|
| 219 |
+
"should_log": false,
|
| 220 |
+
"should_save": true,
|
| 221 |
+
"should_training_stop": false
|
| 222 |
+
},
|
| 223 |
+
"attributes": {}
|
| 224 |
+
}
|
| 225 |
+
},
|
| 226 |
+
"total_flos": 0.0,
|
| 227 |
+
"train_batch_size": 8,
|
| 228 |
+
"trial_name": null,
|
| 229 |
+
"trial_params": null
|
| 230 |
+
}
|
ood-vifactcheck-phobert-topic-domain-00-s44-20260826T110330Z-ab4112e3/trainer/checkpoint-561/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-01-s44-20260826T110548Z-ab4112e3/trainer/checkpoint-1337/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-topic-domain-01-s44-20260826T110548Z-ab4112e3/trainer/checkpoint-1337/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-01-s44-20260826T110548Z-ab4112e3/trainer/checkpoint-1337/special_tokens_map.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": "<mask>",
|
| 6 |
+
"pad_token": "<pad>",
|
| 7 |
+
"sep_token": "</s>",
|
| 8 |
+
"unk_token": "<unk>"
|
| 9 |
+
}
|
ood-vifactcheck-phobert-topic-domain-01-s44-20260826T110548Z-ab4112e3/trainer/checkpoint-1337/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"64000": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 256,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "PhobertTokenizer",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-phobert-topic-domain-01-s44-20260826T110548Z-ab4112e3/trainer/checkpoint-1337/trainer_state.json
ADDED
|
@@ -0,0 +1,491 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1337,
|
| 3 |
+
"best_metric": 0.8468749690930087,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-topic-domain-01-s44-20260826T110548Z-ab4112e3/trainer/checkpoint-1337",
|
| 5 |
+
"epoch": 7.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1337,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.1310615989515072,
|
| 14 |
+
"grad_norm": 6.835499286651611,
|
| 15 |
+
"learning_rate": 1.256544502617801e-06,
|
| 16 |
+
"loss": 4.4524,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.2621231979030144,
|
| 21 |
+
"grad_norm": 7.782848358154297,
|
| 22 |
+
"learning_rate": 2.5654450261780104e-06,
|
| 23 |
+
"loss": 4.394,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.3931847968545216,
|
| 28 |
+
"grad_norm": 6.044909954071045,
|
| 29 |
+
"learning_rate": 3.874345549738221e-06,
|
| 30 |
+
"loss": 4.3521,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5242463958060288,
|
| 35 |
+
"grad_norm": 8.78554916381836,
|
| 36 |
+
"learning_rate": 5.18324607329843e-06,
|
| 37 |
+
"loss": 4.336,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6553079947575361,
|
| 42 |
+
"grad_norm": 10.934264183044434,
|
| 43 |
+
"learning_rate": 6.492146596858639e-06,
|
| 44 |
+
"loss": 4.2358,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.7863695937090432,
|
| 49 |
+
"grad_norm": 13.719097137451172,
|
| 50 |
+
"learning_rate": 7.80104712041885e-06,
|
| 51 |
+
"loss": 4.1304,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9174311926605505,
|
| 56 |
+
"grad_norm": 21.721654891967773,
|
| 57 |
+
"learning_rate": 9.109947643979057e-06,
|
| 58 |
+
"loss": 3.9818,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.6164383561643836,
|
| 64 |
+
"eval_loss": 0.8874578475952148,
|
| 65 |
+
"eval_macro_f1": 0.5915546048952451,
|
| 66 |
+
"eval_runtime": 0.4568,
|
| 67 |
+
"eval_samples_per_second": 1278.566,
|
| 68 |
+
"eval_steps_per_second": 81.005,
|
| 69 |
+
"eval_weighted_f1": 0.5937603838964497,
|
| 70 |
+
"step": 191
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0471821756225426,
|
| 74 |
+
"grad_norm": 25.8647518157959,
|
| 75 |
+
"learning_rate": 1.0418848167539269e-05,
|
| 76 |
+
"loss": 3.6178,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.1782437745740497,
|
| 81 |
+
"grad_norm": 31.343774795532227,
|
| 82 |
+
"learning_rate": 1.1727748691099476e-05,
|
| 83 |
+
"loss": 3.4075,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.309305373525557,
|
| 88 |
+
"grad_norm": 27.553184509277344,
|
| 89 |
+
"learning_rate": 1.3036649214659686e-05,
|
| 90 |
+
"loss": 3.0463,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.4403669724770642,
|
| 95 |
+
"grad_norm": 24.087337493896484,
|
| 96 |
+
"learning_rate": 1.4345549738219897e-05,
|
| 97 |
+
"loss": 2.6957,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.5714285714285714,
|
| 102 |
+
"grad_norm": 46.64403533935547,
|
| 103 |
+
"learning_rate": 1.5654450261780107e-05,
|
| 104 |
+
"loss": 2.5164,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7024901703800785,
|
| 109 |
+
"grad_norm": 41.59882736206055,
|
| 110 |
+
"learning_rate": 1.6963350785340316e-05,
|
| 111 |
+
"loss": 2.5238,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.833551769331586,
|
| 116 |
+
"grad_norm": 46.93950653076172,
|
| 117 |
+
"learning_rate": 1.8272251308900526e-05,
|
| 118 |
+
"loss": 2.2301,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 1.964613368283093,
|
| 123 |
+
"grad_norm": 44.93206024169922,
|
| 124 |
+
"learning_rate": 1.9581151832460736e-05,
|
| 125 |
+
"loss": 2.1671,
|
| 126 |
+
"step": 375
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"epoch": 2.0,
|
| 130 |
+
"eval_accuracy": 0.8116438356164384,
|
| 131 |
+
"eval_loss": 0.47884002327919006,
|
| 132 |
+
"eval_macro_f1": 0.8102593120193068,
|
| 133 |
+
"eval_runtime": 0.4505,
|
| 134 |
+
"eval_samples_per_second": 1296.256,
|
| 135 |
+
"eval_steps_per_second": 82.126,
|
| 136 |
+
"eval_weighted_f1": 0.8106329687469656,
|
| 137 |
+
"step": 382
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.094364351245085,
|
| 141 |
+
"grad_norm": 40.07585525512695,
|
| 142 |
+
"learning_rate": 1.9901105293775453e-05,
|
| 143 |
+
"loss": 1.8435,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.2254259501965925,
|
| 148 |
+
"grad_norm": 32.61440658569336,
|
| 149 |
+
"learning_rate": 1.9755671902268764e-05,
|
| 150 |
+
"loss": 1.8324,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.3564875491480994,
|
| 155 |
+
"grad_norm": 97.53990173339844,
|
| 156 |
+
"learning_rate": 1.961023851076207e-05,
|
| 157 |
+
"loss": 1.6051,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.487549148099607,
|
| 162 |
+
"grad_norm": 47.22065734863281,
|
| 163 |
+
"learning_rate": 1.9464805119255382e-05,
|
| 164 |
+
"loss": 1.5596,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.618610747051114,
|
| 169 |
+
"grad_norm": 39.30236053466797,
|
| 170 |
+
"learning_rate": 1.9319371727748693e-05,
|
| 171 |
+
"loss": 1.5698,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.749672346002621,
|
| 176 |
+
"grad_norm": 68.9538345336914,
|
| 177 |
+
"learning_rate": 1.9173938336242e-05,
|
| 178 |
+
"loss": 1.6676,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 2.8807339449541285,
|
| 183 |
+
"grad_norm": 29.888273239135742,
|
| 184 |
+
"learning_rate": 1.9028504944735315e-05,
|
| 185 |
+
"loss": 1.6416,
|
| 186 |
+
"step": 550
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 3.0,
|
| 190 |
+
"eval_accuracy": 0.8321917808219178,
|
| 191 |
+
"eval_loss": 0.4808124005794525,
|
| 192 |
+
"eval_macro_f1": 0.8312456022796614,
|
| 193 |
+
"eval_runtime": 0.4223,
|
| 194 |
+
"eval_samples_per_second": 1382.94,
|
| 195 |
+
"eval_steps_per_second": 87.618,
|
| 196 |
+
"eval_weighted_f1": 0.8312555870146,
|
| 197 |
+
"step": 573
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.0104849279161208,
|
| 201 |
+
"grad_norm": 13.08161735534668,
|
| 202 |
+
"learning_rate": 1.8883071553228622e-05,
|
| 203 |
+
"loss": 1.5042,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.1415465268676277,
|
| 208 |
+
"grad_norm": 38.01715850830078,
|
| 209 |
+
"learning_rate": 1.8737638161721933e-05,
|
| 210 |
+
"loss": 1.0345,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.272608125819135,
|
| 215 |
+
"grad_norm": 65.74954986572266,
|
| 216 |
+
"learning_rate": 1.8592204770215244e-05,
|
| 217 |
+
"loss": 0.9756,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.4036697247706424,
|
| 222 |
+
"grad_norm": 44.841064453125,
|
| 223 |
+
"learning_rate": 1.844677137870855e-05,
|
| 224 |
+
"loss": 1.0279,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.5347313237221494,
|
| 229 |
+
"grad_norm": 24.73827362060547,
|
| 230 |
+
"learning_rate": 1.8301337987201865e-05,
|
| 231 |
+
"loss": 1.0317,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.6657929226736568,
|
| 236 |
+
"grad_norm": 38.94432067871094,
|
| 237 |
+
"learning_rate": 1.8155904595695173e-05,
|
| 238 |
+
"loss": 1.0829,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.7968545216251637,
|
| 243 |
+
"grad_norm": 76.02899932861328,
|
| 244 |
+
"learning_rate": 1.8010471204188483e-05,
|
| 245 |
+
"loss": 1.0498,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 3.927916120576671,
|
| 250 |
+
"grad_norm": 19.195104598999023,
|
| 251 |
+
"learning_rate": 1.7865037812681794e-05,
|
| 252 |
+
"loss": 1.071,
|
| 253 |
+
"step": 750
|
| 254 |
+
},
|
| 255 |
+
{
|
| 256 |
+
"epoch": 4.0,
|
| 257 |
+
"eval_accuracy": 0.839041095890411,
|
| 258 |
+
"eval_loss": 0.45466530323028564,
|
| 259 |
+
"eval_macro_f1": 0.8390703895384747,
|
| 260 |
+
"eval_runtime": 0.6565,
|
| 261 |
+
"eval_samples_per_second": 889.561,
|
| 262 |
+
"eval_steps_per_second": 56.359,
|
| 263 |
+
"eval_weighted_f1": 0.8391181788025274,
|
| 264 |
+
"step": 764
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.057667103538663,
|
| 268 |
+
"grad_norm": 50.151580810546875,
|
| 269 |
+
"learning_rate": 1.77196044211751e-05,
|
| 270 |
+
"loss": 0.8696,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.18872870249017,
|
| 275 |
+
"grad_norm": 20.54074478149414,
|
| 276 |
+
"learning_rate": 1.7574171029668416e-05,
|
| 277 |
+
"loss": 0.611,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.319790301441677,
|
| 282 |
+
"grad_norm": 82.54438018798828,
|
| 283 |
+
"learning_rate": 1.7428737638161723e-05,
|
| 284 |
+
"loss": 0.7017,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.450851900393185,
|
| 289 |
+
"grad_norm": 44.62449264526367,
|
| 290 |
+
"learning_rate": 1.7283304246655034e-05,
|
| 291 |
+
"loss": 0.6088,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.581913499344692,
|
| 296 |
+
"grad_norm": 56.26701736450195,
|
| 297 |
+
"learning_rate": 1.7137870855148345e-05,
|
| 298 |
+
"loss": 0.5794,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.712975098296199,
|
| 303 |
+
"grad_norm": 71.91743469238281,
|
| 304 |
+
"learning_rate": 1.6992437463641652e-05,
|
| 305 |
+
"loss": 0.5883,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 4.844036697247707,
|
| 310 |
+
"grad_norm": 10.919384002685547,
|
| 311 |
+
"learning_rate": 1.6847004072134963e-05,
|
| 312 |
+
"loss": 0.8431,
|
| 313 |
+
"step": 925
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 4.975098296199214,
|
| 317 |
+
"grad_norm": 15.44614315032959,
|
| 318 |
+
"learning_rate": 1.6701570680628274e-05,
|
| 319 |
+
"loss": 0.7411,
|
| 320 |
+
"step": 950
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 5.0,
|
| 324 |
+
"eval_accuracy": 0.839041095890411,
|
| 325 |
+
"eval_loss": 0.6060966849327087,
|
| 326 |
+
"eval_macro_f1": 0.8396718762569922,
|
| 327 |
+
"eval_runtime": 0.4603,
|
| 328 |
+
"eval_samples_per_second": 1268.668,
|
| 329 |
+
"eval_steps_per_second": 80.378,
|
| 330 |
+
"eval_weighted_f1": 0.8396851310798896,
|
| 331 |
+
"step": 955
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.104849279161206,
|
| 335 |
+
"grad_norm": 1.0175487995147705,
|
| 336 |
+
"learning_rate": 1.6556137289121585e-05,
|
| 337 |
+
"loss": 0.5264,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.235910878112713,
|
| 342 |
+
"grad_norm": 1.4465217590332031,
|
| 343 |
+
"learning_rate": 1.6410703897614895e-05,
|
| 344 |
+
"loss": 0.4807,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.36697247706422,
|
| 349 |
+
"grad_norm": 53.33880615234375,
|
| 350 |
+
"learning_rate": 1.6265270506108203e-05,
|
| 351 |
+
"loss": 0.5422,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.498034076015728,
|
| 356 |
+
"grad_norm": 89.72924041748047,
|
| 357 |
+
"learning_rate": 1.6119837114601514e-05,
|
| 358 |
+
"loss": 0.4899,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.6290956749672345,
|
| 363 |
+
"grad_norm": 37.08826446533203,
|
| 364 |
+
"learning_rate": 1.5974403723094824e-05,
|
| 365 |
+
"loss": 0.3678,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 5.760157273918741,
|
| 370 |
+
"grad_norm": 15.998916625976562,
|
| 371 |
+
"learning_rate": 1.5828970331588135e-05,
|
| 372 |
+
"loss": 0.4723,
|
| 373 |
+
"step": 1100
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 5.891218872870249,
|
| 377 |
+
"grad_norm": 41.07233810424805,
|
| 378 |
+
"learning_rate": 1.5683536940081443e-05,
|
| 379 |
+
"loss": 0.5493,
|
| 380 |
+
"step": 1125
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 6.0,
|
| 384 |
+
"eval_accuracy": 0.8424657534246576,
|
| 385 |
+
"eval_loss": 0.6887933611869812,
|
| 386 |
+
"eval_macro_f1": 0.8419490901512526,
|
| 387 |
+
"eval_runtime": 0.4426,
|
| 388 |
+
"eval_samples_per_second": 1319.619,
|
| 389 |
+
"eval_steps_per_second": 83.606,
|
| 390 |
+
"eval_weighted_f1": 0.8419858826813973,
|
| 391 |
+
"step": 1146
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 6.0209698558322415,
|
| 395 |
+
"grad_norm": 65.35658264160156,
|
| 396 |
+
"learning_rate": 1.5538103548574753e-05,
|
| 397 |
+
"loss": 0.493,
|
| 398 |
+
"step": 1150
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.1520314547837485,
|
| 402 |
+
"grad_norm": 71.60026550292969,
|
| 403 |
+
"learning_rate": 1.5392670157068064e-05,
|
| 404 |
+
"loss": 0.3782,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.283093053735255,
|
| 409 |
+
"grad_norm": 5.998918056488037,
|
| 410 |
+
"learning_rate": 1.5247236765561375e-05,
|
| 411 |
+
"loss": 0.3062,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.414154652686763,
|
| 416 |
+
"grad_norm": 2.930375814437866,
|
| 417 |
+
"learning_rate": 1.5101803374054684e-05,
|
| 418 |
+
"loss": 0.5499,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.54521625163827,
|
| 423 |
+
"grad_norm": 0.9057067036628723,
|
| 424 |
+
"learning_rate": 1.4956369982547993e-05,
|
| 425 |
+
"loss": 0.4047,
|
| 426 |
+
"step": 1250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 6.676277850589777,
|
| 430 |
+
"grad_norm": 140.81387329101562,
|
| 431 |
+
"learning_rate": 1.4810936591041306e-05,
|
| 432 |
+
"loss": 0.4865,
|
| 433 |
+
"step": 1275
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 6.807339449541285,
|
| 437 |
+
"grad_norm": 0.4184958338737488,
|
| 438 |
+
"learning_rate": 1.4665503199534615e-05,
|
| 439 |
+
"loss": 0.5097,
|
| 440 |
+
"step": 1300
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 6.938401048492792,
|
| 444 |
+
"grad_norm": 29.483163833618164,
|
| 445 |
+
"learning_rate": 1.4520069808027924e-05,
|
| 446 |
+
"loss": 0.5279,
|
| 447 |
+
"step": 1325
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 7.0,
|
| 451 |
+
"eval_accuracy": 0.8458904109589042,
|
| 452 |
+
"eval_loss": 0.7323911786079407,
|
| 453 |
+
"eval_macro_f1": 0.8468749690930087,
|
| 454 |
+
"eval_runtime": 0.4847,
|
| 455 |
+
"eval_samples_per_second": 1204.837,
|
| 456 |
+
"eval_steps_per_second": 76.334,
|
| 457 |
+
"eval_weighted_f1": 0.8470004215396969,
|
| 458 |
+
"step": 1337
|
| 459 |
+
}
|
| 460 |
+
],
|
| 461 |
+
"logging_steps": 25,
|
| 462 |
+
"max_steps": 3820,
|
| 463 |
+
"num_input_tokens_seen": 0,
|
| 464 |
+
"num_train_epochs": 20,
|
| 465 |
+
"save_steps": 500,
|
| 466 |
+
"stateful_callbacks": {
|
| 467 |
+
"EarlyStoppingCallback": {
|
| 468 |
+
"args": {
|
| 469 |
+
"early_stopping_patience": 3,
|
| 470 |
+
"early_stopping_threshold": 0.0
|
| 471 |
+
},
|
| 472 |
+
"attributes": {
|
| 473 |
+
"early_stopping_patience_counter": 0
|
| 474 |
+
}
|
| 475 |
+
},
|
| 476 |
+
"TrainerControl": {
|
| 477 |
+
"args": {
|
| 478 |
+
"should_epoch_stop": false,
|
| 479 |
+
"should_evaluate": false,
|
| 480 |
+
"should_log": false,
|
| 481 |
+
"should_save": true,
|
| 482 |
+
"should_training_stop": false
|
| 483 |
+
},
|
| 484 |
+
"attributes": {}
|
| 485 |
+
}
|
| 486 |
+
},
|
| 487 |
+
"total_flos": 0.0,
|
| 488 |
+
"train_batch_size": 8,
|
| 489 |
+
"trial_name": null,
|
| 490 |
+
"trial_params": null
|
| 491 |
+
}
|
ood-vifactcheck-phobert-topic-domain-01-s44-20260826T110548Z-ab4112e3/trainer/checkpoint-1337/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-01-s44-20260826T110548Z-ab4112e3/trainer/checkpoint-1910/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-topic-domain-01-s44-20260826T110548Z-ab4112e3/trainer/checkpoint-1910/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-01-s44-20260826T110548Z-ab4112e3/trainer/checkpoint-1910/special_tokens_map.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": "<mask>",
|
| 6 |
+
"pad_token": "<pad>",
|
| 7 |
+
"sep_token": "</s>",
|
| 8 |
+
"unk_token": "<unk>"
|
| 9 |
+
}
|
ood-vifactcheck-phobert-topic-domain-01-s44-20260826T110548Z-ab4112e3/trainer/checkpoint-1910/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"64000": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 256,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "PhobertTokenizer",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-phobert-topic-domain-01-s44-20260826T110548Z-ab4112e3/trainer/checkpoint-1910/trainer_state.json
ADDED
|
@@ -0,0 +1,685 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1337,
|
| 3 |
+
"best_metric": 0.8468749690930087,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-topic-domain-01-s44-20260826T110548Z-ab4112e3/trainer/checkpoint-1337",
|
| 5 |
+
"epoch": 10.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1910,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.1310615989515072,
|
| 14 |
+
"grad_norm": 6.835499286651611,
|
| 15 |
+
"learning_rate": 1.256544502617801e-06,
|
| 16 |
+
"loss": 4.4524,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.2621231979030144,
|
| 21 |
+
"grad_norm": 7.782848358154297,
|
| 22 |
+
"learning_rate": 2.5654450261780104e-06,
|
| 23 |
+
"loss": 4.394,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.3931847968545216,
|
| 28 |
+
"grad_norm": 6.044909954071045,
|
| 29 |
+
"learning_rate": 3.874345549738221e-06,
|
| 30 |
+
"loss": 4.3521,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5242463958060288,
|
| 35 |
+
"grad_norm": 8.78554916381836,
|
| 36 |
+
"learning_rate": 5.18324607329843e-06,
|
| 37 |
+
"loss": 4.336,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6553079947575361,
|
| 42 |
+
"grad_norm": 10.934264183044434,
|
| 43 |
+
"learning_rate": 6.492146596858639e-06,
|
| 44 |
+
"loss": 4.2358,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.7863695937090432,
|
| 49 |
+
"grad_norm": 13.719097137451172,
|
| 50 |
+
"learning_rate": 7.80104712041885e-06,
|
| 51 |
+
"loss": 4.1304,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9174311926605505,
|
| 56 |
+
"grad_norm": 21.721654891967773,
|
| 57 |
+
"learning_rate": 9.109947643979057e-06,
|
| 58 |
+
"loss": 3.9818,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.6164383561643836,
|
| 64 |
+
"eval_loss": 0.8874578475952148,
|
| 65 |
+
"eval_macro_f1": 0.5915546048952451,
|
| 66 |
+
"eval_runtime": 0.4568,
|
| 67 |
+
"eval_samples_per_second": 1278.566,
|
| 68 |
+
"eval_steps_per_second": 81.005,
|
| 69 |
+
"eval_weighted_f1": 0.5937603838964497,
|
| 70 |
+
"step": 191
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0471821756225426,
|
| 74 |
+
"grad_norm": 25.8647518157959,
|
| 75 |
+
"learning_rate": 1.0418848167539269e-05,
|
| 76 |
+
"loss": 3.6178,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.1782437745740497,
|
| 81 |
+
"grad_norm": 31.343774795532227,
|
| 82 |
+
"learning_rate": 1.1727748691099476e-05,
|
| 83 |
+
"loss": 3.4075,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.309305373525557,
|
| 88 |
+
"grad_norm": 27.553184509277344,
|
| 89 |
+
"learning_rate": 1.3036649214659686e-05,
|
| 90 |
+
"loss": 3.0463,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.4403669724770642,
|
| 95 |
+
"grad_norm": 24.087337493896484,
|
| 96 |
+
"learning_rate": 1.4345549738219897e-05,
|
| 97 |
+
"loss": 2.6957,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.5714285714285714,
|
| 102 |
+
"grad_norm": 46.64403533935547,
|
| 103 |
+
"learning_rate": 1.5654450261780107e-05,
|
| 104 |
+
"loss": 2.5164,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7024901703800785,
|
| 109 |
+
"grad_norm": 41.59882736206055,
|
| 110 |
+
"learning_rate": 1.6963350785340316e-05,
|
| 111 |
+
"loss": 2.5238,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.833551769331586,
|
| 116 |
+
"grad_norm": 46.93950653076172,
|
| 117 |
+
"learning_rate": 1.8272251308900526e-05,
|
| 118 |
+
"loss": 2.2301,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 1.964613368283093,
|
| 123 |
+
"grad_norm": 44.93206024169922,
|
| 124 |
+
"learning_rate": 1.9581151832460736e-05,
|
| 125 |
+
"loss": 2.1671,
|
| 126 |
+
"step": 375
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"epoch": 2.0,
|
| 130 |
+
"eval_accuracy": 0.8116438356164384,
|
| 131 |
+
"eval_loss": 0.47884002327919006,
|
| 132 |
+
"eval_macro_f1": 0.8102593120193068,
|
| 133 |
+
"eval_runtime": 0.4505,
|
| 134 |
+
"eval_samples_per_second": 1296.256,
|
| 135 |
+
"eval_steps_per_second": 82.126,
|
| 136 |
+
"eval_weighted_f1": 0.8106329687469656,
|
| 137 |
+
"step": 382
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.094364351245085,
|
| 141 |
+
"grad_norm": 40.07585525512695,
|
| 142 |
+
"learning_rate": 1.9901105293775453e-05,
|
| 143 |
+
"loss": 1.8435,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.2254259501965925,
|
| 148 |
+
"grad_norm": 32.61440658569336,
|
| 149 |
+
"learning_rate": 1.9755671902268764e-05,
|
| 150 |
+
"loss": 1.8324,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.3564875491480994,
|
| 155 |
+
"grad_norm": 97.53990173339844,
|
| 156 |
+
"learning_rate": 1.961023851076207e-05,
|
| 157 |
+
"loss": 1.6051,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.487549148099607,
|
| 162 |
+
"grad_norm": 47.22065734863281,
|
| 163 |
+
"learning_rate": 1.9464805119255382e-05,
|
| 164 |
+
"loss": 1.5596,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.618610747051114,
|
| 169 |
+
"grad_norm": 39.30236053466797,
|
| 170 |
+
"learning_rate": 1.9319371727748693e-05,
|
| 171 |
+
"loss": 1.5698,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.749672346002621,
|
| 176 |
+
"grad_norm": 68.9538345336914,
|
| 177 |
+
"learning_rate": 1.9173938336242e-05,
|
| 178 |
+
"loss": 1.6676,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 2.8807339449541285,
|
| 183 |
+
"grad_norm": 29.888273239135742,
|
| 184 |
+
"learning_rate": 1.9028504944735315e-05,
|
| 185 |
+
"loss": 1.6416,
|
| 186 |
+
"step": 550
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 3.0,
|
| 190 |
+
"eval_accuracy": 0.8321917808219178,
|
| 191 |
+
"eval_loss": 0.4808124005794525,
|
| 192 |
+
"eval_macro_f1": 0.8312456022796614,
|
| 193 |
+
"eval_runtime": 0.4223,
|
| 194 |
+
"eval_samples_per_second": 1382.94,
|
| 195 |
+
"eval_steps_per_second": 87.618,
|
| 196 |
+
"eval_weighted_f1": 0.8312555870146,
|
| 197 |
+
"step": 573
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.0104849279161208,
|
| 201 |
+
"grad_norm": 13.08161735534668,
|
| 202 |
+
"learning_rate": 1.8883071553228622e-05,
|
| 203 |
+
"loss": 1.5042,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.1415465268676277,
|
| 208 |
+
"grad_norm": 38.01715850830078,
|
| 209 |
+
"learning_rate": 1.8737638161721933e-05,
|
| 210 |
+
"loss": 1.0345,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.272608125819135,
|
| 215 |
+
"grad_norm": 65.74954986572266,
|
| 216 |
+
"learning_rate": 1.8592204770215244e-05,
|
| 217 |
+
"loss": 0.9756,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.4036697247706424,
|
| 222 |
+
"grad_norm": 44.841064453125,
|
| 223 |
+
"learning_rate": 1.844677137870855e-05,
|
| 224 |
+
"loss": 1.0279,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.5347313237221494,
|
| 229 |
+
"grad_norm": 24.73827362060547,
|
| 230 |
+
"learning_rate": 1.8301337987201865e-05,
|
| 231 |
+
"loss": 1.0317,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.6657929226736568,
|
| 236 |
+
"grad_norm": 38.94432067871094,
|
| 237 |
+
"learning_rate": 1.8155904595695173e-05,
|
| 238 |
+
"loss": 1.0829,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.7968545216251637,
|
| 243 |
+
"grad_norm": 76.02899932861328,
|
| 244 |
+
"learning_rate": 1.8010471204188483e-05,
|
| 245 |
+
"loss": 1.0498,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 3.927916120576671,
|
| 250 |
+
"grad_norm": 19.195104598999023,
|
| 251 |
+
"learning_rate": 1.7865037812681794e-05,
|
| 252 |
+
"loss": 1.071,
|
| 253 |
+
"step": 750
|
| 254 |
+
},
|
| 255 |
+
{
|
| 256 |
+
"epoch": 4.0,
|
| 257 |
+
"eval_accuracy": 0.839041095890411,
|
| 258 |
+
"eval_loss": 0.45466530323028564,
|
| 259 |
+
"eval_macro_f1": 0.8390703895384747,
|
| 260 |
+
"eval_runtime": 0.6565,
|
| 261 |
+
"eval_samples_per_second": 889.561,
|
| 262 |
+
"eval_steps_per_second": 56.359,
|
| 263 |
+
"eval_weighted_f1": 0.8391181788025274,
|
| 264 |
+
"step": 764
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.057667103538663,
|
| 268 |
+
"grad_norm": 50.151580810546875,
|
| 269 |
+
"learning_rate": 1.77196044211751e-05,
|
| 270 |
+
"loss": 0.8696,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.18872870249017,
|
| 275 |
+
"grad_norm": 20.54074478149414,
|
| 276 |
+
"learning_rate": 1.7574171029668416e-05,
|
| 277 |
+
"loss": 0.611,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.319790301441677,
|
| 282 |
+
"grad_norm": 82.54438018798828,
|
| 283 |
+
"learning_rate": 1.7428737638161723e-05,
|
| 284 |
+
"loss": 0.7017,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.450851900393185,
|
| 289 |
+
"grad_norm": 44.62449264526367,
|
| 290 |
+
"learning_rate": 1.7283304246655034e-05,
|
| 291 |
+
"loss": 0.6088,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.581913499344692,
|
| 296 |
+
"grad_norm": 56.26701736450195,
|
| 297 |
+
"learning_rate": 1.7137870855148345e-05,
|
| 298 |
+
"loss": 0.5794,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.712975098296199,
|
| 303 |
+
"grad_norm": 71.91743469238281,
|
| 304 |
+
"learning_rate": 1.6992437463641652e-05,
|
| 305 |
+
"loss": 0.5883,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 4.844036697247707,
|
| 310 |
+
"grad_norm": 10.919384002685547,
|
| 311 |
+
"learning_rate": 1.6847004072134963e-05,
|
| 312 |
+
"loss": 0.8431,
|
| 313 |
+
"step": 925
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 4.975098296199214,
|
| 317 |
+
"grad_norm": 15.44614315032959,
|
| 318 |
+
"learning_rate": 1.6701570680628274e-05,
|
| 319 |
+
"loss": 0.7411,
|
| 320 |
+
"step": 950
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 5.0,
|
| 324 |
+
"eval_accuracy": 0.839041095890411,
|
| 325 |
+
"eval_loss": 0.6060966849327087,
|
| 326 |
+
"eval_macro_f1": 0.8396718762569922,
|
| 327 |
+
"eval_runtime": 0.4603,
|
| 328 |
+
"eval_samples_per_second": 1268.668,
|
| 329 |
+
"eval_steps_per_second": 80.378,
|
| 330 |
+
"eval_weighted_f1": 0.8396851310798896,
|
| 331 |
+
"step": 955
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.104849279161206,
|
| 335 |
+
"grad_norm": 1.0175487995147705,
|
| 336 |
+
"learning_rate": 1.6556137289121585e-05,
|
| 337 |
+
"loss": 0.5264,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.235910878112713,
|
| 342 |
+
"grad_norm": 1.4465217590332031,
|
| 343 |
+
"learning_rate": 1.6410703897614895e-05,
|
| 344 |
+
"loss": 0.4807,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.36697247706422,
|
| 349 |
+
"grad_norm": 53.33880615234375,
|
| 350 |
+
"learning_rate": 1.6265270506108203e-05,
|
| 351 |
+
"loss": 0.5422,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.498034076015728,
|
| 356 |
+
"grad_norm": 89.72924041748047,
|
| 357 |
+
"learning_rate": 1.6119837114601514e-05,
|
| 358 |
+
"loss": 0.4899,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.6290956749672345,
|
| 363 |
+
"grad_norm": 37.08826446533203,
|
| 364 |
+
"learning_rate": 1.5974403723094824e-05,
|
| 365 |
+
"loss": 0.3678,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 5.760157273918741,
|
| 370 |
+
"grad_norm": 15.998916625976562,
|
| 371 |
+
"learning_rate": 1.5828970331588135e-05,
|
| 372 |
+
"loss": 0.4723,
|
| 373 |
+
"step": 1100
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 5.891218872870249,
|
| 377 |
+
"grad_norm": 41.07233810424805,
|
| 378 |
+
"learning_rate": 1.5683536940081443e-05,
|
| 379 |
+
"loss": 0.5493,
|
| 380 |
+
"step": 1125
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 6.0,
|
| 384 |
+
"eval_accuracy": 0.8424657534246576,
|
| 385 |
+
"eval_loss": 0.6887933611869812,
|
| 386 |
+
"eval_macro_f1": 0.8419490901512526,
|
| 387 |
+
"eval_runtime": 0.4426,
|
| 388 |
+
"eval_samples_per_second": 1319.619,
|
| 389 |
+
"eval_steps_per_second": 83.606,
|
| 390 |
+
"eval_weighted_f1": 0.8419858826813973,
|
| 391 |
+
"step": 1146
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 6.0209698558322415,
|
| 395 |
+
"grad_norm": 65.35658264160156,
|
| 396 |
+
"learning_rate": 1.5538103548574753e-05,
|
| 397 |
+
"loss": 0.493,
|
| 398 |
+
"step": 1150
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.1520314547837485,
|
| 402 |
+
"grad_norm": 71.60026550292969,
|
| 403 |
+
"learning_rate": 1.5392670157068064e-05,
|
| 404 |
+
"loss": 0.3782,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.283093053735255,
|
| 409 |
+
"grad_norm": 5.998918056488037,
|
| 410 |
+
"learning_rate": 1.5247236765561375e-05,
|
| 411 |
+
"loss": 0.3062,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.414154652686763,
|
| 416 |
+
"grad_norm": 2.930375814437866,
|
| 417 |
+
"learning_rate": 1.5101803374054684e-05,
|
| 418 |
+
"loss": 0.5499,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.54521625163827,
|
| 423 |
+
"grad_norm": 0.9057067036628723,
|
| 424 |
+
"learning_rate": 1.4956369982547993e-05,
|
| 425 |
+
"loss": 0.4047,
|
| 426 |
+
"step": 1250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 6.676277850589777,
|
| 430 |
+
"grad_norm": 140.81387329101562,
|
| 431 |
+
"learning_rate": 1.4810936591041306e-05,
|
| 432 |
+
"loss": 0.4865,
|
| 433 |
+
"step": 1275
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 6.807339449541285,
|
| 437 |
+
"grad_norm": 0.4184958338737488,
|
| 438 |
+
"learning_rate": 1.4665503199534615e-05,
|
| 439 |
+
"loss": 0.5097,
|
| 440 |
+
"step": 1300
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 6.938401048492792,
|
| 444 |
+
"grad_norm": 29.483163833618164,
|
| 445 |
+
"learning_rate": 1.4520069808027924e-05,
|
| 446 |
+
"loss": 0.5279,
|
| 447 |
+
"step": 1325
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 7.0,
|
| 451 |
+
"eval_accuracy": 0.8458904109589042,
|
| 452 |
+
"eval_loss": 0.7323911786079407,
|
| 453 |
+
"eval_macro_f1": 0.8468749690930087,
|
| 454 |
+
"eval_runtime": 0.4847,
|
| 455 |
+
"eval_samples_per_second": 1204.837,
|
| 456 |
+
"eval_steps_per_second": 76.334,
|
| 457 |
+
"eval_weighted_f1": 0.8470004215396969,
|
| 458 |
+
"step": 1337
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 7.068152031454784,
|
| 462 |
+
"grad_norm": 0.747008740901947,
|
| 463 |
+
"learning_rate": 1.4374636416521235e-05,
|
| 464 |
+
"loss": 0.206,
|
| 465 |
+
"step": 1350
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 7.199213630406291,
|
| 469 |
+
"grad_norm": 2.603377342224121,
|
| 470 |
+
"learning_rate": 1.4229203025014544e-05,
|
| 471 |
+
"loss": 0.3824,
|
| 472 |
+
"step": 1375
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 7.330275229357798,
|
| 476 |
+
"grad_norm": 98.1949462890625,
|
| 477 |
+
"learning_rate": 1.4083769633507855e-05,
|
| 478 |
+
"loss": 0.3611,
|
| 479 |
+
"step": 1400
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 7.461336828309306,
|
| 483 |
+
"grad_norm": 0.2691371738910675,
|
| 484 |
+
"learning_rate": 1.3938336242001166e-05,
|
| 485 |
+
"loss": 0.3325,
|
| 486 |
+
"step": 1425
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 7.592398427260813,
|
| 490 |
+
"grad_norm": 0.576208233833313,
|
| 491 |
+
"learning_rate": 1.3792902850494475e-05,
|
| 492 |
+
"loss": 0.2876,
|
| 493 |
+
"step": 1450
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 7.72346002621232,
|
| 497 |
+
"grad_norm": 115.74394226074219,
|
| 498 |
+
"learning_rate": 1.3647469458987785e-05,
|
| 499 |
+
"loss": 0.4616,
|
| 500 |
+
"step": 1475
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 7.854521625163827,
|
| 504 |
+
"grad_norm": 94.09638977050781,
|
| 505 |
+
"learning_rate": 1.3502036067481095e-05,
|
| 506 |
+
"loss": 0.2895,
|
| 507 |
+
"step": 1500
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 7.985583224115334,
|
| 511 |
+
"grad_norm": 116.08271026611328,
|
| 512 |
+
"learning_rate": 1.3356602675974404e-05,
|
| 513 |
+
"loss": 0.3586,
|
| 514 |
+
"step": 1525
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 8.0,
|
| 518 |
+
"eval_accuracy": 0.839041095890411,
|
| 519 |
+
"eval_loss": 0.810036301612854,
|
| 520 |
+
"eval_macro_f1": 0.8403842984543092,
|
| 521 |
+
"eval_runtime": 0.4426,
|
| 522 |
+
"eval_samples_per_second": 1319.462,
|
| 523 |
+
"eval_steps_per_second": 83.596,
|
| 524 |
+
"eval_weighted_f1": 0.8405339427253326,
|
| 525 |
+
"step": 1528
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 8.115334207077327,
|
| 529 |
+
"grad_norm": 0.22293846309185028,
|
| 530 |
+
"learning_rate": 1.3211169284467716e-05,
|
| 531 |
+
"loss": 0.2152,
|
| 532 |
+
"step": 1550
|
| 533 |
+
},
|
| 534 |
+
{
|
| 535 |
+
"epoch": 8.246395806028833,
|
| 536 |
+
"grad_norm": 99.01307678222656,
|
| 537 |
+
"learning_rate": 1.3065735892961025e-05,
|
| 538 |
+
"loss": 0.3097,
|
| 539 |
+
"step": 1575
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 8.37745740498034,
|
| 543 |
+
"grad_norm": 10.559774398803711,
|
| 544 |
+
"learning_rate": 1.2920302501454336e-05,
|
| 545 |
+
"loss": 0.2277,
|
| 546 |
+
"step": 1600
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 8.508519003931848,
|
| 550 |
+
"grad_norm": 33.459228515625,
|
| 551 |
+
"learning_rate": 1.2774869109947645e-05,
|
| 552 |
+
"loss": 0.1517,
|
| 553 |
+
"step": 1625
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 8.639580602883354,
|
| 557 |
+
"grad_norm": 3.62520694732666,
|
| 558 |
+
"learning_rate": 1.2629435718440954e-05,
|
| 559 |
+
"loss": 0.3604,
|
| 560 |
+
"step": 1650
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 8.770642201834862,
|
| 564 |
+
"grad_norm": 0.7166460752487183,
|
| 565 |
+
"learning_rate": 1.2484002326934265e-05,
|
| 566 |
+
"loss": 0.3739,
|
| 567 |
+
"step": 1675
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 8.90170380078637,
|
| 571 |
+
"grad_norm": 107.67479705810547,
|
| 572 |
+
"learning_rate": 1.2338568935427576e-05,
|
| 573 |
+
"loss": 0.2157,
|
| 574 |
+
"step": 1700
|
| 575 |
+
},
|
| 576 |
+
{
|
| 577 |
+
"epoch": 9.0,
|
| 578 |
+
"eval_accuracy": 0.8407534246575342,
|
| 579 |
+
"eval_loss": 0.8536344766616821,
|
| 580 |
+
"eval_macro_f1": 0.8407841431483424,
|
| 581 |
+
"eval_runtime": 0.4237,
|
| 582 |
+
"eval_samples_per_second": 1378.461,
|
| 583 |
+
"eval_steps_per_second": 87.334,
|
| 584 |
+
"eval_weighted_f1": 0.8409023306212734,
|
| 585 |
+
"step": 1719
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 9.031454783748362,
|
| 589 |
+
"grad_norm": 9.85992431640625,
|
| 590 |
+
"learning_rate": 1.2193135543920885e-05,
|
| 591 |
+
"loss": 0.371,
|
| 592 |
+
"step": 1725
|
| 593 |
+
},
|
| 594 |
+
{
|
| 595 |
+
"epoch": 9.162516382699868,
|
| 596 |
+
"grad_norm": 0.29206278920173645,
|
| 597 |
+
"learning_rate": 1.2047702152414196e-05,
|
| 598 |
+
"loss": 0.2741,
|
| 599 |
+
"step": 1750
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 9.293577981651376,
|
| 603 |
+
"grad_norm": 1.4016391038894653,
|
| 604 |
+
"learning_rate": 1.1902268760907505e-05,
|
| 605 |
+
"loss": 0.214,
|
| 606 |
+
"step": 1775
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"epoch": 9.424639580602884,
|
| 610 |
+
"grad_norm": 0.14982491731643677,
|
| 611 |
+
"learning_rate": 1.1756835369400816e-05,
|
| 612 |
+
"loss": 0.1837,
|
| 613 |
+
"step": 1800
|
| 614 |
+
},
|
| 615 |
+
{
|
| 616 |
+
"epoch": 9.55570117955439,
|
| 617 |
+
"grad_norm": 0.143387109041214,
|
| 618 |
+
"learning_rate": 1.1611401977894125e-05,
|
| 619 |
+
"loss": 0.2936,
|
| 620 |
+
"step": 1825
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 9.686762778505898,
|
| 624 |
+
"grad_norm": 0.19346462190151215,
|
| 625 |
+
"learning_rate": 1.1465968586387436e-05,
|
| 626 |
+
"loss": 0.2196,
|
| 627 |
+
"step": 1850
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 9.817824377457406,
|
| 631 |
+
"grad_norm": 0.17706957459449768,
|
| 632 |
+
"learning_rate": 1.1320535194880746e-05,
|
| 633 |
+
"loss": 0.2475,
|
| 634 |
+
"step": 1875
|
| 635 |
+
},
|
| 636 |
+
{
|
| 637 |
+
"epoch": 9.948885976408912,
|
| 638 |
+
"grad_norm": 0.1315685361623764,
|
| 639 |
+
"learning_rate": 1.1175101803374056e-05,
|
| 640 |
+
"loss": 0.0871,
|
| 641 |
+
"step": 1900
|
| 642 |
+
},
|
| 643 |
+
{
|
| 644 |
+
"epoch": 10.0,
|
| 645 |
+
"eval_accuracy": 0.8424657534246576,
|
| 646 |
+
"eval_loss": 0.8965077996253967,
|
| 647 |
+
"eval_macro_f1": 0.8424415910050449,
|
| 648 |
+
"eval_runtime": 0.438,
|
| 649 |
+
"eval_samples_per_second": 1333.421,
|
| 650 |
+
"eval_steps_per_second": 84.48,
|
| 651 |
+
"eval_weighted_f1": 0.8424040675789648,
|
| 652 |
+
"step": 1910
|
| 653 |
+
}
|
| 654 |
+
],
|
| 655 |
+
"logging_steps": 25,
|
| 656 |
+
"max_steps": 3820,
|
| 657 |
+
"num_input_tokens_seen": 0,
|
| 658 |
+
"num_train_epochs": 20,
|
| 659 |
+
"save_steps": 500,
|
| 660 |
+
"stateful_callbacks": {
|
| 661 |
+
"EarlyStoppingCallback": {
|
| 662 |
+
"args": {
|
| 663 |
+
"early_stopping_patience": 3,
|
| 664 |
+
"early_stopping_threshold": 0.0
|
| 665 |
+
},
|
| 666 |
+
"attributes": {
|
| 667 |
+
"early_stopping_patience_counter": 3
|
| 668 |
+
}
|
| 669 |
+
},
|
| 670 |
+
"TrainerControl": {
|
| 671 |
+
"args": {
|
| 672 |
+
"should_epoch_stop": false,
|
| 673 |
+
"should_evaluate": false,
|
| 674 |
+
"should_log": false,
|
| 675 |
+
"should_save": true,
|
| 676 |
+
"should_training_stop": true
|
| 677 |
+
},
|
| 678 |
+
"attributes": {}
|
| 679 |
+
}
|
| 680 |
+
},
|
| 681 |
+
"total_flos": 0.0,
|
| 682 |
+
"train_batch_size": 8,
|
| 683 |
+
"trial_name": null,
|
| 684 |
+
"trial_params": null
|
| 685 |
+
}
|
ood-vifactcheck-phobert-topic-domain-01-s44-20260826T110548Z-ab4112e3/trainer/checkpoint-1910/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-02-s44-20260826T110927Z-ab4112e3/trainer/checkpoint-1274/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-topic-domain-02-s44-20260826T110927Z-ab4112e3/trainer/checkpoint-1274/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-02-s44-20260826T110927Z-ab4112e3/trainer/checkpoint-1274/special_tokens_map.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": "<mask>",
|
| 6 |
+
"pad_token": "<pad>",
|
| 7 |
+
"sep_token": "</s>",
|
| 8 |
+
"unk_token": "<unk>"
|
| 9 |
+
}
|
ood-vifactcheck-phobert-topic-domain-02-s44-20260826T110927Z-ab4112e3/trainer/checkpoint-1274/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"64000": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 256,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "PhobertTokenizer",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-phobert-topic-domain-02-s44-20260826T110927Z-ab4112e3/trainer/checkpoint-1274/trainer_state.json
ADDED
|
@@ -0,0 +1,470 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 728,
|
| 3 |
+
"best_metric": 0.8544001042022827,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-topic-domain-02-s44-20260826T110927Z-ab4112e3/trainer/checkpoint-728",
|
| 5 |
+
"epoch": 7.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1274,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.13793103448275862,
|
| 14 |
+
"grad_norm": 6.872440814971924,
|
| 15 |
+
"learning_rate": 1.3186813186813187e-06,
|
| 16 |
+
"loss": 4.4262,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.27586206896551724,
|
| 21 |
+
"grad_norm": 11.044365882873535,
|
| 22 |
+
"learning_rate": 2.6923076923076923e-06,
|
| 23 |
+
"loss": 4.4173,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.41379310344827586,
|
| 28 |
+
"grad_norm": 8.799114227294922,
|
| 29 |
+
"learning_rate": 4.065934065934066e-06,
|
| 30 |
+
"loss": 4.3499,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5517241379310345,
|
| 35 |
+
"grad_norm": 6.819574356079102,
|
| 36 |
+
"learning_rate": 5.43956043956044e-06,
|
| 37 |
+
"loss": 4.3331,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6896551724137931,
|
| 42 |
+
"grad_norm": 9.603765487670898,
|
| 43 |
+
"learning_rate": 6.813186813186814e-06,
|
| 44 |
+
"loss": 4.2531,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8275862068965517,
|
| 49 |
+
"grad_norm": 15.39643383026123,
|
| 50 |
+
"learning_rate": 8.186813186813188e-06,
|
| 51 |
+
"loss": 4.1812,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9655172413793104,
|
| 56 |
+
"grad_norm": 33.583702087402344,
|
| 57 |
+
"learning_rate": 9.560439560439562e-06,
|
| 58 |
+
"loss": 3.9498,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.5506958250497018,
|
| 64 |
+
"eval_loss": 0.9701381325721741,
|
| 65 |
+
"eval_macro_f1": 0.5309041126146968,
|
| 66 |
+
"eval_runtime": 0.401,
|
| 67 |
+
"eval_samples_per_second": 1254.212,
|
| 68 |
+
"eval_steps_per_second": 79.791,
|
| 69 |
+
"eval_weighted_f1": 0.5325482882964178,
|
| 70 |
+
"step": 182
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0993103448275863,
|
| 74 |
+
"grad_norm": 19.458309173583984,
|
| 75 |
+
"learning_rate": 1.0934065934065933e-05,
|
| 76 |
+
"loss": 3.6788,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.237241379310345,
|
| 81 |
+
"grad_norm": 26.95549201965332,
|
| 82 |
+
"learning_rate": 1.230769230769231e-05,
|
| 83 |
+
"loss": 3.3084,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.3751724137931034,
|
| 88 |
+
"grad_norm": 27.57164764404297,
|
| 89 |
+
"learning_rate": 1.3681318681318683e-05,
|
| 90 |
+
"loss": 3.1112,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.513103448275862,
|
| 95 |
+
"grad_norm": 23.33271598815918,
|
| 96 |
+
"learning_rate": 1.5054945054945056e-05,
|
| 97 |
+
"loss": 2.7621,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.6510344827586207,
|
| 102 |
+
"grad_norm": 45.245548248291016,
|
| 103 |
+
"learning_rate": 1.642857142857143e-05,
|
| 104 |
+
"loss": 2.562,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7889655172413792,
|
| 109 |
+
"grad_norm": 34.93929672241211,
|
| 110 |
+
"learning_rate": 1.78021978021978e-05,
|
| 111 |
+
"loss": 2.4925,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.926896551724138,
|
| 116 |
+
"grad_norm": 32.8734130859375,
|
| 117 |
+
"learning_rate": 1.9175824175824178e-05,
|
| 118 |
+
"loss": 2.3083,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.8011928429423459,
|
| 124 |
+
"eval_loss": 0.4829344153404236,
|
| 125 |
+
"eval_macro_f1": 0.7950994844873506,
|
| 126 |
+
"eval_runtime": 0.4394,
|
| 127 |
+
"eval_samples_per_second": 1144.737,
|
| 128 |
+
"eval_steps_per_second": 72.826,
|
| 129 |
+
"eval_weighted_f1": 0.7955514729742386,
|
| 130 |
+
"step": 364
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.0606896551724136,
|
| 134 |
+
"grad_norm": 23.684185028076172,
|
| 135 |
+
"learning_rate": 1.993894993894994e-05,
|
| 136 |
+
"loss": 1.7555,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.1986206896551725,
|
| 141 |
+
"grad_norm": 38.884883880615234,
|
| 142 |
+
"learning_rate": 1.9786324786324788e-05,
|
| 143 |
+
"loss": 1.7633,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.336551724137931,
|
| 148 |
+
"grad_norm": 39.6777458190918,
|
| 149 |
+
"learning_rate": 1.9633699633699634e-05,
|
| 150 |
+
"loss": 1.6774,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.47448275862069,
|
| 155 |
+
"grad_norm": 22.136768341064453,
|
| 156 |
+
"learning_rate": 1.948107448107448e-05,
|
| 157 |
+
"loss": 1.7496,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.6124137931034483,
|
| 162 |
+
"grad_norm": 34.138729095458984,
|
| 163 |
+
"learning_rate": 1.932844932844933e-05,
|
| 164 |
+
"loss": 1.75,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.750344827586207,
|
| 169 |
+
"grad_norm": 21.8614501953125,
|
| 170 |
+
"learning_rate": 1.9175824175824178e-05,
|
| 171 |
+
"loss": 1.8152,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.8882758620689657,
|
| 176 |
+
"grad_norm": 28.91727066040039,
|
| 177 |
+
"learning_rate": 1.9023199023199025e-05,
|
| 178 |
+
"loss": 1.523,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 3.0,
|
| 183 |
+
"eval_accuracy": 0.8469184890656064,
|
| 184 |
+
"eval_loss": 0.40458178520202637,
|
| 185 |
+
"eval_macro_f1": 0.8465678071015414,
|
| 186 |
+
"eval_runtime": 0.4117,
|
| 187 |
+
"eval_samples_per_second": 1221.904,
|
| 188 |
+
"eval_steps_per_second": 77.735,
|
| 189 |
+
"eval_weighted_f1": 0.8469680747634899,
|
| 190 |
+
"step": 546
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.0220689655172412,
|
| 194 |
+
"grad_norm": 30.68318748474121,
|
| 195 |
+
"learning_rate": 1.887057387057387e-05,
|
| 196 |
+
"loss": 1.4116,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.16,
|
| 201 |
+
"grad_norm": 16.432405471801758,
|
| 202 |
+
"learning_rate": 1.8717948717948718e-05,
|
| 203 |
+
"loss": 1.0403,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.2979310344827586,
|
| 208 |
+
"grad_norm": 27.419424057006836,
|
| 209 |
+
"learning_rate": 1.8565323565323568e-05,
|
| 210 |
+
"loss": 1.1421,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.435862068965517,
|
| 215 |
+
"grad_norm": 67.66222381591797,
|
| 216 |
+
"learning_rate": 1.8412698412698415e-05,
|
| 217 |
+
"loss": 1.2681,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.573793103448276,
|
| 222 |
+
"grad_norm": 32.518653869628906,
|
| 223 |
+
"learning_rate": 1.826007326007326e-05,
|
| 224 |
+
"loss": 1.1212,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.7117241379310344,
|
| 229 |
+
"grad_norm": 64.80052185058594,
|
| 230 |
+
"learning_rate": 1.8107448107448108e-05,
|
| 231 |
+
"loss": 1.0342,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.849655172413793,
|
| 236 |
+
"grad_norm": 54.30545425415039,
|
| 237 |
+
"learning_rate": 1.7954822954822958e-05,
|
| 238 |
+
"loss": 1.1861,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.9875862068965517,
|
| 243 |
+
"grad_norm": 61.86852264404297,
|
| 244 |
+
"learning_rate": 1.78021978021978e-05,
|
| 245 |
+
"loss": 1.1703,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 4.0,
|
| 250 |
+
"eval_accuracy": 0.856858846918489,
|
| 251 |
+
"eval_loss": 0.3991502523422241,
|
| 252 |
+
"eval_macro_f1": 0.8544001042022827,
|
| 253 |
+
"eval_runtime": 0.4014,
|
| 254 |
+
"eval_samples_per_second": 1253.022,
|
| 255 |
+
"eval_steps_per_second": 79.715,
|
| 256 |
+
"eval_weighted_f1": 0.8548356226146273,
|
| 257 |
+
"step": 728
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.121379310344827,
|
| 261 |
+
"grad_norm": 30.64313316345215,
|
| 262 |
+
"learning_rate": 1.764957264957265e-05,
|
| 263 |
+
"loss": 0.7988,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.259310344827586,
|
| 268 |
+
"grad_norm": 25.8349609375,
|
| 269 |
+
"learning_rate": 1.74969474969475e-05,
|
| 270 |
+
"loss": 0.8099,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.397241379310345,
|
| 275 |
+
"grad_norm": 57.631553649902344,
|
| 276 |
+
"learning_rate": 1.7344322344322345e-05,
|
| 277 |
+
"loss": 0.835,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.535172413793103,
|
| 282 |
+
"grad_norm": 39.93080520629883,
|
| 283 |
+
"learning_rate": 1.7191697191697195e-05,
|
| 284 |
+
"loss": 0.7322,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.673103448275862,
|
| 289 |
+
"grad_norm": 57.785396575927734,
|
| 290 |
+
"learning_rate": 1.703907203907204e-05,
|
| 291 |
+
"loss": 0.8504,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.811034482758621,
|
| 296 |
+
"grad_norm": 74.5561752319336,
|
| 297 |
+
"learning_rate": 1.688644688644689e-05,
|
| 298 |
+
"loss": 0.6595,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.94896551724138,
|
| 303 |
+
"grad_norm": 44.875144958496094,
|
| 304 |
+
"learning_rate": 1.6733821733821735e-05,
|
| 305 |
+
"loss": 0.6292,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 5.0,
|
| 310 |
+
"eval_accuracy": 0.8469184890656064,
|
| 311 |
+
"eval_loss": 0.4885314702987671,
|
| 312 |
+
"eval_macro_f1": 0.8471428146899719,
|
| 313 |
+
"eval_runtime": 0.3828,
|
| 314 |
+
"eval_samples_per_second": 1313.982,
|
| 315 |
+
"eval_steps_per_second": 83.593,
|
| 316 |
+
"eval_weighted_f1": 0.8472744001719064,
|
| 317 |
+
"step": 910
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.082758620689655,
|
| 321 |
+
"grad_norm": 12.291990280151367,
|
| 322 |
+
"learning_rate": 1.6581196581196585e-05,
|
| 323 |
+
"loss": 0.5211,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.220689655172414,
|
| 328 |
+
"grad_norm": 39.85081100463867,
|
| 329 |
+
"learning_rate": 1.642857142857143e-05,
|
| 330 |
+
"loss": 0.5671,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.358620689655172,
|
| 335 |
+
"grad_norm": 33.413169860839844,
|
| 336 |
+
"learning_rate": 1.627594627594628e-05,
|
| 337 |
+
"loss": 0.3554,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.496551724137931,
|
| 342 |
+
"grad_norm": 1.0892884731292725,
|
| 343 |
+
"learning_rate": 1.6123321123321125e-05,
|
| 344 |
+
"loss": 0.6082,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.63448275862069,
|
| 349 |
+
"grad_norm": 89.72523498535156,
|
| 350 |
+
"learning_rate": 1.5970695970695972e-05,
|
| 351 |
+
"loss": 0.5899,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.772413793103448,
|
| 356 |
+
"grad_norm": 55.45600891113281,
|
| 357 |
+
"learning_rate": 1.581807081807082e-05,
|
| 358 |
+
"loss": 0.6102,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.910344827586207,
|
| 363 |
+
"grad_norm": 36.698997497558594,
|
| 364 |
+
"learning_rate": 1.5665445665445665e-05,
|
| 365 |
+
"loss": 0.715,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 6.0,
|
| 370 |
+
"eval_accuracy": 0.856858846918489,
|
| 371 |
+
"eval_loss": 0.5660998821258545,
|
| 372 |
+
"eval_macro_f1": 0.8543716595079601,
|
| 373 |
+
"eval_runtime": 0.4035,
|
| 374 |
+
"eval_samples_per_second": 1246.465,
|
| 375 |
+
"eval_steps_per_second": 79.298,
|
| 376 |
+
"eval_weighted_f1": 0.8546059917170999,
|
| 377 |
+
"step": 1092
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 6.0441379310344825,
|
| 381 |
+
"grad_norm": 1.9779577255249023,
|
| 382 |
+
"learning_rate": 1.5512820512820516e-05,
|
| 383 |
+
"loss": 0.5175,
|
| 384 |
+
"step": 1100
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 6.182068965517241,
|
| 388 |
+
"grad_norm": 17.767475128173828,
|
| 389 |
+
"learning_rate": 1.5360195360195362e-05,
|
| 390 |
+
"loss": 0.3986,
|
| 391 |
+
"step": 1125
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 6.32,
|
| 395 |
+
"grad_norm": 69.13174438476562,
|
| 396 |
+
"learning_rate": 1.5207570207570207e-05,
|
| 397 |
+
"loss": 0.4695,
|
| 398 |
+
"step": 1150
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.457931034482758,
|
| 402 |
+
"grad_norm": 3.363964319229126,
|
| 403 |
+
"learning_rate": 1.5054945054945056e-05,
|
| 404 |
+
"loss": 0.3081,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.595862068965517,
|
| 409 |
+
"grad_norm": 1.140702486038208,
|
| 410 |
+
"learning_rate": 1.4902319902319904e-05,
|
| 411 |
+
"loss": 0.4707,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.733793103448276,
|
| 416 |
+
"grad_norm": 29.39591407775879,
|
| 417 |
+
"learning_rate": 1.4749694749694752e-05,
|
| 418 |
+
"loss": 0.5393,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.871724137931034,
|
| 423 |
+
"grad_norm": 0.6890884637832642,
|
| 424 |
+
"learning_rate": 1.4597069597069597e-05,
|
| 425 |
+
"loss": 0.4053,
|
| 426 |
+
"step": 1250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 7.0,
|
| 430 |
+
"eval_accuracy": 0.831013916500994,
|
| 431 |
+
"eval_loss": 0.7072443962097168,
|
| 432 |
+
"eval_macro_f1": 0.8315162574950369,
|
| 433 |
+
"eval_runtime": 0.3956,
|
| 434 |
+
"eval_samples_per_second": 1271.545,
|
| 435 |
+
"eval_steps_per_second": 80.894,
|
| 436 |
+
"eval_weighted_f1": 0.8314949998191277,
|
| 437 |
+
"step": 1274
|
| 438 |
+
}
|
| 439 |
+
],
|
| 440 |
+
"logging_steps": 25,
|
| 441 |
+
"max_steps": 3640,
|
| 442 |
+
"num_input_tokens_seen": 0,
|
| 443 |
+
"num_train_epochs": 20,
|
| 444 |
+
"save_steps": 500,
|
| 445 |
+
"stateful_callbacks": {
|
| 446 |
+
"EarlyStoppingCallback": {
|
| 447 |
+
"args": {
|
| 448 |
+
"early_stopping_patience": 3,
|
| 449 |
+
"early_stopping_threshold": 0.0
|
| 450 |
+
},
|
| 451 |
+
"attributes": {
|
| 452 |
+
"early_stopping_patience_counter": 3
|
| 453 |
+
}
|
| 454 |
+
},
|
| 455 |
+
"TrainerControl": {
|
| 456 |
+
"args": {
|
| 457 |
+
"should_epoch_stop": false,
|
| 458 |
+
"should_evaluate": false,
|
| 459 |
+
"should_log": false,
|
| 460 |
+
"should_save": true,
|
| 461 |
+
"should_training_stop": true
|
| 462 |
+
},
|
| 463 |
+
"attributes": {}
|
| 464 |
+
}
|
| 465 |
+
},
|
| 466 |
+
"total_flos": 0.0,
|
| 467 |
+
"train_batch_size": 8,
|
| 468 |
+
"trial_name": null,
|
| 469 |
+
"trial_params": null
|
| 470 |
+
}
|
ood-vifactcheck-phobert-topic-domain-02-s44-20260826T110927Z-ab4112e3/trainer/checkpoint-1274/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-02-s44-20260826T110927Z-ab4112e3/trainer/checkpoint-728/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-topic-domain-02-s44-20260826T110927Z-ab4112e3/trainer/checkpoint-728/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-02-s44-20260826T110927Z-ab4112e3/trainer/checkpoint-728/special_tokens_map.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": "<mask>",
|
| 6 |
+
"pad_token": "<pad>",
|
| 7 |
+
"sep_token": "</s>",
|
| 8 |
+
"unk_token": "<unk>"
|
| 9 |
+
}
|
ood-vifactcheck-phobert-topic-domain-02-s44-20260826T110927Z-ab4112e3/trainer/checkpoint-728/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"64000": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 256,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "PhobertTokenizer",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-phobert-topic-domain-02-s44-20260826T110927Z-ab4112e3/trainer/checkpoint-728/trainer_state.json
ADDED
|
@@ -0,0 +1,290 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 728,
|
| 3 |
+
"best_metric": 0.8544001042022827,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-topic-domain-02-s44-20260826T110927Z-ab4112e3/trainer/checkpoint-728",
|
| 5 |
+
"epoch": 4.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 728,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.13793103448275862,
|
| 14 |
+
"grad_norm": 6.872440814971924,
|
| 15 |
+
"learning_rate": 1.3186813186813187e-06,
|
| 16 |
+
"loss": 4.4262,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.27586206896551724,
|
| 21 |
+
"grad_norm": 11.044365882873535,
|
| 22 |
+
"learning_rate": 2.6923076923076923e-06,
|
| 23 |
+
"loss": 4.4173,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.41379310344827586,
|
| 28 |
+
"grad_norm": 8.799114227294922,
|
| 29 |
+
"learning_rate": 4.065934065934066e-06,
|
| 30 |
+
"loss": 4.3499,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5517241379310345,
|
| 35 |
+
"grad_norm": 6.819574356079102,
|
| 36 |
+
"learning_rate": 5.43956043956044e-06,
|
| 37 |
+
"loss": 4.3331,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6896551724137931,
|
| 42 |
+
"grad_norm": 9.603765487670898,
|
| 43 |
+
"learning_rate": 6.813186813186814e-06,
|
| 44 |
+
"loss": 4.2531,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8275862068965517,
|
| 49 |
+
"grad_norm": 15.39643383026123,
|
| 50 |
+
"learning_rate": 8.186813186813188e-06,
|
| 51 |
+
"loss": 4.1812,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9655172413793104,
|
| 56 |
+
"grad_norm": 33.583702087402344,
|
| 57 |
+
"learning_rate": 9.560439560439562e-06,
|
| 58 |
+
"loss": 3.9498,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.5506958250497018,
|
| 64 |
+
"eval_loss": 0.9701381325721741,
|
| 65 |
+
"eval_macro_f1": 0.5309041126146968,
|
| 66 |
+
"eval_runtime": 0.401,
|
| 67 |
+
"eval_samples_per_second": 1254.212,
|
| 68 |
+
"eval_steps_per_second": 79.791,
|
| 69 |
+
"eval_weighted_f1": 0.5325482882964178,
|
| 70 |
+
"step": 182
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0993103448275863,
|
| 74 |
+
"grad_norm": 19.458309173583984,
|
| 75 |
+
"learning_rate": 1.0934065934065933e-05,
|
| 76 |
+
"loss": 3.6788,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.237241379310345,
|
| 81 |
+
"grad_norm": 26.95549201965332,
|
| 82 |
+
"learning_rate": 1.230769230769231e-05,
|
| 83 |
+
"loss": 3.3084,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.3751724137931034,
|
| 88 |
+
"grad_norm": 27.57164764404297,
|
| 89 |
+
"learning_rate": 1.3681318681318683e-05,
|
| 90 |
+
"loss": 3.1112,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.513103448275862,
|
| 95 |
+
"grad_norm": 23.33271598815918,
|
| 96 |
+
"learning_rate": 1.5054945054945056e-05,
|
| 97 |
+
"loss": 2.7621,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.6510344827586207,
|
| 102 |
+
"grad_norm": 45.245548248291016,
|
| 103 |
+
"learning_rate": 1.642857142857143e-05,
|
| 104 |
+
"loss": 2.562,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7889655172413792,
|
| 109 |
+
"grad_norm": 34.93929672241211,
|
| 110 |
+
"learning_rate": 1.78021978021978e-05,
|
| 111 |
+
"loss": 2.4925,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.926896551724138,
|
| 116 |
+
"grad_norm": 32.8734130859375,
|
| 117 |
+
"learning_rate": 1.9175824175824178e-05,
|
| 118 |
+
"loss": 2.3083,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.8011928429423459,
|
| 124 |
+
"eval_loss": 0.4829344153404236,
|
| 125 |
+
"eval_macro_f1": 0.7950994844873506,
|
| 126 |
+
"eval_runtime": 0.4394,
|
| 127 |
+
"eval_samples_per_second": 1144.737,
|
| 128 |
+
"eval_steps_per_second": 72.826,
|
| 129 |
+
"eval_weighted_f1": 0.7955514729742386,
|
| 130 |
+
"step": 364
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.0606896551724136,
|
| 134 |
+
"grad_norm": 23.684185028076172,
|
| 135 |
+
"learning_rate": 1.993894993894994e-05,
|
| 136 |
+
"loss": 1.7555,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.1986206896551725,
|
| 141 |
+
"grad_norm": 38.884883880615234,
|
| 142 |
+
"learning_rate": 1.9786324786324788e-05,
|
| 143 |
+
"loss": 1.7633,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.336551724137931,
|
| 148 |
+
"grad_norm": 39.6777458190918,
|
| 149 |
+
"learning_rate": 1.9633699633699634e-05,
|
| 150 |
+
"loss": 1.6774,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.47448275862069,
|
| 155 |
+
"grad_norm": 22.136768341064453,
|
| 156 |
+
"learning_rate": 1.948107448107448e-05,
|
| 157 |
+
"loss": 1.7496,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.6124137931034483,
|
| 162 |
+
"grad_norm": 34.138729095458984,
|
| 163 |
+
"learning_rate": 1.932844932844933e-05,
|
| 164 |
+
"loss": 1.75,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.750344827586207,
|
| 169 |
+
"grad_norm": 21.8614501953125,
|
| 170 |
+
"learning_rate": 1.9175824175824178e-05,
|
| 171 |
+
"loss": 1.8152,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.8882758620689657,
|
| 176 |
+
"grad_norm": 28.91727066040039,
|
| 177 |
+
"learning_rate": 1.9023199023199025e-05,
|
| 178 |
+
"loss": 1.523,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 3.0,
|
| 183 |
+
"eval_accuracy": 0.8469184890656064,
|
| 184 |
+
"eval_loss": 0.40458178520202637,
|
| 185 |
+
"eval_macro_f1": 0.8465678071015414,
|
| 186 |
+
"eval_runtime": 0.4117,
|
| 187 |
+
"eval_samples_per_second": 1221.904,
|
| 188 |
+
"eval_steps_per_second": 77.735,
|
| 189 |
+
"eval_weighted_f1": 0.8469680747634899,
|
| 190 |
+
"step": 546
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.0220689655172412,
|
| 194 |
+
"grad_norm": 30.68318748474121,
|
| 195 |
+
"learning_rate": 1.887057387057387e-05,
|
| 196 |
+
"loss": 1.4116,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.16,
|
| 201 |
+
"grad_norm": 16.432405471801758,
|
| 202 |
+
"learning_rate": 1.8717948717948718e-05,
|
| 203 |
+
"loss": 1.0403,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.2979310344827586,
|
| 208 |
+
"grad_norm": 27.419424057006836,
|
| 209 |
+
"learning_rate": 1.8565323565323568e-05,
|
| 210 |
+
"loss": 1.1421,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.435862068965517,
|
| 215 |
+
"grad_norm": 67.66222381591797,
|
| 216 |
+
"learning_rate": 1.8412698412698415e-05,
|
| 217 |
+
"loss": 1.2681,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.573793103448276,
|
| 222 |
+
"grad_norm": 32.518653869628906,
|
| 223 |
+
"learning_rate": 1.826007326007326e-05,
|
| 224 |
+
"loss": 1.1212,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.7117241379310344,
|
| 229 |
+
"grad_norm": 64.80052185058594,
|
| 230 |
+
"learning_rate": 1.8107448107448108e-05,
|
| 231 |
+
"loss": 1.0342,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.849655172413793,
|
| 236 |
+
"grad_norm": 54.30545425415039,
|
| 237 |
+
"learning_rate": 1.7954822954822958e-05,
|
| 238 |
+
"loss": 1.1861,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.9875862068965517,
|
| 243 |
+
"grad_norm": 61.86852264404297,
|
| 244 |
+
"learning_rate": 1.78021978021978e-05,
|
| 245 |
+
"loss": 1.1703,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 4.0,
|
| 250 |
+
"eval_accuracy": 0.856858846918489,
|
| 251 |
+
"eval_loss": 0.3991502523422241,
|
| 252 |
+
"eval_macro_f1": 0.8544001042022827,
|
| 253 |
+
"eval_runtime": 0.4014,
|
| 254 |
+
"eval_samples_per_second": 1253.022,
|
| 255 |
+
"eval_steps_per_second": 79.715,
|
| 256 |
+
"eval_weighted_f1": 0.8548356226146273,
|
| 257 |
+
"step": 728
|
| 258 |
+
}
|
| 259 |
+
],
|
| 260 |
+
"logging_steps": 25,
|
| 261 |
+
"max_steps": 3640,
|
| 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-02-s44-20260826T110927Z-ab4112e3/trainer/checkpoint-728/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-03-s44-20260826T111158Z-ab4112e3/trainer/checkpoint-1170/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-topic-domain-03-s44-20260826T111158Z-ab4112e3/trainer/checkpoint-1170/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-03-s44-20260826T111158Z-ab4112e3/trainer/checkpoint-1170/special_tokens_map.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": "<mask>",
|
| 6 |
+
"pad_token": "<pad>",
|
| 7 |
+
"sep_token": "</s>",
|
| 8 |
+
"unk_token": "<unk>"
|
| 9 |
+
}
|
ood-vifactcheck-phobert-topic-domain-03-s44-20260826T111158Z-ab4112e3/trainer/checkpoint-1170/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"64000": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 256,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "PhobertTokenizer",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-phobert-topic-domain-03-s44-20260826T111158Z-ab4112e3/trainer/checkpoint-1170/trainer_state.json
ADDED
|
@@ -0,0 +1,431 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1170,
|
| 3 |
+
"best_metric": 0.8557374806643535,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-topic-domain-03-s44-20260826T111158Z-ab4112e3/trainer/checkpoint-1170",
|
| 5 |
+
"epoch": 6.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1170,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.12836970474967907,
|
| 14 |
+
"grad_norm": 7.355264186859131,
|
| 15 |
+
"learning_rate": 1.230769230769231e-06,
|
| 16 |
+
"loss": 4.42,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.25673940949935814,
|
| 21 |
+
"grad_norm": 7.894757270812988,
|
| 22 |
+
"learning_rate": 2.512820512820513e-06,
|
| 23 |
+
"loss": 4.3766,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.3851091142490372,
|
| 28 |
+
"grad_norm": 7.210986137390137,
|
| 29 |
+
"learning_rate": 3.794871794871795e-06,
|
| 30 |
+
"loss": 4.3868,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5134788189987163,
|
| 35 |
+
"grad_norm": 9.210958480834961,
|
| 36 |
+
"learning_rate": 5.076923076923077e-06,
|
| 37 |
+
"loss": 4.3375,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6418485237483954,
|
| 42 |
+
"grad_norm": 10.58199405670166,
|
| 43 |
+
"learning_rate": 6.358974358974359e-06,
|
| 44 |
+
"loss": 4.2921,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.7702182284980744,
|
| 49 |
+
"grad_norm": 15.043291091918945,
|
| 50 |
+
"learning_rate": 7.641025641025641e-06,
|
| 51 |
+
"loss": 4.1014,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.8985879332477535,
|
| 56 |
+
"grad_norm": 21.792621612548828,
|
| 57 |
+
"learning_rate": 8.923076923076925e-06,
|
| 58 |
+
"loss": 4.0205,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.6366782006920415,
|
| 64 |
+
"eval_loss": 0.8276789784431458,
|
| 65 |
+
"eval_macro_f1": 0.6315092173660761,
|
| 66 |
+
"eval_runtime": 0.4133,
|
| 67 |
+
"eval_samples_per_second": 1398.41,
|
| 68 |
+
"eval_steps_per_second": 89.518,
|
| 69 |
+
"eval_weighted_f1": 0.6345556061280964,
|
| 70 |
+
"step": 195
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0256739409499358,
|
| 74 |
+
"grad_norm": 45.431007385253906,
|
| 75 |
+
"learning_rate": 1.0205128205128205e-05,
|
| 76 |
+
"loss": 3.6374,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.1540436456996148,
|
| 81 |
+
"grad_norm": 38.621925354003906,
|
| 82 |
+
"learning_rate": 1.1487179487179487e-05,
|
| 83 |
+
"loss": 3.2732,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.282413350449294,
|
| 88 |
+
"grad_norm": 41.117637634277344,
|
| 89 |
+
"learning_rate": 1.2769230769230769e-05,
|
| 90 |
+
"loss": 3.1589,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.410783055198973,
|
| 95 |
+
"grad_norm": 44.818180084228516,
|
| 96 |
+
"learning_rate": 1.405128205128205e-05,
|
| 97 |
+
"loss": 2.8765,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.539152759948652,
|
| 102 |
+
"grad_norm": 46.80485916137695,
|
| 103 |
+
"learning_rate": 1.5333333333333334e-05,
|
| 104 |
+
"loss": 2.6511,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.6675224646983313,
|
| 109 |
+
"grad_norm": 33.0837287902832,
|
| 110 |
+
"learning_rate": 1.6615384615384618e-05,
|
| 111 |
+
"loss": 2.4328,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.7958921694480101,
|
| 116 |
+
"grad_norm": 34.05903244018555,
|
| 117 |
+
"learning_rate": 1.78974358974359e-05,
|
| 118 |
+
"loss": 2.1861,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 1.9242618741976893,
|
| 123 |
+
"grad_norm": 33.80216598510742,
|
| 124 |
+
"learning_rate": 1.9179487179487182e-05,
|
| 125 |
+
"loss": 2.1043,
|
| 126 |
+
"step": 375
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"epoch": 2.0,
|
| 130 |
+
"eval_accuracy": 0.7958477508650519,
|
| 131 |
+
"eval_loss": 0.5350261330604553,
|
| 132 |
+
"eval_macro_f1": 0.786443465956267,
|
| 133 |
+
"eval_runtime": 0.4296,
|
| 134 |
+
"eval_samples_per_second": 1345.294,
|
| 135 |
+
"eval_steps_per_second": 86.117,
|
| 136 |
+
"eval_weighted_f1": 0.7878405892780137,
|
| 137 |
+
"step": 390
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.0513478818998716,
|
| 141 |
+
"grad_norm": 49.66297912597656,
|
| 142 |
+
"learning_rate": 1.994871794871795e-05,
|
| 143 |
+
"loss": 2.0522,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.179717586649551,
|
| 148 |
+
"grad_norm": 56.68103790283203,
|
| 149 |
+
"learning_rate": 1.9806267806267806e-05,
|
| 150 |
+
"loss": 1.8106,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.3080872913992296,
|
| 155 |
+
"grad_norm": 54.557979583740234,
|
| 156 |
+
"learning_rate": 1.9663817663817667e-05,
|
| 157 |
+
"loss": 1.7066,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.436456996148909,
|
| 162 |
+
"grad_norm": 36.9579963684082,
|
| 163 |
+
"learning_rate": 1.952136752136752e-05,
|
| 164 |
+
"loss": 1.6575,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.564826700898588,
|
| 169 |
+
"grad_norm": 38.816524505615234,
|
| 170 |
+
"learning_rate": 1.9378917378917382e-05,
|
| 171 |
+
"loss": 1.5141,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.693196405648267,
|
| 176 |
+
"grad_norm": 18.72123146057129,
|
| 177 |
+
"learning_rate": 1.923646723646724e-05,
|
| 178 |
+
"loss": 1.3195,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 2.821566110397946,
|
| 183 |
+
"grad_norm": 31.898258209228516,
|
| 184 |
+
"learning_rate": 1.9094017094017097e-05,
|
| 185 |
+
"loss": 1.5703,
|
| 186 |
+
"step": 550
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 2.949935815147625,
|
| 190 |
+
"grad_norm": 33.38069534301758,
|
| 191 |
+
"learning_rate": 1.8951566951566954e-05,
|
| 192 |
+
"loss": 1.772,
|
| 193 |
+
"step": 575
|
| 194 |
+
},
|
| 195 |
+
{
|
| 196 |
+
"epoch": 3.0,
|
| 197 |
+
"eval_accuracy": 0.8529411764705882,
|
| 198 |
+
"eval_loss": 0.42055270075798035,
|
| 199 |
+
"eval_macro_f1": 0.8528889392261485,
|
| 200 |
+
"eval_runtime": 0.4458,
|
| 201 |
+
"eval_samples_per_second": 1296.579,
|
| 202 |
+
"eval_steps_per_second": 82.999,
|
| 203 |
+
"eval_weighted_f1": 0.8529198876416628,
|
| 204 |
+
"step": 585
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.0770218228498076,
|
| 208 |
+
"grad_norm": 49.937992095947266,
|
| 209 |
+
"learning_rate": 1.880911680911681e-05,
|
| 210 |
+
"loss": 1.2775,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.2053915275994864,
|
| 215 |
+
"grad_norm": 37.18613815307617,
|
| 216 |
+
"learning_rate": 1.866666666666667e-05,
|
| 217 |
+
"loss": 0.9974,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.3337612323491657,
|
| 222 |
+
"grad_norm": 37.34135818481445,
|
| 223 |
+
"learning_rate": 1.8524216524216526e-05,
|
| 224 |
+
"loss": 1.0888,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.462130937098845,
|
| 229 |
+
"grad_norm": 59.23209762573242,
|
| 230 |
+
"learning_rate": 1.8381766381766383e-05,
|
| 231 |
+
"loss": 1.0523,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.5905006418485237,
|
| 236 |
+
"grad_norm": 33.81317901611328,
|
| 237 |
+
"learning_rate": 1.823931623931624e-05,
|
| 238 |
+
"loss": 1.0592,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.718870346598203,
|
| 243 |
+
"grad_norm": 32.36256408691406,
|
| 244 |
+
"learning_rate": 1.8096866096866098e-05,
|
| 245 |
+
"loss": 1.0479,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 3.8472400513478817,
|
| 250 |
+
"grad_norm": 20.547197341918945,
|
| 251 |
+
"learning_rate": 1.7954415954415955e-05,
|
| 252 |
+
"loss": 0.9133,
|
| 253 |
+
"step": 750
|
| 254 |
+
},
|
| 255 |
+
{
|
| 256 |
+
"epoch": 3.975609756097561,
|
| 257 |
+
"grad_norm": 43.12257766723633,
|
| 258 |
+
"learning_rate": 1.7811965811965813e-05,
|
| 259 |
+
"loss": 1.0289,
|
| 260 |
+
"step": 775
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 4.0,
|
| 264 |
+
"eval_accuracy": 0.8356401384083045,
|
| 265 |
+
"eval_loss": 0.5289112329483032,
|
| 266 |
+
"eval_macro_f1": 0.8344789536399748,
|
| 267 |
+
"eval_runtime": 0.4057,
|
| 268 |
+
"eval_samples_per_second": 1424.581,
|
| 269 |
+
"eval_steps_per_second": 91.193,
|
| 270 |
+
"eval_weighted_f1": 0.834630388996351,
|
| 271 |
+
"step": 780
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.102695763799743,
|
| 275 |
+
"grad_norm": 13.654150009155273,
|
| 276 |
+
"learning_rate": 1.766951566951567e-05,
|
| 277 |
+
"loss": 0.8242,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.2310654685494224,
|
| 282 |
+
"grad_norm": 137.63372802734375,
|
| 283 |
+
"learning_rate": 1.7527065527065528e-05,
|
| 284 |
+
"loss": 0.7267,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.359435173299102,
|
| 289 |
+
"grad_norm": 15.335307121276855,
|
| 290 |
+
"learning_rate": 1.7384615384615385e-05,
|
| 291 |
+
"loss": 0.4614,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.487804878048781,
|
| 296 |
+
"grad_norm": 95.56256866455078,
|
| 297 |
+
"learning_rate": 1.7242165242165242e-05,
|
| 298 |
+
"loss": 0.4938,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.616174582798459,
|
| 303 |
+
"grad_norm": 21.486291885375977,
|
| 304 |
+
"learning_rate": 1.7099715099715103e-05,
|
| 305 |
+
"loss": 0.7232,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 4.7445442875481385,
|
| 310 |
+
"grad_norm": 38.590484619140625,
|
| 311 |
+
"learning_rate": 1.6957264957264957e-05,
|
| 312 |
+
"loss": 0.73,
|
| 313 |
+
"step": 925
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 4.872913992297818,
|
| 317 |
+
"grad_norm": 69.3209228515625,
|
| 318 |
+
"learning_rate": 1.6814814814814818e-05,
|
| 319 |
+
"loss": 0.7072,
|
| 320 |
+
"step": 950
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 5.0,
|
| 324 |
+
"grad_norm": 45.240501403808594,
|
| 325 |
+
"learning_rate": 1.6672364672364675e-05,
|
| 326 |
+
"loss": 0.8244,
|
| 327 |
+
"step": 975
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 5.0,
|
| 331 |
+
"eval_accuracy": 0.8529411764705882,
|
| 332 |
+
"eval_loss": 0.5391636490821838,
|
| 333 |
+
"eval_macro_f1": 0.8530635352428276,
|
| 334 |
+
"eval_runtime": 0.4554,
|
| 335 |
+
"eval_samples_per_second": 1269.324,
|
| 336 |
+
"eval_steps_per_second": 81.254,
|
| 337 |
+
"eval_weighted_f1": 0.8534522026712293,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.128369704749679,
|
| 342 |
+
"grad_norm": 117.1485366821289,
|
| 343 |
+
"learning_rate": 1.652991452991453e-05,
|
| 344 |
+
"loss": 0.5359,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.2567394094993585,
|
| 349 |
+
"grad_norm": 1.4075114727020264,
|
| 350 |
+
"learning_rate": 1.638746438746439e-05,
|
| 351 |
+
"loss": 0.4168,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.385109114249037,
|
| 356 |
+
"grad_norm": 70.00537872314453,
|
| 357 |
+
"learning_rate": 1.6245014245014247e-05,
|
| 358 |
+
"loss": 0.6482,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.513478818998716,
|
| 363 |
+
"grad_norm": 0.5642402768135071,
|
| 364 |
+
"learning_rate": 1.6102564102564105e-05,
|
| 365 |
+
"loss": 0.5052,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 5.641848523748395,
|
| 370 |
+
"grad_norm": 32.7194938659668,
|
| 371 |
+
"learning_rate": 1.5960113960113962e-05,
|
| 372 |
+
"loss": 0.6093,
|
| 373 |
+
"step": 1100
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 5.7702182284980745,
|
| 377 |
+
"grad_norm": 57.54438400268555,
|
| 378 |
+
"learning_rate": 1.581766381766382e-05,
|
| 379 |
+
"loss": 0.5007,
|
| 380 |
+
"step": 1125
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 5.898587933247754,
|
| 384 |
+
"grad_norm": 70.30224609375,
|
| 385 |
+
"learning_rate": 1.5675213675213677e-05,
|
| 386 |
+
"loss": 0.3988,
|
| 387 |
+
"step": 1150
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 6.0,
|
| 391 |
+
"eval_accuracy": 0.856401384083045,
|
| 392 |
+
"eval_loss": 0.6412298679351807,
|
| 393 |
+
"eval_macro_f1": 0.8557374806643535,
|
| 394 |
+
"eval_runtime": 0.4399,
|
| 395 |
+
"eval_samples_per_second": 1313.994,
|
| 396 |
+
"eval_steps_per_second": 84.114,
|
| 397 |
+
"eval_weighted_f1": 0.8557890854307254,
|
| 398 |
+
"step": 1170
|
| 399 |
+
}
|
| 400 |
+
],
|
| 401 |
+
"logging_steps": 25,
|
| 402 |
+
"max_steps": 3900,
|
| 403 |
+
"num_input_tokens_seen": 0,
|
| 404 |
+
"num_train_epochs": 20,
|
| 405 |
+
"save_steps": 500,
|
| 406 |
+
"stateful_callbacks": {
|
| 407 |
+
"EarlyStoppingCallback": {
|
| 408 |
+
"args": {
|
| 409 |
+
"early_stopping_patience": 3,
|
| 410 |
+
"early_stopping_threshold": 0.0
|
| 411 |
+
},
|
| 412 |
+
"attributes": {
|
| 413 |
+
"early_stopping_patience_counter": 0
|
| 414 |
+
}
|
| 415 |
+
},
|
| 416 |
+
"TrainerControl": {
|
| 417 |
+
"args": {
|
| 418 |
+
"should_epoch_stop": false,
|
| 419 |
+
"should_evaluate": false,
|
| 420 |
+
"should_log": false,
|
| 421 |
+
"should_save": true,
|
| 422 |
+
"should_training_stop": false
|
| 423 |
+
},
|
| 424 |
+
"attributes": {}
|
| 425 |
+
}
|
| 426 |
+
},
|
| 427 |
+
"total_flos": 0.0,
|
| 428 |
+
"train_batch_size": 8,
|
| 429 |
+
"trial_name": null,
|
| 430 |
+
"trial_params": null
|
| 431 |
+
}
|
ood-vifactcheck-phobert-topic-domain-03-s44-20260826T111158Z-ab4112e3/trainer/checkpoint-1170/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-03-s44-20260826T111158Z-ab4112e3/trainer/checkpoint-1755/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-topic-domain-03-s44-20260826T111158Z-ab4112e3/trainer/checkpoint-1755/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-03-s44-20260826T111158Z-ab4112e3/trainer/checkpoint-1755/special_tokens_map.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": "<mask>",
|
| 6 |
+
"pad_token": "<pad>",
|
| 7 |
+
"sep_token": "</s>",
|
| 8 |
+
"unk_token": "<unk>"
|
| 9 |
+
}
|
ood-vifactcheck-phobert-topic-domain-03-s44-20260826T111158Z-ab4112e3/trainer/checkpoint-1755/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"64000": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 256,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "PhobertTokenizer",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-phobert-topic-domain-03-s44-20260826T111158Z-ab4112e3/trainer/checkpoint-1755/trainer_state.json
ADDED
|
@@ -0,0 +1,632 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1170,
|
| 3 |
+
"best_metric": 0.8557374806643535,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-topic-domain-03-s44-20260826T111158Z-ab4112e3/trainer/checkpoint-1170",
|
| 5 |
+
"epoch": 9.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1755,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.12836970474967907,
|
| 14 |
+
"grad_norm": 7.355264186859131,
|
| 15 |
+
"learning_rate": 1.230769230769231e-06,
|
| 16 |
+
"loss": 4.42,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.25673940949935814,
|
| 21 |
+
"grad_norm": 7.894757270812988,
|
| 22 |
+
"learning_rate": 2.512820512820513e-06,
|
| 23 |
+
"loss": 4.3766,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.3851091142490372,
|
| 28 |
+
"grad_norm": 7.210986137390137,
|
| 29 |
+
"learning_rate": 3.794871794871795e-06,
|
| 30 |
+
"loss": 4.3868,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5134788189987163,
|
| 35 |
+
"grad_norm": 9.210958480834961,
|
| 36 |
+
"learning_rate": 5.076923076923077e-06,
|
| 37 |
+
"loss": 4.3375,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6418485237483954,
|
| 42 |
+
"grad_norm": 10.58199405670166,
|
| 43 |
+
"learning_rate": 6.358974358974359e-06,
|
| 44 |
+
"loss": 4.2921,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.7702182284980744,
|
| 49 |
+
"grad_norm": 15.043291091918945,
|
| 50 |
+
"learning_rate": 7.641025641025641e-06,
|
| 51 |
+
"loss": 4.1014,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.8985879332477535,
|
| 56 |
+
"grad_norm": 21.792621612548828,
|
| 57 |
+
"learning_rate": 8.923076923076925e-06,
|
| 58 |
+
"loss": 4.0205,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.6366782006920415,
|
| 64 |
+
"eval_loss": 0.8276789784431458,
|
| 65 |
+
"eval_macro_f1": 0.6315092173660761,
|
| 66 |
+
"eval_runtime": 0.4133,
|
| 67 |
+
"eval_samples_per_second": 1398.41,
|
| 68 |
+
"eval_steps_per_second": 89.518,
|
| 69 |
+
"eval_weighted_f1": 0.6345556061280964,
|
| 70 |
+
"step": 195
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0256739409499358,
|
| 74 |
+
"grad_norm": 45.431007385253906,
|
| 75 |
+
"learning_rate": 1.0205128205128205e-05,
|
| 76 |
+
"loss": 3.6374,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.1540436456996148,
|
| 81 |
+
"grad_norm": 38.621925354003906,
|
| 82 |
+
"learning_rate": 1.1487179487179487e-05,
|
| 83 |
+
"loss": 3.2732,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.282413350449294,
|
| 88 |
+
"grad_norm": 41.117637634277344,
|
| 89 |
+
"learning_rate": 1.2769230769230769e-05,
|
| 90 |
+
"loss": 3.1589,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.410783055198973,
|
| 95 |
+
"grad_norm": 44.818180084228516,
|
| 96 |
+
"learning_rate": 1.405128205128205e-05,
|
| 97 |
+
"loss": 2.8765,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.539152759948652,
|
| 102 |
+
"grad_norm": 46.80485916137695,
|
| 103 |
+
"learning_rate": 1.5333333333333334e-05,
|
| 104 |
+
"loss": 2.6511,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.6675224646983313,
|
| 109 |
+
"grad_norm": 33.0837287902832,
|
| 110 |
+
"learning_rate": 1.6615384615384618e-05,
|
| 111 |
+
"loss": 2.4328,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.7958921694480101,
|
| 116 |
+
"grad_norm": 34.05903244018555,
|
| 117 |
+
"learning_rate": 1.78974358974359e-05,
|
| 118 |
+
"loss": 2.1861,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 1.9242618741976893,
|
| 123 |
+
"grad_norm": 33.80216598510742,
|
| 124 |
+
"learning_rate": 1.9179487179487182e-05,
|
| 125 |
+
"loss": 2.1043,
|
| 126 |
+
"step": 375
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"epoch": 2.0,
|
| 130 |
+
"eval_accuracy": 0.7958477508650519,
|
| 131 |
+
"eval_loss": 0.5350261330604553,
|
| 132 |
+
"eval_macro_f1": 0.786443465956267,
|
| 133 |
+
"eval_runtime": 0.4296,
|
| 134 |
+
"eval_samples_per_second": 1345.294,
|
| 135 |
+
"eval_steps_per_second": 86.117,
|
| 136 |
+
"eval_weighted_f1": 0.7878405892780137,
|
| 137 |
+
"step": 390
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.0513478818998716,
|
| 141 |
+
"grad_norm": 49.66297912597656,
|
| 142 |
+
"learning_rate": 1.994871794871795e-05,
|
| 143 |
+
"loss": 2.0522,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.179717586649551,
|
| 148 |
+
"grad_norm": 56.68103790283203,
|
| 149 |
+
"learning_rate": 1.9806267806267806e-05,
|
| 150 |
+
"loss": 1.8106,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.3080872913992296,
|
| 155 |
+
"grad_norm": 54.557979583740234,
|
| 156 |
+
"learning_rate": 1.9663817663817667e-05,
|
| 157 |
+
"loss": 1.7066,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.436456996148909,
|
| 162 |
+
"grad_norm": 36.9579963684082,
|
| 163 |
+
"learning_rate": 1.952136752136752e-05,
|
| 164 |
+
"loss": 1.6575,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.564826700898588,
|
| 169 |
+
"grad_norm": 38.816524505615234,
|
| 170 |
+
"learning_rate": 1.9378917378917382e-05,
|
| 171 |
+
"loss": 1.5141,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.693196405648267,
|
| 176 |
+
"grad_norm": 18.72123146057129,
|
| 177 |
+
"learning_rate": 1.923646723646724e-05,
|
| 178 |
+
"loss": 1.3195,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 2.821566110397946,
|
| 183 |
+
"grad_norm": 31.898258209228516,
|
| 184 |
+
"learning_rate": 1.9094017094017097e-05,
|
| 185 |
+
"loss": 1.5703,
|
| 186 |
+
"step": 550
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 2.949935815147625,
|
| 190 |
+
"grad_norm": 33.38069534301758,
|
| 191 |
+
"learning_rate": 1.8951566951566954e-05,
|
| 192 |
+
"loss": 1.772,
|
| 193 |
+
"step": 575
|
| 194 |
+
},
|
| 195 |
+
{
|
| 196 |
+
"epoch": 3.0,
|
| 197 |
+
"eval_accuracy": 0.8529411764705882,
|
| 198 |
+
"eval_loss": 0.42055270075798035,
|
| 199 |
+
"eval_macro_f1": 0.8528889392261485,
|
| 200 |
+
"eval_runtime": 0.4458,
|
| 201 |
+
"eval_samples_per_second": 1296.579,
|
| 202 |
+
"eval_steps_per_second": 82.999,
|
| 203 |
+
"eval_weighted_f1": 0.8529198876416628,
|
| 204 |
+
"step": 585
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.0770218228498076,
|
| 208 |
+
"grad_norm": 49.937992095947266,
|
| 209 |
+
"learning_rate": 1.880911680911681e-05,
|
| 210 |
+
"loss": 1.2775,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.2053915275994864,
|
| 215 |
+
"grad_norm": 37.18613815307617,
|
| 216 |
+
"learning_rate": 1.866666666666667e-05,
|
| 217 |
+
"loss": 0.9974,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.3337612323491657,
|
| 222 |
+
"grad_norm": 37.34135818481445,
|
| 223 |
+
"learning_rate": 1.8524216524216526e-05,
|
| 224 |
+
"loss": 1.0888,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.462130937098845,
|
| 229 |
+
"grad_norm": 59.23209762573242,
|
| 230 |
+
"learning_rate": 1.8381766381766383e-05,
|
| 231 |
+
"loss": 1.0523,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.5905006418485237,
|
| 236 |
+
"grad_norm": 33.81317901611328,
|
| 237 |
+
"learning_rate": 1.823931623931624e-05,
|
| 238 |
+
"loss": 1.0592,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.718870346598203,
|
| 243 |
+
"grad_norm": 32.36256408691406,
|
| 244 |
+
"learning_rate": 1.8096866096866098e-05,
|
| 245 |
+
"loss": 1.0479,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 3.8472400513478817,
|
| 250 |
+
"grad_norm": 20.547197341918945,
|
| 251 |
+
"learning_rate": 1.7954415954415955e-05,
|
| 252 |
+
"loss": 0.9133,
|
| 253 |
+
"step": 750
|
| 254 |
+
},
|
| 255 |
+
{
|
| 256 |
+
"epoch": 3.975609756097561,
|
| 257 |
+
"grad_norm": 43.12257766723633,
|
| 258 |
+
"learning_rate": 1.7811965811965813e-05,
|
| 259 |
+
"loss": 1.0289,
|
| 260 |
+
"step": 775
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 4.0,
|
| 264 |
+
"eval_accuracy": 0.8356401384083045,
|
| 265 |
+
"eval_loss": 0.5289112329483032,
|
| 266 |
+
"eval_macro_f1": 0.8344789536399748,
|
| 267 |
+
"eval_runtime": 0.4057,
|
| 268 |
+
"eval_samples_per_second": 1424.581,
|
| 269 |
+
"eval_steps_per_second": 91.193,
|
| 270 |
+
"eval_weighted_f1": 0.834630388996351,
|
| 271 |
+
"step": 780
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.102695763799743,
|
| 275 |
+
"grad_norm": 13.654150009155273,
|
| 276 |
+
"learning_rate": 1.766951566951567e-05,
|
| 277 |
+
"loss": 0.8242,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.2310654685494224,
|
| 282 |
+
"grad_norm": 137.63372802734375,
|
| 283 |
+
"learning_rate": 1.7527065527065528e-05,
|
| 284 |
+
"loss": 0.7267,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.359435173299102,
|
| 289 |
+
"grad_norm": 15.335307121276855,
|
| 290 |
+
"learning_rate": 1.7384615384615385e-05,
|
| 291 |
+
"loss": 0.4614,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.487804878048781,
|
| 296 |
+
"grad_norm": 95.56256866455078,
|
| 297 |
+
"learning_rate": 1.7242165242165242e-05,
|
| 298 |
+
"loss": 0.4938,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.616174582798459,
|
| 303 |
+
"grad_norm": 21.486291885375977,
|
| 304 |
+
"learning_rate": 1.7099715099715103e-05,
|
| 305 |
+
"loss": 0.7232,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 4.7445442875481385,
|
| 310 |
+
"grad_norm": 38.590484619140625,
|
| 311 |
+
"learning_rate": 1.6957264957264957e-05,
|
| 312 |
+
"loss": 0.73,
|
| 313 |
+
"step": 925
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 4.872913992297818,
|
| 317 |
+
"grad_norm": 69.3209228515625,
|
| 318 |
+
"learning_rate": 1.6814814814814818e-05,
|
| 319 |
+
"loss": 0.7072,
|
| 320 |
+
"step": 950
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 5.0,
|
| 324 |
+
"grad_norm": 45.240501403808594,
|
| 325 |
+
"learning_rate": 1.6672364672364675e-05,
|
| 326 |
+
"loss": 0.8244,
|
| 327 |
+
"step": 975
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 5.0,
|
| 331 |
+
"eval_accuracy": 0.8529411764705882,
|
| 332 |
+
"eval_loss": 0.5391636490821838,
|
| 333 |
+
"eval_macro_f1": 0.8530635352428276,
|
| 334 |
+
"eval_runtime": 0.4554,
|
| 335 |
+
"eval_samples_per_second": 1269.324,
|
| 336 |
+
"eval_steps_per_second": 81.254,
|
| 337 |
+
"eval_weighted_f1": 0.8534522026712293,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.128369704749679,
|
| 342 |
+
"grad_norm": 117.1485366821289,
|
| 343 |
+
"learning_rate": 1.652991452991453e-05,
|
| 344 |
+
"loss": 0.5359,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.2567394094993585,
|
| 349 |
+
"grad_norm": 1.4075114727020264,
|
| 350 |
+
"learning_rate": 1.638746438746439e-05,
|
| 351 |
+
"loss": 0.4168,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.385109114249037,
|
| 356 |
+
"grad_norm": 70.00537872314453,
|
| 357 |
+
"learning_rate": 1.6245014245014247e-05,
|
| 358 |
+
"loss": 0.6482,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.513478818998716,
|
| 363 |
+
"grad_norm": 0.5642402768135071,
|
| 364 |
+
"learning_rate": 1.6102564102564105e-05,
|
| 365 |
+
"loss": 0.5052,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 5.641848523748395,
|
| 370 |
+
"grad_norm": 32.7194938659668,
|
| 371 |
+
"learning_rate": 1.5960113960113962e-05,
|
| 372 |
+
"loss": 0.6093,
|
| 373 |
+
"step": 1100
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 5.7702182284980745,
|
| 377 |
+
"grad_norm": 57.54438400268555,
|
| 378 |
+
"learning_rate": 1.581766381766382e-05,
|
| 379 |
+
"loss": 0.5007,
|
| 380 |
+
"step": 1125
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 5.898587933247754,
|
| 384 |
+
"grad_norm": 70.30224609375,
|
| 385 |
+
"learning_rate": 1.5675213675213677e-05,
|
| 386 |
+
"loss": 0.3988,
|
| 387 |
+
"step": 1150
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 6.0,
|
| 391 |
+
"eval_accuracy": 0.856401384083045,
|
| 392 |
+
"eval_loss": 0.6412298679351807,
|
| 393 |
+
"eval_macro_f1": 0.8557374806643535,
|
| 394 |
+
"eval_runtime": 0.4399,
|
| 395 |
+
"eval_samples_per_second": 1313.994,
|
| 396 |
+
"eval_steps_per_second": 84.114,
|
| 397 |
+
"eval_weighted_f1": 0.8557890854307254,
|
| 398 |
+
"step": 1170
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.025673940949936,
|
| 402 |
+
"grad_norm": 144.1610107421875,
|
| 403 |
+
"learning_rate": 1.5532763532763534e-05,
|
| 404 |
+
"loss": 0.4983,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.154043645699615,
|
| 409 |
+
"grad_norm": 205.6548309326172,
|
| 410 |
+
"learning_rate": 1.539031339031339e-05,
|
| 411 |
+
"loss": 0.3568,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.282413350449294,
|
| 416 |
+
"grad_norm": 2.15409779548645,
|
| 417 |
+
"learning_rate": 1.5247863247863249e-05,
|
| 418 |
+
"loss": 0.3162,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.410783055198973,
|
| 423 |
+
"grad_norm": 84.37388610839844,
|
| 424 |
+
"learning_rate": 1.5105413105413106e-05,
|
| 425 |
+
"loss": 0.4437,
|
| 426 |
+
"step": 1250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 6.539152759948652,
|
| 430 |
+
"grad_norm": 18.134357452392578,
|
| 431 |
+
"learning_rate": 1.4962962962962964e-05,
|
| 432 |
+
"loss": 0.4622,
|
| 433 |
+
"step": 1275
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 6.667522464698331,
|
| 437 |
+
"grad_norm": 5.952671051025391,
|
| 438 |
+
"learning_rate": 1.4820512820512823e-05,
|
| 439 |
+
"loss": 0.1966,
|
| 440 |
+
"step": 1300
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 6.7958921694480106,
|
| 444 |
+
"grad_norm": 13.544970512390137,
|
| 445 |
+
"learning_rate": 1.4678062678062678e-05,
|
| 446 |
+
"loss": 0.5034,
|
| 447 |
+
"step": 1325
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 6.92426187419769,
|
| 451 |
+
"grad_norm": 229.30418395996094,
|
| 452 |
+
"learning_rate": 1.4535612535612537e-05,
|
| 453 |
+
"loss": 0.4141,
|
| 454 |
+
"step": 1350
|
| 455 |
+
},
|
| 456 |
+
{
|
| 457 |
+
"epoch": 7.0,
|
| 458 |
+
"eval_accuracy": 0.8408304498269896,
|
| 459 |
+
"eval_loss": 0.7622020244598389,
|
| 460 |
+
"eval_macro_f1": 0.8411512706304088,
|
| 461 |
+
"eval_runtime": 0.4426,
|
| 462 |
+
"eval_samples_per_second": 1306.02,
|
| 463 |
+
"eval_steps_per_second": 83.603,
|
| 464 |
+
"eval_weighted_f1": 0.8411083252116578,
|
| 465 |
+
"step": 1365
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 7.051347881899872,
|
| 469 |
+
"grad_norm": 0.2923320531845093,
|
| 470 |
+
"learning_rate": 1.4393162393162395e-05,
|
| 471 |
+
"loss": 0.3775,
|
| 472 |
+
"step": 1375
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 7.17971758664955,
|
| 476 |
+
"grad_norm": 12.381291389465332,
|
| 477 |
+
"learning_rate": 1.425071225071225e-05,
|
| 478 |
+
"loss": 0.5123,
|
| 479 |
+
"step": 1400
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 7.30808729139923,
|
| 483 |
+
"grad_norm": 0.31667593121528625,
|
| 484 |
+
"learning_rate": 1.410826210826211e-05,
|
| 485 |
+
"loss": 0.3195,
|
| 486 |
+
"step": 1425
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 7.436456996148909,
|
| 490 |
+
"grad_norm": 3.5099740028381348,
|
| 491 |
+
"learning_rate": 1.3965811965811967e-05,
|
| 492 |
+
"loss": 0.3023,
|
| 493 |
+
"step": 1450
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 7.564826700898588,
|
| 497 |
+
"grad_norm": 28.75172996520996,
|
| 498 |
+
"learning_rate": 1.3823361823361824e-05,
|
| 499 |
+
"loss": 0.2253,
|
| 500 |
+
"step": 1475
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 7.693196405648267,
|
| 504 |
+
"grad_norm": 46.74394607543945,
|
| 505 |
+
"learning_rate": 1.3680911680911682e-05,
|
| 506 |
+
"loss": 0.3522,
|
| 507 |
+
"step": 1500
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 7.821566110397946,
|
| 511 |
+
"grad_norm": 26.004684448242188,
|
| 512 |
+
"learning_rate": 1.353846153846154e-05,
|
| 513 |
+
"loss": 0.2312,
|
| 514 |
+
"step": 1525
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 7.949935815147625,
|
| 518 |
+
"grad_norm": 2.0983738899230957,
|
| 519 |
+
"learning_rate": 1.3396011396011396e-05,
|
| 520 |
+
"loss": 0.2739,
|
| 521 |
+
"step": 1550
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 8.0,
|
| 525 |
+
"eval_accuracy": 0.842560553633218,
|
| 526 |
+
"eval_loss": 0.8228955864906311,
|
| 527 |
+
"eval_macro_f1": 0.8429158700193823,
|
| 528 |
+
"eval_runtime": 0.4522,
|
| 529 |
+
"eval_samples_per_second": 1278.298,
|
| 530 |
+
"eval_steps_per_second": 81.829,
|
| 531 |
+
"eval_weighted_f1": 0.8429341956453134,
|
| 532 |
+
"step": 1560
|
| 533 |
+
},
|
| 534 |
+
{
|
| 535 |
+
"epoch": 8.077021822849808,
|
| 536 |
+
"grad_norm": 435.05645751953125,
|
| 537 |
+
"learning_rate": 1.3253561253561255e-05,
|
| 538 |
+
"loss": 0.2854,
|
| 539 |
+
"step": 1575
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 8.205391527599486,
|
| 543 |
+
"grad_norm": 0.21021690964698792,
|
| 544 |
+
"learning_rate": 1.3111111111111113e-05,
|
| 545 |
+
"loss": 0.1693,
|
| 546 |
+
"step": 1600
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 8.333761232349165,
|
| 550 |
+
"grad_norm": 0.22363586723804474,
|
| 551 |
+
"learning_rate": 1.2968660968660968e-05,
|
| 552 |
+
"loss": 0.2353,
|
| 553 |
+
"step": 1625
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 8.462130937098845,
|
| 557 |
+
"grad_norm": 0.18580549955368042,
|
| 558 |
+
"learning_rate": 1.2826210826210828e-05,
|
| 559 |
+
"loss": 0.126,
|
| 560 |
+
"step": 1650
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 8.590500641848523,
|
| 564 |
+
"grad_norm": 38.587364196777344,
|
| 565 |
+
"learning_rate": 1.2683760683760685e-05,
|
| 566 |
+
"loss": 0.227,
|
| 567 |
+
"step": 1675
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 8.718870346598203,
|
| 571 |
+
"grad_norm": 147.95643615722656,
|
| 572 |
+
"learning_rate": 1.2541310541310542e-05,
|
| 573 |
+
"loss": 0.3492,
|
| 574 |
+
"step": 1700
|
| 575 |
+
},
|
| 576 |
+
{
|
| 577 |
+
"epoch": 8.847240051347882,
|
| 578 |
+
"grad_norm": 0.1929789036512375,
|
| 579 |
+
"learning_rate": 1.23988603988604e-05,
|
| 580 |
+
"loss": 0.1601,
|
| 581 |
+
"step": 1725
|
| 582 |
+
},
|
| 583 |
+
{
|
| 584 |
+
"epoch": 8.975609756097562,
|
| 585 |
+
"grad_norm": 0.3023139536380768,
|
| 586 |
+
"learning_rate": 1.2256410256410259e-05,
|
| 587 |
+
"loss": 0.265,
|
| 588 |
+
"step": 1750
|
| 589 |
+
},
|
| 590 |
+
{
|
| 591 |
+
"epoch": 9.0,
|
| 592 |
+
"eval_accuracy": 0.8477508650519031,
|
| 593 |
+
"eval_loss": 0.8869156837463379,
|
| 594 |
+
"eval_macro_f1": 0.8470594903704449,
|
| 595 |
+
"eval_runtime": 0.4406,
|
| 596 |
+
"eval_samples_per_second": 1311.743,
|
| 597 |
+
"eval_steps_per_second": 83.97,
|
| 598 |
+
"eval_weighted_f1": 0.8469133417703978,
|
| 599 |
+
"step": 1755
|
| 600 |
+
}
|
| 601 |
+
],
|
| 602 |
+
"logging_steps": 25,
|
| 603 |
+
"max_steps": 3900,
|
| 604 |
+
"num_input_tokens_seen": 0,
|
| 605 |
+
"num_train_epochs": 20,
|
| 606 |
+
"save_steps": 500,
|
| 607 |
+
"stateful_callbacks": {
|
| 608 |
+
"EarlyStoppingCallback": {
|
| 609 |
+
"args": {
|
| 610 |
+
"early_stopping_patience": 3,
|
| 611 |
+
"early_stopping_threshold": 0.0
|
| 612 |
+
},
|
| 613 |
+
"attributes": {
|
| 614 |
+
"early_stopping_patience_counter": 3
|
| 615 |
+
}
|
| 616 |
+
},
|
| 617 |
+
"TrainerControl": {
|
| 618 |
+
"args": {
|
| 619 |
+
"should_epoch_stop": false,
|
| 620 |
+
"should_evaluate": false,
|
| 621 |
+
"should_log": false,
|
| 622 |
+
"should_save": true,
|
| 623 |
+
"should_training_stop": true
|
| 624 |
+
},
|
| 625 |
+
"attributes": {}
|
| 626 |
+
}
|
| 627 |
+
},
|
| 628 |
+
"total_flos": 0.0,
|
| 629 |
+
"train_batch_size": 8,
|
| 630 |
+
"trial_name": null,
|
| 631 |
+
"trial_params": null
|
| 632 |
+
}
|
ood-vifactcheck-phobert-topic-domain-03-s44-20260826T111158Z-ab4112e3/trainer/checkpoint-1755/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-04-s44-20260826T111516Z-ab4112e3/trainer/checkpoint-1316/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-topic-domain-04-s44-20260826T111516Z-ab4112e3/trainer/checkpoint-1316/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|