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-s43-20260814T131947Z-bdd2c308/trainer/checkpoint-1309/added_tokens.json +3 -0
- ood-vifactcheck-phobert-topic-domain-00-s43-20260814T131947Z-bdd2c308/trainer/checkpoint-1309/bpe.codes +0 -0
- ood-vifactcheck-phobert-topic-domain-00-s43-20260814T131947Z-bdd2c308/trainer/checkpoint-1309/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-topic-domain-00-s43-20260814T131947Z-bdd2c308/trainer/checkpoint-1309/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-topic-domain-00-s43-20260814T131947Z-bdd2c308/trainer/checkpoint-1309/trainer_state.json +484 -0
- ood-vifactcheck-phobert-topic-domain-00-s43-20260814T131947Z-bdd2c308/trainer/checkpoint-1309/vocab.txt +0 -0
- ood-vifactcheck-phobert-topic-domain-00-s43-20260814T131947Z-bdd2c308/trainer/checkpoint-1870/added_tokens.json +3 -0
- ood-vifactcheck-phobert-topic-domain-00-s43-20260814T131947Z-bdd2c308/trainer/checkpoint-1870/bpe.codes +0 -0
- ood-vifactcheck-phobert-topic-domain-00-s43-20260814T131947Z-bdd2c308/trainer/checkpoint-1870/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-topic-domain-00-s43-20260814T131947Z-bdd2c308/trainer/checkpoint-1870/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-topic-domain-00-s43-20260814T131947Z-bdd2c308/trainer/checkpoint-1870/trainer_state.json +671 -0
- ood-vifactcheck-phobert-topic-domain-00-s43-20260814T131947Z-bdd2c308/trainer/checkpoint-1870/vocab.txt +0 -0
- ood-vifactcheck-phobert-topic-domain-01-s43-20260814T133200Z-bdd2c308/trainer/checkpoint-1146/added_tokens.json +3 -0
- ood-vifactcheck-phobert-topic-domain-01-s43-20260814T133200Z-bdd2c308/trainer/checkpoint-1146/bpe.codes +0 -0
- ood-vifactcheck-phobert-topic-domain-01-s43-20260814T133200Z-bdd2c308/trainer/checkpoint-1146/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-topic-domain-01-s43-20260814T133200Z-bdd2c308/trainer/checkpoint-1146/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-topic-domain-01-s43-20260814T133200Z-bdd2c308/trainer/checkpoint-1146/trainer_state.json +424 -0
- ood-vifactcheck-phobert-topic-domain-01-s43-20260814T133200Z-bdd2c308/trainer/checkpoint-1146/vocab.txt +0 -0
- ood-vifactcheck-phobert-topic-domain-01-s43-20260814T133200Z-bdd2c308/trainer/checkpoint-1719/added_tokens.json +3 -0
- ood-vifactcheck-phobert-topic-domain-01-s43-20260814T133200Z-bdd2c308/trainer/checkpoint-1719/bpe.codes +0 -0
- ood-vifactcheck-phobert-topic-domain-01-s43-20260814T133200Z-bdd2c308/trainer/checkpoint-1719/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-topic-domain-01-s43-20260814T133200Z-bdd2c308/trainer/checkpoint-1719/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-topic-domain-01-s43-20260814T133200Z-bdd2c308/trainer/checkpoint-1719/trainer_state.json +618 -0
- ood-vifactcheck-phobert-topic-domain-01-s43-20260814T133200Z-bdd2c308/trainer/checkpoint-1719/vocab.txt +0 -0
- ood-vifactcheck-phobert-topic-domain-02-s43-20260814T134559Z-bdd2c308/trainer/checkpoint-1092/added_tokens.json +3 -0
- ood-vifactcheck-phobert-topic-domain-02-s43-20260814T134559Z-bdd2c308/trainer/checkpoint-1092/bpe.codes +0 -0
- ood-vifactcheck-phobert-topic-domain-02-s43-20260814T134559Z-bdd2c308/trainer/checkpoint-1092/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-topic-domain-02-s43-20260814T134559Z-bdd2c308/trainer/checkpoint-1092/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-topic-domain-02-s43-20260814T134559Z-bdd2c308/trainer/checkpoint-1092/trainer_state.json +410 -0
- ood-vifactcheck-phobert-topic-domain-02-s43-20260814T134559Z-bdd2c308/trainer/checkpoint-1092/vocab.txt +0 -0
- ood-vifactcheck-phobert-topic-domain-02-s43-20260814T134559Z-bdd2c308/trainer/checkpoint-1638/added_tokens.json +3 -0
- ood-vifactcheck-phobert-topic-domain-02-s43-20260814T134559Z-bdd2c308/trainer/checkpoint-1638/bpe.codes +0 -0
- ood-vifactcheck-phobert-topic-domain-02-s43-20260814T134559Z-bdd2c308/trainer/checkpoint-1638/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-topic-domain-02-s43-20260814T134559Z-bdd2c308/trainer/checkpoint-1638/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-topic-domain-02-s43-20260814T134559Z-bdd2c308/trainer/checkpoint-1638/trainer_state.json +597 -0
- ood-vifactcheck-phobert-topic-domain-02-s43-20260814T134559Z-bdd2c308/trainer/checkpoint-1638/vocab.txt +0 -0
- ood-vifactcheck-phobert-topic-domain-03-s43-20260814T135955Z-bdd2c308/trainer/checkpoint-1755/added_tokens.json +3 -0
- ood-vifactcheck-phobert-topic-domain-03-s43-20260814T135955Z-bdd2c308/trainer/checkpoint-1755/bpe.codes +0 -0
- ood-vifactcheck-phobert-topic-domain-03-s43-20260814T135955Z-bdd2c308/trainer/checkpoint-1755/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-topic-domain-03-s43-20260814T135955Z-bdd2c308/trainer/checkpoint-1755/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-topic-domain-03-s43-20260814T135955Z-bdd2c308/trainer/checkpoint-1755/trainer_state.json +632 -0
- ood-vifactcheck-phobert-topic-domain-03-s43-20260814T135955Z-bdd2c308/trainer/checkpoint-1755/vocab.txt +0 -0
- ood-vifactcheck-phobert-topic-domain-03-s43-20260814T135955Z-bdd2c308/trainer/checkpoint-2340/added_tokens.json +3 -0
- ood-vifactcheck-phobert-topic-domain-03-s43-20260814T135955Z-bdd2c308/trainer/checkpoint-2340/bpe.codes +0 -0
- ood-vifactcheck-phobert-topic-domain-03-s43-20260814T135955Z-bdd2c308/trainer/checkpoint-2340/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-topic-domain-03-s43-20260814T135955Z-bdd2c308/trainer/checkpoint-2340/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-topic-domain-03-s43-20260814T135955Z-bdd2c308/trainer/checkpoint-2340/trainer_state.json +826 -0
- ood-vifactcheck-phobert-topic-domain-03-s43-20260814T135955Z-bdd2c308/trainer/checkpoint-2340/vocab.txt +0 -0
- ood-vifactcheck-phobert-topic-domain-04-s43-20260814T141531Z-bdd2c308/trainer/checkpoint-1880/added_tokens.json +3 -0
- ood-vifactcheck-phobert-topic-domain-04-s43-20260814T141531Z-bdd2c308/trainer/checkpoint-1880/bpe.codes +0 -0
ood-vifactcheck-phobert-topic-domain-00-s43-20260814T131947Z-bdd2c308/trainer/checkpoint-1309/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-topic-domain-00-s43-20260814T131947Z-bdd2c308/trainer/checkpoint-1309/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-00-s43-20260814T131947Z-bdd2c308/trainer/checkpoint-1309/special_tokens_map.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": "<mask>",
|
| 6 |
+
"pad_token": "<pad>",
|
| 7 |
+
"sep_token": "</s>",
|
| 8 |
+
"unk_token": "<unk>"
|
| 9 |
+
}
|
ood-vifactcheck-phobert-topic-domain-00-s43-20260814T131947Z-bdd2c308/trainer/checkpoint-1309/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"64000": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 256,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "PhobertTokenizer",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-phobert-topic-domain-00-s43-20260814T131947Z-bdd2c308/trainer/checkpoint-1309/trainer_state.json
ADDED
|
@@ -0,0 +1,484 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1309,
|
| 3 |
+
"best_metric": 0.8649961495621286,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-topic-domain-00-s43-20260814T131947Z-bdd2c308/trainer/checkpoint-1309",
|
| 5 |
+
"epoch": 7.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1309,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.13368983957219252,
|
| 14 |
+
"grad_norm": 8.550273895263672,
|
| 15 |
+
"learning_rate": 1.2834224598930483e-06,
|
| 16 |
+
"loss": 4.458,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.26737967914438504,
|
| 21 |
+
"grad_norm": 8.35202693939209,
|
| 22 |
+
"learning_rate": 2.6203208556149735e-06,
|
| 23 |
+
"loss": 4.4168,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.40106951871657753,
|
| 28 |
+
"grad_norm": 8.180274963378906,
|
| 29 |
+
"learning_rate": 3.957219251336899e-06,
|
| 30 |
+
"loss": 4.415,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5347593582887701,
|
| 35 |
+
"grad_norm": 6.830117702484131,
|
| 36 |
+
"learning_rate": 5.294117647058824e-06,
|
| 37 |
+
"loss": 4.3738,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6684491978609626,
|
| 42 |
+
"grad_norm": 10.448054313659668,
|
| 43 |
+
"learning_rate": 6.631016042780749e-06,
|
| 44 |
+
"loss": 4.3396,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8021390374331551,
|
| 49 |
+
"grad_norm": 6.9119157791137695,
|
| 50 |
+
"learning_rate": 7.967914438502674e-06,
|
| 51 |
+
"loss": 4.2713,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9358288770053476,
|
| 56 |
+
"grad_norm": 17.67502784729004,
|
| 57 |
+
"learning_rate": 9.3048128342246e-06,
|
| 58 |
+
"loss": 4.2205,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.5192307692307693,
|
| 64 |
+
"eval_loss": 1.0020179748535156,
|
| 65 |
+
"eval_macro_f1": 0.4799346180393955,
|
| 66 |
+
"eval_runtime": 1.2266,
|
| 67 |
+
"eval_samples_per_second": 466.315,
|
| 68 |
+
"eval_steps_per_second": 29.348,
|
| 69 |
+
"eval_weighted_f1": 0.4837086195672974,
|
| 70 |
+
"step": 187
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0695187165775402,
|
| 74 |
+
"grad_norm": 10.691020965576172,
|
| 75 |
+
"learning_rate": 1.0641711229946525e-05,
|
| 76 |
+
"loss": 4.0073,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.2032085561497325,
|
| 81 |
+
"grad_norm": 28.809295654296875,
|
| 82 |
+
"learning_rate": 1.197860962566845e-05,
|
| 83 |
+
"loss": 3.6183,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.3368983957219251,
|
| 88 |
+
"grad_norm": 20.3957576751709,
|
| 89 |
+
"learning_rate": 1.3315508021390375e-05,
|
| 90 |
+
"loss": 3.2054,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.4705882352941178,
|
| 95 |
+
"grad_norm": 35.9618034362793,
|
| 96 |
+
"learning_rate": 1.4652406417112302e-05,
|
| 97 |
+
"loss": 2.8019,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.6042780748663101,
|
| 102 |
+
"grad_norm": 52.8614387512207,
|
| 103 |
+
"learning_rate": 1.5989304812834226e-05,
|
| 104 |
+
"loss": 2.6812,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7379679144385025,
|
| 109 |
+
"grad_norm": 36.53756332397461,
|
| 110 |
+
"learning_rate": 1.7326203208556153e-05,
|
| 111 |
+
"loss": 2.6444,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.8716577540106951,
|
| 116 |
+
"grad_norm": 49.83660125732422,
|
| 117 |
+
"learning_rate": 1.8663101604278075e-05,
|
| 118 |
+
"loss": 2.2564,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.7884615384615384,
|
| 124 |
+
"eval_loss": 0.5643421411514282,
|
| 125 |
+
"eval_macro_f1": 0.7900153309394997,
|
| 126 |
+
"eval_runtime": 1.2157,
|
| 127 |
+
"eval_samples_per_second": 470.522,
|
| 128 |
+
"eval_steps_per_second": 29.613,
|
| 129 |
+
"eval_weighted_f1": 0.7902700827318445,
|
| 130 |
+
"step": 374
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.0053475935828877,
|
| 134 |
+
"grad_norm": 22.349103927612305,
|
| 135 |
+
"learning_rate": 2e-05,
|
| 136 |
+
"loss": 2.2273,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.1390374331550803,
|
| 141 |
+
"grad_norm": 30.592769622802734,
|
| 142 |
+
"learning_rate": 1.9851455733808676e-05,
|
| 143 |
+
"loss": 1.69,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.2727272727272725,
|
| 148 |
+
"grad_norm": 57.95469665527344,
|
| 149 |
+
"learning_rate": 1.9702911467617353e-05,
|
| 150 |
+
"loss": 1.709,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.406417112299465,
|
| 155 |
+
"grad_norm": 25.25126838684082,
|
| 156 |
+
"learning_rate": 1.9554367201426024e-05,
|
| 157 |
+
"loss": 1.6286,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.5401069518716577,
|
| 162 |
+
"grad_norm": 22.47257423400879,
|
| 163 |
+
"learning_rate": 1.94058229352347e-05,
|
| 164 |
+
"loss": 1.5817,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.6737967914438503,
|
| 169 |
+
"grad_norm": 30.61155891418457,
|
| 170 |
+
"learning_rate": 1.9257278669043376e-05,
|
| 171 |
+
"loss": 1.4679,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.807486631016043,
|
| 176 |
+
"grad_norm": 42.0126953125,
|
| 177 |
+
"learning_rate": 1.9108734402852053e-05,
|
| 178 |
+
"loss": 1.8517,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 2.9411764705882355,
|
| 183 |
+
"grad_norm": 23.88939666748047,
|
| 184 |
+
"learning_rate": 1.8960190136660727e-05,
|
| 185 |
+
"loss": 1.76,
|
| 186 |
+
"step": 550
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 3.0,
|
| 190 |
+
"eval_accuracy": 0.8426573426573427,
|
| 191 |
+
"eval_loss": 0.409542053937912,
|
| 192 |
+
"eval_macro_f1": 0.8417380338584355,
|
| 193 |
+
"eval_runtime": 1.2187,
|
| 194 |
+
"eval_samples_per_second": 469.35,
|
| 195 |
+
"eval_steps_per_second": 29.539,
|
| 196 |
+
"eval_weighted_f1": 0.8417467513181709,
|
| 197 |
+
"step": 561
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.0748663101604277,
|
| 201 |
+
"grad_norm": 50.94145965576172,
|
| 202 |
+
"learning_rate": 1.88116458704694e-05,
|
| 203 |
+
"loss": 1.4248,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.2085561497326203,
|
| 208 |
+
"grad_norm": 25.70808219909668,
|
| 209 |
+
"learning_rate": 1.8663101604278075e-05,
|
| 210 |
+
"loss": 0.9205,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.342245989304813,
|
| 215 |
+
"grad_norm": 34.09783172607422,
|
| 216 |
+
"learning_rate": 1.8514557338086753e-05,
|
| 217 |
+
"loss": 1.0759,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.4759358288770055,
|
| 222 |
+
"grad_norm": 49.7615966796875,
|
| 223 |
+
"learning_rate": 1.8366013071895427e-05,
|
| 224 |
+
"loss": 1.3098,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.6096256684491976,
|
| 229 |
+
"grad_norm": 37.38193893432617,
|
| 230 |
+
"learning_rate": 1.82174688057041e-05,
|
| 231 |
+
"loss": 0.9969,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.7433155080213902,
|
| 236 |
+
"grad_norm": 14.047014236450195,
|
| 237 |
+
"learning_rate": 1.8068924539512775e-05,
|
| 238 |
+
"loss": 0.9892,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.877005347593583,
|
| 243 |
+
"grad_norm": 35.82085418701172,
|
| 244 |
+
"learning_rate": 1.7920380273321453e-05,
|
| 245 |
+
"loss": 1.1764,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 4.0,
|
| 250 |
+
"eval_accuracy": 0.8496503496503497,
|
| 251 |
+
"eval_loss": 0.4479869604110718,
|
| 252 |
+
"eval_macro_f1": 0.8506644603130006,
|
| 253 |
+
"eval_runtime": 1.2528,
|
| 254 |
+
"eval_samples_per_second": 456.591,
|
| 255 |
+
"eval_steps_per_second": 28.737,
|
| 256 |
+
"eval_weighted_f1": 0.8503804119050217,
|
| 257 |
+
"step": 748
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.010695187165775,
|
| 261 |
+
"grad_norm": 48.36317825317383,
|
| 262 |
+
"learning_rate": 1.7771836007130127e-05,
|
| 263 |
+
"loss": 1.1172,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.144385026737968,
|
| 268 |
+
"grad_norm": 109.0439682006836,
|
| 269 |
+
"learning_rate": 1.76232917409388e-05,
|
| 270 |
+
"loss": 0.8959,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.278074866310161,
|
| 275 |
+
"grad_norm": 34.39694595336914,
|
| 276 |
+
"learning_rate": 1.7474747474747475e-05,
|
| 277 |
+
"loss": 0.7605,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.411764705882353,
|
| 282 |
+
"grad_norm": 50.23285675048828,
|
| 283 |
+
"learning_rate": 1.7326203208556153e-05,
|
| 284 |
+
"loss": 0.6136,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.545454545454545,
|
| 289 |
+
"grad_norm": 31.331445693969727,
|
| 290 |
+
"learning_rate": 1.7177658942364827e-05,
|
| 291 |
+
"loss": 0.8215,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.6791443850267385,
|
| 296 |
+
"grad_norm": 10.790857315063477,
|
| 297 |
+
"learning_rate": 1.70291146761735e-05,
|
| 298 |
+
"loss": 0.8176,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.81283422459893,
|
| 303 |
+
"grad_norm": 74.65180969238281,
|
| 304 |
+
"learning_rate": 1.6880570409982175e-05,
|
| 305 |
+
"loss": 0.6816,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 4.946524064171123,
|
| 310 |
+
"grad_norm": 52.105438232421875,
|
| 311 |
+
"learning_rate": 1.6732026143790852e-05,
|
| 312 |
+
"loss": 0.7524,
|
| 313 |
+
"step": 925
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 5.0,
|
| 317 |
+
"eval_accuracy": 0.8583916083916084,
|
| 318 |
+
"eval_loss": 0.4474256932735443,
|
| 319 |
+
"eval_macro_f1": 0.8584367006806669,
|
| 320 |
+
"eval_runtime": 1.2162,
|
| 321 |
+
"eval_samples_per_second": 470.308,
|
| 322 |
+
"eval_steps_per_second": 29.6,
|
| 323 |
+
"eval_weighted_f1": 0.85794731020364,
|
| 324 |
+
"step": 935
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.080213903743315,
|
| 328 |
+
"grad_norm": 12.245149612426758,
|
| 329 |
+
"learning_rate": 1.6583481877599526e-05,
|
| 330 |
+
"loss": 0.7131,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.213903743315508,
|
| 335 |
+
"grad_norm": 59.7825813293457,
|
| 336 |
+
"learning_rate": 1.64349376114082e-05,
|
| 337 |
+
"loss": 0.5904,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.347593582887701,
|
| 342 |
+
"grad_norm": 49.57136535644531,
|
| 343 |
+
"learning_rate": 1.6286393345216875e-05,
|
| 344 |
+
"loss": 0.6863,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.481283422459893,
|
| 349 |
+
"grad_norm": 43.128318786621094,
|
| 350 |
+
"learning_rate": 1.6137849079025552e-05,
|
| 351 |
+
"loss": 0.5114,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.614973262032086,
|
| 356 |
+
"grad_norm": 103.40721893310547,
|
| 357 |
+
"learning_rate": 1.5989304812834226e-05,
|
| 358 |
+
"loss": 0.5661,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.748663101604278,
|
| 363 |
+
"grad_norm": 18.81574821472168,
|
| 364 |
+
"learning_rate": 1.58407605466429e-05,
|
| 365 |
+
"loss": 0.5256,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 5.882352941176471,
|
| 370 |
+
"grad_norm": 1.978880763053894,
|
| 371 |
+
"learning_rate": 1.5692216280451574e-05,
|
| 372 |
+
"loss": 0.7984,
|
| 373 |
+
"step": 1100
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 6.0,
|
| 377 |
+
"eval_accuracy": 0.8409090909090909,
|
| 378 |
+
"eval_loss": 0.5604342818260193,
|
| 379 |
+
"eval_macro_f1": 0.8410231545083371,
|
| 380 |
+
"eval_runtime": 1.2259,
|
| 381 |
+
"eval_samples_per_second": 466.583,
|
| 382 |
+
"eval_steps_per_second": 29.365,
|
| 383 |
+
"eval_weighted_f1": 0.840753267812526,
|
| 384 |
+
"step": 1122
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 6.016042780748663,
|
| 388 |
+
"grad_norm": 23.75819206237793,
|
| 389 |
+
"learning_rate": 1.5543672014260252e-05,
|
| 390 |
+
"loss": 0.6205,
|
| 391 |
+
"step": 1125
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 6.149732620320855,
|
| 395 |
+
"grad_norm": 1.4443494081497192,
|
| 396 |
+
"learning_rate": 1.5395127748068926e-05,
|
| 397 |
+
"loss": 0.4256,
|
| 398 |
+
"step": 1150
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.283422459893048,
|
| 402 |
+
"grad_norm": 9.22177505493164,
|
| 403 |
+
"learning_rate": 1.5246583481877602e-05,
|
| 404 |
+
"loss": 0.4217,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.4171122994652405,
|
| 409 |
+
"grad_norm": 160.48672485351562,
|
| 410 |
+
"learning_rate": 1.5098039215686276e-05,
|
| 411 |
+
"loss": 0.4283,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.550802139037433,
|
| 416 |
+
"grad_norm": 100.88615417480469,
|
| 417 |
+
"learning_rate": 1.4949494949494952e-05,
|
| 418 |
+
"loss": 0.4926,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.684491978609626,
|
| 423 |
+
"grad_norm": 36.497230529785156,
|
| 424 |
+
"learning_rate": 1.4800950683303626e-05,
|
| 425 |
+
"loss": 0.4978,
|
| 426 |
+
"step": 1250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 6.818181818181818,
|
| 430 |
+
"grad_norm": 0.34046390652656555,
|
| 431 |
+
"learning_rate": 1.4652406417112302e-05,
|
| 432 |
+
"loss": 0.4061,
|
| 433 |
+
"step": 1275
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 6.951871657754011,
|
| 437 |
+
"grad_norm": 4.94968843460083,
|
| 438 |
+
"learning_rate": 1.4503862150920976e-05,
|
| 439 |
+
"loss": 0.2933,
|
| 440 |
+
"step": 1300
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 7.0,
|
| 444 |
+
"eval_accuracy": 0.8653846153846154,
|
| 445 |
+
"eval_loss": 0.5979101061820984,
|
| 446 |
+
"eval_macro_f1": 0.8649961495621286,
|
| 447 |
+
"eval_runtime": 1.1963,
|
| 448 |
+
"eval_samples_per_second": 478.148,
|
| 449 |
+
"eval_steps_per_second": 30.093,
|
| 450 |
+
"eval_weighted_f1": 0.8647390579624559,
|
| 451 |
+
"step": 1309
|
| 452 |
+
}
|
| 453 |
+
],
|
| 454 |
+
"logging_steps": 25,
|
| 455 |
+
"max_steps": 3740,
|
| 456 |
+
"num_input_tokens_seen": 0,
|
| 457 |
+
"num_train_epochs": 20,
|
| 458 |
+
"save_steps": 500,
|
| 459 |
+
"stateful_callbacks": {
|
| 460 |
+
"EarlyStoppingCallback": {
|
| 461 |
+
"args": {
|
| 462 |
+
"early_stopping_patience": 3,
|
| 463 |
+
"early_stopping_threshold": 0.0
|
| 464 |
+
},
|
| 465 |
+
"attributes": {
|
| 466 |
+
"early_stopping_patience_counter": 0
|
| 467 |
+
}
|
| 468 |
+
},
|
| 469 |
+
"TrainerControl": {
|
| 470 |
+
"args": {
|
| 471 |
+
"should_epoch_stop": false,
|
| 472 |
+
"should_evaluate": false,
|
| 473 |
+
"should_log": false,
|
| 474 |
+
"should_save": true,
|
| 475 |
+
"should_training_stop": false
|
| 476 |
+
},
|
| 477 |
+
"attributes": {}
|
| 478 |
+
}
|
| 479 |
+
},
|
| 480 |
+
"total_flos": 0.0,
|
| 481 |
+
"train_batch_size": 8,
|
| 482 |
+
"trial_name": null,
|
| 483 |
+
"trial_params": null
|
| 484 |
+
}
|
ood-vifactcheck-phobert-topic-domain-00-s43-20260814T131947Z-bdd2c308/trainer/checkpoint-1309/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-00-s43-20260814T131947Z-bdd2c308/trainer/checkpoint-1870/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-topic-domain-00-s43-20260814T131947Z-bdd2c308/trainer/checkpoint-1870/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-00-s43-20260814T131947Z-bdd2c308/trainer/checkpoint-1870/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-s43-20260814T131947Z-bdd2c308/trainer/checkpoint-1870/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-s43-20260814T131947Z-bdd2c308/trainer/checkpoint-1870/trainer_state.json
ADDED
|
@@ -0,0 +1,671 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1309,
|
| 3 |
+
"best_metric": 0.8649961495621286,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-topic-domain-00-s43-20260814T131947Z-bdd2c308/trainer/checkpoint-1309",
|
| 5 |
+
"epoch": 10.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1870,
|
| 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": 8.550273895263672,
|
| 15 |
+
"learning_rate": 1.2834224598930483e-06,
|
| 16 |
+
"loss": 4.458,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.26737967914438504,
|
| 21 |
+
"grad_norm": 8.35202693939209,
|
| 22 |
+
"learning_rate": 2.6203208556149735e-06,
|
| 23 |
+
"loss": 4.4168,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.40106951871657753,
|
| 28 |
+
"grad_norm": 8.180274963378906,
|
| 29 |
+
"learning_rate": 3.957219251336899e-06,
|
| 30 |
+
"loss": 4.415,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5347593582887701,
|
| 35 |
+
"grad_norm": 6.830117702484131,
|
| 36 |
+
"learning_rate": 5.294117647058824e-06,
|
| 37 |
+
"loss": 4.3738,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6684491978609626,
|
| 42 |
+
"grad_norm": 10.448054313659668,
|
| 43 |
+
"learning_rate": 6.631016042780749e-06,
|
| 44 |
+
"loss": 4.3396,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8021390374331551,
|
| 49 |
+
"grad_norm": 6.9119157791137695,
|
| 50 |
+
"learning_rate": 7.967914438502674e-06,
|
| 51 |
+
"loss": 4.2713,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9358288770053476,
|
| 56 |
+
"grad_norm": 17.67502784729004,
|
| 57 |
+
"learning_rate": 9.3048128342246e-06,
|
| 58 |
+
"loss": 4.2205,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.5192307692307693,
|
| 64 |
+
"eval_loss": 1.0020179748535156,
|
| 65 |
+
"eval_macro_f1": 0.4799346180393955,
|
| 66 |
+
"eval_runtime": 1.2266,
|
| 67 |
+
"eval_samples_per_second": 466.315,
|
| 68 |
+
"eval_steps_per_second": 29.348,
|
| 69 |
+
"eval_weighted_f1": 0.4837086195672974,
|
| 70 |
+
"step": 187
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0695187165775402,
|
| 74 |
+
"grad_norm": 10.691020965576172,
|
| 75 |
+
"learning_rate": 1.0641711229946525e-05,
|
| 76 |
+
"loss": 4.0073,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.2032085561497325,
|
| 81 |
+
"grad_norm": 28.809295654296875,
|
| 82 |
+
"learning_rate": 1.197860962566845e-05,
|
| 83 |
+
"loss": 3.6183,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.3368983957219251,
|
| 88 |
+
"grad_norm": 20.3957576751709,
|
| 89 |
+
"learning_rate": 1.3315508021390375e-05,
|
| 90 |
+
"loss": 3.2054,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.4705882352941178,
|
| 95 |
+
"grad_norm": 35.9618034362793,
|
| 96 |
+
"learning_rate": 1.4652406417112302e-05,
|
| 97 |
+
"loss": 2.8019,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.6042780748663101,
|
| 102 |
+
"grad_norm": 52.8614387512207,
|
| 103 |
+
"learning_rate": 1.5989304812834226e-05,
|
| 104 |
+
"loss": 2.6812,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7379679144385025,
|
| 109 |
+
"grad_norm": 36.53756332397461,
|
| 110 |
+
"learning_rate": 1.7326203208556153e-05,
|
| 111 |
+
"loss": 2.6444,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.8716577540106951,
|
| 116 |
+
"grad_norm": 49.83660125732422,
|
| 117 |
+
"learning_rate": 1.8663101604278075e-05,
|
| 118 |
+
"loss": 2.2564,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.7884615384615384,
|
| 124 |
+
"eval_loss": 0.5643421411514282,
|
| 125 |
+
"eval_macro_f1": 0.7900153309394997,
|
| 126 |
+
"eval_runtime": 1.2157,
|
| 127 |
+
"eval_samples_per_second": 470.522,
|
| 128 |
+
"eval_steps_per_second": 29.613,
|
| 129 |
+
"eval_weighted_f1": 0.7902700827318445,
|
| 130 |
+
"step": 374
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.0053475935828877,
|
| 134 |
+
"grad_norm": 22.349103927612305,
|
| 135 |
+
"learning_rate": 2e-05,
|
| 136 |
+
"loss": 2.2273,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.1390374331550803,
|
| 141 |
+
"grad_norm": 30.592769622802734,
|
| 142 |
+
"learning_rate": 1.9851455733808676e-05,
|
| 143 |
+
"loss": 1.69,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.2727272727272725,
|
| 148 |
+
"grad_norm": 57.95469665527344,
|
| 149 |
+
"learning_rate": 1.9702911467617353e-05,
|
| 150 |
+
"loss": 1.709,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.406417112299465,
|
| 155 |
+
"grad_norm": 25.25126838684082,
|
| 156 |
+
"learning_rate": 1.9554367201426024e-05,
|
| 157 |
+
"loss": 1.6286,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.5401069518716577,
|
| 162 |
+
"grad_norm": 22.47257423400879,
|
| 163 |
+
"learning_rate": 1.94058229352347e-05,
|
| 164 |
+
"loss": 1.5817,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.6737967914438503,
|
| 169 |
+
"grad_norm": 30.61155891418457,
|
| 170 |
+
"learning_rate": 1.9257278669043376e-05,
|
| 171 |
+
"loss": 1.4679,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.807486631016043,
|
| 176 |
+
"grad_norm": 42.0126953125,
|
| 177 |
+
"learning_rate": 1.9108734402852053e-05,
|
| 178 |
+
"loss": 1.8517,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 2.9411764705882355,
|
| 183 |
+
"grad_norm": 23.88939666748047,
|
| 184 |
+
"learning_rate": 1.8960190136660727e-05,
|
| 185 |
+
"loss": 1.76,
|
| 186 |
+
"step": 550
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 3.0,
|
| 190 |
+
"eval_accuracy": 0.8426573426573427,
|
| 191 |
+
"eval_loss": 0.409542053937912,
|
| 192 |
+
"eval_macro_f1": 0.8417380338584355,
|
| 193 |
+
"eval_runtime": 1.2187,
|
| 194 |
+
"eval_samples_per_second": 469.35,
|
| 195 |
+
"eval_steps_per_second": 29.539,
|
| 196 |
+
"eval_weighted_f1": 0.8417467513181709,
|
| 197 |
+
"step": 561
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.0748663101604277,
|
| 201 |
+
"grad_norm": 50.94145965576172,
|
| 202 |
+
"learning_rate": 1.88116458704694e-05,
|
| 203 |
+
"loss": 1.4248,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.2085561497326203,
|
| 208 |
+
"grad_norm": 25.70808219909668,
|
| 209 |
+
"learning_rate": 1.8663101604278075e-05,
|
| 210 |
+
"loss": 0.9205,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.342245989304813,
|
| 215 |
+
"grad_norm": 34.09783172607422,
|
| 216 |
+
"learning_rate": 1.8514557338086753e-05,
|
| 217 |
+
"loss": 1.0759,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.4759358288770055,
|
| 222 |
+
"grad_norm": 49.7615966796875,
|
| 223 |
+
"learning_rate": 1.8366013071895427e-05,
|
| 224 |
+
"loss": 1.3098,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.6096256684491976,
|
| 229 |
+
"grad_norm": 37.38193893432617,
|
| 230 |
+
"learning_rate": 1.82174688057041e-05,
|
| 231 |
+
"loss": 0.9969,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.7433155080213902,
|
| 236 |
+
"grad_norm": 14.047014236450195,
|
| 237 |
+
"learning_rate": 1.8068924539512775e-05,
|
| 238 |
+
"loss": 0.9892,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.877005347593583,
|
| 243 |
+
"grad_norm": 35.82085418701172,
|
| 244 |
+
"learning_rate": 1.7920380273321453e-05,
|
| 245 |
+
"loss": 1.1764,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 4.0,
|
| 250 |
+
"eval_accuracy": 0.8496503496503497,
|
| 251 |
+
"eval_loss": 0.4479869604110718,
|
| 252 |
+
"eval_macro_f1": 0.8506644603130006,
|
| 253 |
+
"eval_runtime": 1.2528,
|
| 254 |
+
"eval_samples_per_second": 456.591,
|
| 255 |
+
"eval_steps_per_second": 28.737,
|
| 256 |
+
"eval_weighted_f1": 0.8503804119050217,
|
| 257 |
+
"step": 748
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.010695187165775,
|
| 261 |
+
"grad_norm": 48.36317825317383,
|
| 262 |
+
"learning_rate": 1.7771836007130127e-05,
|
| 263 |
+
"loss": 1.1172,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.144385026737968,
|
| 268 |
+
"grad_norm": 109.0439682006836,
|
| 269 |
+
"learning_rate": 1.76232917409388e-05,
|
| 270 |
+
"loss": 0.8959,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.278074866310161,
|
| 275 |
+
"grad_norm": 34.39694595336914,
|
| 276 |
+
"learning_rate": 1.7474747474747475e-05,
|
| 277 |
+
"loss": 0.7605,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.411764705882353,
|
| 282 |
+
"grad_norm": 50.23285675048828,
|
| 283 |
+
"learning_rate": 1.7326203208556153e-05,
|
| 284 |
+
"loss": 0.6136,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.545454545454545,
|
| 289 |
+
"grad_norm": 31.331445693969727,
|
| 290 |
+
"learning_rate": 1.7177658942364827e-05,
|
| 291 |
+
"loss": 0.8215,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.6791443850267385,
|
| 296 |
+
"grad_norm": 10.790857315063477,
|
| 297 |
+
"learning_rate": 1.70291146761735e-05,
|
| 298 |
+
"loss": 0.8176,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.81283422459893,
|
| 303 |
+
"grad_norm": 74.65180969238281,
|
| 304 |
+
"learning_rate": 1.6880570409982175e-05,
|
| 305 |
+
"loss": 0.6816,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 4.946524064171123,
|
| 310 |
+
"grad_norm": 52.105438232421875,
|
| 311 |
+
"learning_rate": 1.6732026143790852e-05,
|
| 312 |
+
"loss": 0.7524,
|
| 313 |
+
"step": 925
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 5.0,
|
| 317 |
+
"eval_accuracy": 0.8583916083916084,
|
| 318 |
+
"eval_loss": 0.4474256932735443,
|
| 319 |
+
"eval_macro_f1": 0.8584367006806669,
|
| 320 |
+
"eval_runtime": 1.2162,
|
| 321 |
+
"eval_samples_per_second": 470.308,
|
| 322 |
+
"eval_steps_per_second": 29.6,
|
| 323 |
+
"eval_weighted_f1": 0.85794731020364,
|
| 324 |
+
"step": 935
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.080213903743315,
|
| 328 |
+
"grad_norm": 12.245149612426758,
|
| 329 |
+
"learning_rate": 1.6583481877599526e-05,
|
| 330 |
+
"loss": 0.7131,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.213903743315508,
|
| 335 |
+
"grad_norm": 59.7825813293457,
|
| 336 |
+
"learning_rate": 1.64349376114082e-05,
|
| 337 |
+
"loss": 0.5904,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.347593582887701,
|
| 342 |
+
"grad_norm": 49.57136535644531,
|
| 343 |
+
"learning_rate": 1.6286393345216875e-05,
|
| 344 |
+
"loss": 0.6863,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.481283422459893,
|
| 349 |
+
"grad_norm": 43.128318786621094,
|
| 350 |
+
"learning_rate": 1.6137849079025552e-05,
|
| 351 |
+
"loss": 0.5114,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.614973262032086,
|
| 356 |
+
"grad_norm": 103.40721893310547,
|
| 357 |
+
"learning_rate": 1.5989304812834226e-05,
|
| 358 |
+
"loss": 0.5661,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.748663101604278,
|
| 363 |
+
"grad_norm": 18.81574821472168,
|
| 364 |
+
"learning_rate": 1.58407605466429e-05,
|
| 365 |
+
"loss": 0.5256,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 5.882352941176471,
|
| 370 |
+
"grad_norm": 1.978880763053894,
|
| 371 |
+
"learning_rate": 1.5692216280451574e-05,
|
| 372 |
+
"loss": 0.7984,
|
| 373 |
+
"step": 1100
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 6.0,
|
| 377 |
+
"eval_accuracy": 0.8409090909090909,
|
| 378 |
+
"eval_loss": 0.5604342818260193,
|
| 379 |
+
"eval_macro_f1": 0.8410231545083371,
|
| 380 |
+
"eval_runtime": 1.2259,
|
| 381 |
+
"eval_samples_per_second": 466.583,
|
| 382 |
+
"eval_steps_per_second": 29.365,
|
| 383 |
+
"eval_weighted_f1": 0.840753267812526,
|
| 384 |
+
"step": 1122
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 6.016042780748663,
|
| 388 |
+
"grad_norm": 23.75819206237793,
|
| 389 |
+
"learning_rate": 1.5543672014260252e-05,
|
| 390 |
+
"loss": 0.6205,
|
| 391 |
+
"step": 1125
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 6.149732620320855,
|
| 395 |
+
"grad_norm": 1.4443494081497192,
|
| 396 |
+
"learning_rate": 1.5395127748068926e-05,
|
| 397 |
+
"loss": 0.4256,
|
| 398 |
+
"step": 1150
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.283422459893048,
|
| 402 |
+
"grad_norm": 9.22177505493164,
|
| 403 |
+
"learning_rate": 1.5246583481877602e-05,
|
| 404 |
+
"loss": 0.4217,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.4171122994652405,
|
| 409 |
+
"grad_norm": 160.48672485351562,
|
| 410 |
+
"learning_rate": 1.5098039215686276e-05,
|
| 411 |
+
"loss": 0.4283,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.550802139037433,
|
| 416 |
+
"grad_norm": 100.88615417480469,
|
| 417 |
+
"learning_rate": 1.4949494949494952e-05,
|
| 418 |
+
"loss": 0.4926,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.684491978609626,
|
| 423 |
+
"grad_norm": 36.497230529785156,
|
| 424 |
+
"learning_rate": 1.4800950683303626e-05,
|
| 425 |
+
"loss": 0.4978,
|
| 426 |
+
"step": 1250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 6.818181818181818,
|
| 430 |
+
"grad_norm": 0.34046390652656555,
|
| 431 |
+
"learning_rate": 1.4652406417112302e-05,
|
| 432 |
+
"loss": 0.4061,
|
| 433 |
+
"step": 1275
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 6.951871657754011,
|
| 437 |
+
"grad_norm": 4.94968843460083,
|
| 438 |
+
"learning_rate": 1.4503862150920976e-05,
|
| 439 |
+
"loss": 0.2933,
|
| 440 |
+
"step": 1300
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 7.0,
|
| 444 |
+
"eval_accuracy": 0.8653846153846154,
|
| 445 |
+
"eval_loss": 0.5979101061820984,
|
| 446 |
+
"eval_macro_f1": 0.8649961495621286,
|
| 447 |
+
"eval_runtime": 1.1963,
|
| 448 |
+
"eval_samples_per_second": 478.148,
|
| 449 |
+
"eval_steps_per_second": 30.093,
|
| 450 |
+
"eval_weighted_f1": 0.8647390579624559,
|
| 451 |
+
"step": 1309
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 7.0855614973262036,
|
| 455 |
+
"grad_norm": 83.05696868896484,
|
| 456 |
+
"learning_rate": 1.4355317884729652e-05,
|
| 457 |
+
"loss": 0.4728,
|
| 458 |
+
"step": 1325
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 7.219251336898395,
|
| 462 |
+
"grad_norm": 14.914695739746094,
|
| 463 |
+
"learning_rate": 1.4206773618538326e-05,
|
| 464 |
+
"loss": 0.4154,
|
| 465 |
+
"step": 1350
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 7.352941176470588,
|
| 469 |
+
"grad_norm": 0.9840183258056641,
|
| 470 |
+
"learning_rate": 1.4058229352347002e-05,
|
| 471 |
+
"loss": 0.291,
|
| 472 |
+
"step": 1375
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 7.4866310160427805,
|
| 476 |
+
"grad_norm": 209.08303833007812,
|
| 477 |
+
"learning_rate": 1.3909685086155676e-05,
|
| 478 |
+
"loss": 0.2542,
|
| 479 |
+
"step": 1400
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 7.620320855614973,
|
| 483 |
+
"grad_norm": 78.11043548583984,
|
| 484 |
+
"learning_rate": 1.3761140819964351e-05,
|
| 485 |
+
"loss": 0.3005,
|
| 486 |
+
"step": 1425
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 7.754010695187166,
|
| 490 |
+
"grad_norm": 0.5458040237426758,
|
| 491 |
+
"learning_rate": 1.3612596553773026e-05,
|
| 492 |
+
"loss": 0.3645,
|
| 493 |
+
"step": 1450
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 7.887700534759358,
|
| 497 |
+
"grad_norm": 0.3649214804172516,
|
| 498 |
+
"learning_rate": 1.3464052287581701e-05,
|
| 499 |
+
"loss": 0.399,
|
| 500 |
+
"step": 1475
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 8.0,
|
| 504 |
+
"eval_accuracy": 0.8479020979020979,
|
| 505 |
+
"eval_loss": 0.7353163957595825,
|
| 506 |
+
"eval_macro_f1": 0.8481413569048634,
|
| 507 |
+
"eval_runtime": 1.2248,
|
| 508 |
+
"eval_samples_per_second": 467.028,
|
| 509 |
+
"eval_steps_per_second": 29.393,
|
| 510 |
+
"eval_weighted_f1": 0.8475335295021763,
|
| 511 |
+
"step": 1496
|
| 512 |
+
},
|
| 513 |
+
{
|
| 514 |
+
"epoch": 8.02139037433155,
|
| 515 |
+
"grad_norm": 0.2816140353679657,
|
| 516 |
+
"learning_rate": 1.3315508021390375e-05,
|
| 517 |
+
"loss": 0.4812,
|
| 518 |
+
"step": 1500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 8.155080213903743,
|
| 522 |
+
"grad_norm": 1.3347516059875488,
|
| 523 |
+
"learning_rate": 1.3166963755199051e-05,
|
| 524 |
+
"loss": 0.3616,
|
| 525 |
+
"step": 1525
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 8.288770053475936,
|
| 529 |
+
"grad_norm": 0.2449711412191391,
|
| 530 |
+
"learning_rate": 1.3018419489007725e-05,
|
| 531 |
+
"loss": 0.2338,
|
| 532 |
+
"step": 1550
|
| 533 |
+
},
|
| 534 |
+
{
|
| 535 |
+
"epoch": 8.422459893048128,
|
| 536 |
+
"grad_norm": 63.414058685302734,
|
| 537 |
+
"learning_rate": 1.2869875222816401e-05,
|
| 538 |
+
"loss": 0.1533,
|
| 539 |
+
"step": 1575
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 8.556149732620321,
|
| 543 |
+
"grad_norm": 0.23334269225597382,
|
| 544 |
+
"learning_rate": 1.2721330956625075e-05,
|
| 545 |
+
"loss": 0.3094,
|
| 546 |
+
"step": 1600
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 8.689839572192513,
|
| 550 |
+
"grad_norm": 1.5575984716415405,
|
| 551 |
+
"learning_rate": 1.257278669043375e-05,
|
| 552 |
+
"loss": 0.2063,
|
| 553 |
+
"step": 1625
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 8.823529411764707,
|
| 557 |
+
"grad_norm": 0.2828432619571686,
|
| 558 |
+
"learning_rate": 1.2424242424242425e-05,
|
| 559 |
+
"loss": 0.3134,
|
| 560 |
+
"step": 1650
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 8.957219251336898,
|
| 564 |
+
"grad_norm": 0.2469726800918579,
|
| 565 |
+
"learning_rate": 1.22756981580511e-05,
|
| 566 |
+
"loss": 0.2491,
|
| 567 |
+
"step": 1675
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 9.0,
|
| 571 |
+
"eval_accuracy": 0.8426573426573427,
|
| 572 |
+
"eval_loss": 0.8236244916915894,
|
| 573 |
+
"eval_macro_f1": 0.843112440925141,
|
| 574 |
+
"eval_runtime": 1.2088,
|
| 575 |
+
"eval_samples_per_second": 473.208,
|
| 576 |
+
"eval_steps_per_second": 29.782,
|
| 577 |
+
"eval_weighted_f1": 0.8424271137629507,
|
| 578 |
+
"step": 1683
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 9.090909090909092,
|
| 582 |
+
"grad_norm": 8.258333206176758,
|
| 583 |
+
"learning_rate": 1.2127153891859775e-05,
|
| 584 |
+
"loss": 0.2262,
|
| 585 |
+
"step": 1700
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 9.224598930481283,
|
| 589 |
+
"grad_norm": 0.4637715816497803,
|
| 590 |
+
"learning_rate": 1.197860962566845e-05,
|
| 591 |
+
"loss": 0.1847,
|
| 592 |
+
"step": 1725
|
| 593 |
+
},
|
| 594 |
+
{
|
| 595 |
+
"epoch": 9.358288770053475,
|
| 596 |
+
"grad_norm": 0.20395129919052124,
|
| 597 |
+
"learning_rate": 1.1830065359477125e-05,
|
| 598 |
+
"loss": 0.1121,
|
| 599 |
+
"step": 1750
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 9.491978609625669,
|
| 603 |
+
"grad_norm": 0.22899523377418518,
|
| 604 |
+
"learning_rate": 1.1681521093285799e-05,
|
| 605 |
+
"loss": 0.1835,
|
| 606 |
+
"step": 1775
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"epoch": 9.62566844919786,
|
| 610 |
+
"grad_norm": 0.1722222864627838,
|
| 611 |
+
"learning_rate": 1.1532976827094475e-05,
|
| 612 |
+
"loss": 0.2646,
|
| 613 |
+
"step": 1800
|
| 614 |
+
},
|
| 615 |
+
{
|
| 616 |
+
"epoch": 9.759358288770054,
|
| 617 |
+
"grad_norm": 0.1547284573316574,
|
| 618 |
+
"learning_rate": 1.1384432560903149e-05,
|
| 619 |
+
"loss": 0.2177,
|
| 620 |
+
"step": 1825
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 9.893048128342246,
|
| 624 |
+
"grad_norm": 144.07351684570312,
|
| 625 |
+
"learning_rate": 1.1235888294711825e-05,
|
| 626 |
+
"loss": 0.1941,
|
| 627 |
+
"step": 1850
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 10.0,
|
| 631 |
+
"eval_accuracy": 0.8461538461538461,
|
| 632 |
+
"eval_loss": 0.8470482230186462,
|
| 633 |
+
"eval_macro_f1": 0.8464161446254975,
|
| 634 |
+
"eval_runtime": 1.1933,
|
| 635 |
+
"eval_samples_per_second": 479.332,
|
| 636 |
+
"eval_steps_per_second": 30.168,
|
| 637 |
+
"eval_weighted_f1": 0.8457504742925153,
|
| 638 |
+
"step": 1870
|
| 639 |
+
}
|
| 640 |
+
],
|
| 641 |
+
"logging_steps": 25,
|
| 642 |
+
"max_steps": 3740,
|
| 643 |
+
"num_input_tokens_seen": 0,
|
| 644 |
+
"num_train_epochs": 20,
|
| 645 |
+
"save_steps": 500,
|
| 646 |
+
"stateful_callbacks": {
|
| 647 |
+
"EarlyStoppingCallback": {
|
| 648 |
+
"args": {
|
| 649 |
+
"early_stopping_patience": 3,
|
| 650 |
+
"early_stopping_threshold": 0.0
|
| 651 |
+
},
|
| 652 |
+
"attributes": {
|
| 653 |
+
"early_stopping_patience_counter": 3
|
| 654 |
+
}
|
| 655 |
+
},
|
| 656 |
+
"TrainerControl": {
|
| 657 |
+
"args": {
|
| 658 |
+
"should_epoch_stop": false,
|
| 659 |
+
"should_evaluate": false,
|
| 660 |
+
"should_log": false,
|
| 661 |
+
"should_save": true,
|
| 662 |
+
"should_training_stop": true
|
| 663 |
+
},
|
| 664 |
+
"attributes": {}
|
| 665 |
+
}
|
| 666 |
+
},
|
| 667 |
+
"total_flos": 0.0,
|
| 668 |
+
"train_batch_size": 8,
|
| 669 |
+
"trial_name": null,
|
| 670 |
+
"trial_params": null
|
| 671 |
+
}
|
ood-vifactcheck-phobert-topic-domain-00-s43-20260814T131947Z-bdd2c308/trainer/checkpoint-1870/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-01-s43-20260814T133200Z-bdd2c308/trainer/checkpoint-1146/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-topic-domain-01-s43-20260814T133200Z-bdd2c308/trainer/checkpoint-1146/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-01-s43-20260814T133200Z-bdd2c308/trainer/checkpoint-1146/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-s43-20260814T133200Z-bdd2c308/trainer/checkpoint-1146/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-s43-20260814T133200Z-bdd2c308/trainer/checkpoint-1146/trainer_state.json
ADDED
|
@@ -0,0 +1,424 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1146,
|
| 3 |
+
"best_metric": 0.8437205023046266,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-topic-domain-01-s43-20260814T133200Z-bdd2c308/trainer/checkpoint-1146",
|
| 5 |
+
"epoch": 6.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1146,
|
| 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": 6.507308483123779,
|
| 15 |
+
"learning_rate": 1.256544502617801e-06,
|
| 16 |
+
"loss": 4.4621,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.2621231979030144,
|
| 21 |
+
"grad_norm": 10.711575508117676,
|
| 22 |
+
"learning_rate": 2.5654450261780104e-06,
|
| 23 |
+
"loss": 4.4524,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.3931847968545216,
|
| 28 |
+
"grad_norm": 5.922560214996338,
|
| 29 |
+
"learning_rate": 3.874345549738221e-06,
|
| 30 |
+
"loss": 4.4021,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5242463958060288,
|
| 35 |
+
"grad_norm": 6.39362907409668,
|
| 36 |
+
"learning_rate": 5.18324607329843e-06,
|
| 37 |
+
"loss": 4.3856,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6553079947575361,
|
| 42 |
+
"grad_norm": 7.5367960929870605,
|
| 43 |
+
"learning_rate": 6.492146596858639e-06,
|
| 44 |
+
"loss": 4.3043,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.7863695937090432,
|
| 49 |
+
"grad_norm": 9.4915189743042,
|
| 50 |
+
"learning_rate": 7.80104712041885e-06,
|
| 51 |
+
"loss": 4.3187,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9174311926605505,
|
| 56 |
+
"grad_norm": 8.589058876037598,
|
| 57 |
+
"learning_rate": 9.109947643979057e-06,
|
| 58 |
+
"loss": 4.2166,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.5445205479452054,
|
| 64 |
+
"eval_loss": 0.9759103655815125,
|
| 65 |
+
"eval_macro_f1": 0.5322518162179672,
|
| 66 |
+
"eval_runtime": 1.1826,
|
| 67 |
+
"eval_samples_per_second": 493.836,
|
| 68 |
+
"eval_steps_per_second": 31.288,
|
| 69 |
+
"eval_weighted_f1": 0.5336594975823489,
|
| 70 |
+
"step": 191
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0471821756225426,
|
| 74 |
+
"grad_norm": 11.926389694213867,
|
| 75 |
+
"learning_rate": 1.0418848167539269e-05,
|
| 76 |
+
"loss": 3.9699,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.1782437745740497,
|
| 81 |
+
"grad_norm": 24.30746078491211,
|
| 82 |
+
"learning_rate": 1.1727748691099476e-05,
|
| 83 |
+
"loss": 3.7977,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.309305373525557,
|
| 88 |
+
"grad_norm": 46.72258377075195,
|
| 89 |
+
"learning_rate": 1.3036649214659686e-05,
|
| 90 |
+
"loss": 3.4673,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.4403669724770642,
|
| 95 |
+
"grad_norm": 16.41703224182129,
|
| 96 |
+
"learning_rate": 1.4345549738219897e-05,
|
| 97 |
+
"loss": 3.0033,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.5714285714285714,
|
| 102 |
+
"grad_norm": 113.51791381835938,
|
| 103 |
+
"learning_rate": 1.5654450261780107e-05,
|
| 104 |
+
"loss": 2.6431,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7024901703800785,
|
| 109 |
+
"grad_norm": 41.03617858886719,
|
| 110 |
+
"learning_rate": 1.6963350785340316e-05,
|
| 111 |
+
"loss": 2.4171,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.833551769331586,
|
| 116 |
+
"grad_norm": 40.75381851196289,
|
| 117 |
+
"learning_rate": 1.8272251308900526e-05,
|
| 118 |
+
"loss": 2.3819,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 1.964613368283093,
|
| 123 |
+
"grad_norm": 37.86177062988281,
|
| 124 |
+
"learning_rate": 1.9581151832460736e-05,
|
| 125 |
+
"loss": 2.2975,
|
| 126 |
+
"step": 375
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"epoch": 2.0,
|
| 130 |
+
"eval_accuracy": 0.8236301369863014,
|
| 131 |
+
"eval_loss": 0.5043326616287231,
|
| 132 |
+
"eval_macro_f1": 0.8232426802524113,
|
| 133 |
+
"eval_runtime": 1.1943,
|
| 134 |
+
"eval_samples_per_second": 488.98,
|
| 135 |
+
"eval_steps_per_second": 30.98,
|
| 136 |
+
"eval_weighted_f1": 0.8233629823810918,
|
| 137 |
+
"step": 382
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.094364351245085,
|
| 141 |
+
"grad_norm": 41.09190368652344,
|
| 142 |
+
"learning_rate": 1.9901105293775453e-05,
|
| 143 |
+
"loss": 1.7503,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.2254259501965925,
|
| 148 |
+
"grad_norm": 59.534854888916016,
|
| 149 |
+
"learning_rate": 1.9755671902268764e-05,
|
| 150 |
+
"loss": 2.0186,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.3564875491480994,
|
| 155 |
+
"grad_norm": 42.519718170166016,
|
| 156 |
+
"learning_rate": 1.961023851076207e-05,
|
| 157 |
+
"loss": 1.7058,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.487549148099607,
|
| 162 |
+
"grad_norm": 65.17015838623047,
|
| 163 |
+
"learning_rate": 1.9464805119255382e-05,
|
| 164 |
+
"loss": 1.8268,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.618610747051114,
|
| 169 |
+
"grad_norm": 42.40889358520508,
|
| 170 |
+
"learning_rate": 1.9319371727748693e-05,
|
| 171 |
+
"loss": 1.6248,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.749672346002621,
|
| 176 |
+
"grad_norm": 59.507144927978516,
|
| 177 |
+
"learning_rate": 1.9173938336242e-05,
|
| 178 |
+
"loss": 1.5336,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 2.8807339449541285,
|
| 183 |
+
"grad_norm": 56.277626037597656,
|
| 184 |
+
"learning_rate": 1.9028504944735315e-05,
|
| 185 |
+
"loss": 1.7832,
|
| 186 |
+
"step": 550
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 3.0,
|
| 190 |
+
"eval_accuracy": 0.8287671232876712,
|
| 191 |
+
"eval_loss": 0.4839979112148285,
|
| 192 |
+
"eval_macro_f1": 0.8277774662495601,
|
| 193 |
+
"eval_runtime": 1.1573,
|
| 194 |
+
"eval_samples_per_second": 504.644,
|
| 195 |
+
"eval_steps_per_second": 31.972,
|
| 196 |
+
"eval_weighted_f1": 0.8279726033495205,
|
| 197 |
+
"step": 573
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.0104849279161208,
|
| 201 |
+
"grad_norm": 16.086835861206055,
|
| 202 |
+
"learning_rate": 1.8883071553228622e-05,
|
| 203 |
+
"loss": 1.5576,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.1415465268676277,
|
| 208 |
+
"grad_norm": 38.5900764465332,
|
| 209 |
+
"learning_rate": 1.8737638161721933e-05,
|
| 210 |
+
"loss": 1.2188,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.272608125819135,
|
| 215 |
+
"grad_norm": 62.46556091308594,
|
| 216 |
+
"learning_rate": 1.8592204770215244e-05,
|
| 217 |
+
"loss": 1.1246,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.4036697247706424,
|
| 222 |
+
"grad_norm": 33.22957229614258,
|
| 223 |
+
"learning_rate": 1.844677137870855e-05,
|
| 224 |
+
"loss": 1.1218,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.5347313237221494,
|
| 229 |
+
"grad_norm": 51.590599060058594,
|
| 230 |
+
"learning_rate": 1.8301337987201865e-05,
|
| 231 |
+
"loss": 1.2173,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.6657929226736568,
|
| 236 |
+
"grad_norm": 60.49343490600586,
|
| 237 |
+
"learning_rate": 1.8155904595695173e-05,
|
| 238 |
+
"loss": 1.1069,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.7968545216251637,
|
| 243 |
+
"grad_norm": 13.992277145385742,
|
| 244 |
+
"learning_rate": 1.8010471204188483e-05,
|
| 245 |
+
"loss": 1.2053,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 3.927916120576671,
|
| 250 |
+
"grad_norm": 25.929916381835938,
|
| 251 |
+
"learning_rate": 1.7865037812681794e-05,
|
| 252 |
+
"loss": 1.2599,
|
| 253 |
+
"step": 750
|
| 254 |
+
},
|
| 255 |
+
{
|
| 256 |
+
"epoch": 4.0,
|
| 257 |
+
"eval_accuracy": 0.833904109589041,
|
| 258 |
+
"eval_loss": 0.46194958686828613,
|
| 259 |
+
"eval_macro_f1": 0.8338684891422288,
|
| 260 |
+
"eval_runtime": 1.1895,
|
| 261 |
+
"eval_samples_per_second": 490.968,
|
| 262 |
+
"eval_steps_per_second": 31.106,
|
| 263 |
+
"eval_weighted_f1": 0.8339768721491453,
|
| 264 |
+
"step": 764
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.057667103538663,
|
| 268 |
+
"grad_norm": 28.464736938476562,
|
| 269 |
+
"learning_rate": 1.77196044211751e-05,
|
| 270 |
+
"loss": 0.9678,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.18872870249017,
|
| 275 |
+
"grad_norm": 10.684721946716309,
|
| 276 |
+
"learning_rate": 1.7574171029668416e-05,
|
| 277 |
+
"loss": 0.7347,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.319790301441677,
|
| 282 |
+
"grad_norm": 28.156612396240234,
|
| 283 |
+
"learning_rate": 1.7428737638161723e-05,
|
| 284 |
+
"loss": 0.6759,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.450851900393185,
|
| 289 |
+
"grad_norm": 44.93278503417969,
|
| 290 |
+
"learning_rate": 1.7283304246655034e-05,
|
| 291 |
+
"loss": 0.7375,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.581913499344692,
|
| 296 |
+
"grad_norm": 65.99188232421875,
|
| 297 |
+
"learning_rate": 1.7137870855148345e-05,
|
| 298 |
+
"loss": 0.784,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.712975098296199,
|
| 303 |
+
"grad_norm": 121.46563720703125,
|
| 304 |
+
"learning_rate": 1.6992437463641652e-05,
|
| 305 |
+
"loss": 0.7848,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 4.844036697247707,
|
| 310 |
+
"grad_norm": 15.833542823791504,
|
| 311 |
+
"learning_rate": 1.6847004072134963e-05,
|
| 312 |
+
"loss": 0.7481,
|
| 313 |
+
"step": 925
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 4.975098296199214,
|
| 317 |
+
"grad_norm": 29.14629554748535,
|
| 318 |
+
"learning_rate": 1.6701570680628274e-05,
|
| 319 |
+
"loss": 0.8473,
|
| 320 |
+
"step": 950
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 5.0,
|
| 324 |
+
"eval_accuracy": 0.8253424657534246,
|
| 325 |
+
"eval_loss": 0.5701684951782227,
|
| 326 |
+
"eval_macro_f1": 0.8249473143027406,
|
| 327 |
+
"eval_runtime": 1.1368,
|
| 328 |
+
"eval_samples_per_second": 513.704,
|
| 329 |
+
"eval_steps_per_second": 32.546,
|
| 330 |
+
"eval_weighted_f1": 0.8252094165190567,
|
| 331 |
+
"step": 955
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.104849279161206,
|
| 335 |
+
"grad_norm": 38.70903015136719,
|
| 336 |
+
"learning_rate": 1.6556137289121585e-05,
|
| 337 |
+
"loss": 0.5724,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.235910878112713,
|
| 342 |
+
"grad_norm": 70.35674285888672,
|
| 343 |
+
"learning_rate": 1.6410703897614895e-05,
|
| 344 |
+
"loss": 0.5921,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.36697247706422,
|
| 349 |
+
"grad_norm": 0.7167691588401794,
|
| 350 |
+
"learning_rate": 1.6265270506108203e-05,
|
| 351 |
+
"loss": 0.471,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.498034076015728,
|
| 356 |
+
"grad_norm": 113.53590393066406,
|
| 357 |
+
"learning_rate": 1.6119837114601514e-05,
|
| 358 |
+
"loss": 0.5086,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.6290956749672345,
|
| 363 |
+
"grad_norm": 106.8655014038086,
|
| 364 |
+
"learning_rate": 1.5974403723094824e-05,
|
| 365 |
+
"loss": 0.5397,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 5.760157273918741,
|
| 370 |
+
"grad_norm": 60.988162994384766,
|
| 371 |
+
"learning_rate": 1.5828970331588135e-05,
|
| 372 |
+
"loss": 0.6833,
|
| 373 |
+
"step": 1100
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 5.891218872870249,
|
| 377 |
+
"grad_norm": 59.26218795776367,
|
| 378 |
+
"learning_rate": 1.5683536940081443e-05,
|
| 379 |
+
"loss": 0.6606,
|
| 380 |
+
"step": 1125
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 6.0,
|
| 384 |
+
"eval_accuracy": 0.8424657534246576,
|
| 385 |
+
"eval_loss": 0.6837798953056335,
|
| 386 |
+
"eval_macro_f1": 0.8437205023046266,
|
| 387 |
+
"eval_runtime": 1.1415,
|
| 388 |
+
"eval_samples_per_second": 511.606,
|
| 389 |
+
"eval_steps_per_second": 32.413,
|
| 390 |
+
"eval_weighted_f1": 0.8435703502320667,
|
| 391 |
+
"step": 1146
|
| 392 |
+
}
|
| 393 |
+
],
|
| 394 |
+
"logging_steps": 25,
|
| 395 |
+
"max_steps": 3820,
|
| 396 |
+
"num_input_tokens_seen": 0,
|
| 397 |
+
"num_train_epochs": 20,
|
| 398 |
+
"save_steps": 500,
|
| 399 |
+
"stateful_callbacks": {
|
| 400 |
+
"EarlyStoppingCallback": {
|
| 401 |
+
"args": {
|
| 402 |
+
"early_stopping_patience": 3,
|
| 403 |
+
"early_stopping_threshold": 0.0
|
| 404 |
+
},
|
| 405 |
+
"attributes": {
|
| 406 |
+
"early_stopping_patience_counter": 0
|
| 407 |
+
}
|
| 408 |
+
},
|
| 409 |
+
"TrainerControl": {
|
| 410 |
+
"args": {
|
| 411 |
+
"should_epoch_stop": false,
|
| 412 |
+
"should_evaluate": false,
|
| 413 |
+
"should_log": false,
|
| 414 |
+
"should_save": true,
|
| 415 |
+
"should_training_stop": false
|
| 416 |
+
},
|
| 417 |
+
"attributes": {}
|
| 418 |
+
}
|
| 419 |
+
},
|
| 420 |
+
"total_flos": 0.0,
|
| 421 |
+
"train_batch_size": 8,
|
| 422 |
+
"trial_name": null,
|
| 423 |
+
"trial_params": null
|
| 424 |
+
}
|
ood-vifactcheck-phobert-topic-domain-01-s43-20260814T133200Z-bdd2c308/trainer/checkpoint-1146/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-01-s43-20260814T133200Z-bdd2c308/trainer/checkpoint-1719/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-topic-domain-01-s43-20260814T133200Z-bdd2c308/trainer/checkpoint-1719/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-01-s43-20260814T133200Z-bdd2c308/trainer/checkpoint-1719/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-s43-20260814T133200Z-bdd2c308/trainer/checkpoint-1719/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-s43-20260814T133200Z-bdd2c308/trainer/checkpoint-1719/trainer_state.json
ADDED
|
@@ -0,0 +1,618 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1146,
|
| 3 |
+
"best_metric": 0.8437205023046266,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-topic-domain-01-s43-20260814T133200Z-bdd2c308/trainer/checkpoint-1146",
|
| 5 |
+
"epoch": 9.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1719,
|
| 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": 6.507308483123779,
|
| 15 |
+
"learning_rate": 1.256544502617801e-06,
|
| 16 |
+
"loss": 4.4621,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.2621231979030144,
|
| 21 |
+
"grad_norm": 10.711575508117676,
|
| 22 |
+
"learning_rate": 2.5654450261780104e-06,
|
| 23 |
+
"loss": 4.4524,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.3931847968545216,
|
| 28 |
+
"grad_norm": 5.922560214996338,
|
| 29 |
+
"learning_rate": 3.874345549738221e-06,
|
| 30 |
+
"loss": 4.4021,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5242463958060288,
|
| 35 |
+
"grad_norm": 6.39362907409668,
|
| 36 |
+
"learning_rate": 5.18324607329843e-06,
|
| 37 |
+
"loss": 4.3856,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6553079947575361,
|
| 42 |
+
"grad_norm": 7.5367960929870605,
|
| 43 |
+
"learning_rate": 6.492146596858639e-06,
|
| 44 |
+
"loss": 4.3043,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.7863695937090432,
|
| 49 |
+
"grad_norm": 9.4915189743042,
|
| 50 |
+
"learning_rate": 7.80104712041885e-06,
|
| 51 |
+
"loss": 4.3187,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9174311926605505,
|
| 56 |
+
"grad_norm": 8.589058876037598,
|
| 57 |
+
"learning_rate": 9.109947643979057e-06,
|
| 58 |
+
"loss": 4.2166,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.5445205479452054,
|
| 64 |
+
"eval_loss": 0.9759103655815125,
|
| 65 |
+
"eval_macro_f1": 0.5322518162179672,
|
| 66 |
+
"eval_runtime": 1.1826,
|
| 67 |
+
"eval_samples_per_second": 493.836,
|
| 68 |
+
"eval_steps_per_second": 31.288,
|
| 69 |
+
"eval_weighted_f1": 0.5336594975823489,
|
| 70 |
+
"step": 191
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0471821756225426,
|
| 74 |
+
"grad_norm": 11.926389694213867,
|
| 75 |
+
"learning_rate": 1.0418848167539269e-05,
|
| 76 |
+
"loss": 3.9699,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.1782437745740497,
|
| 81 |
+
"grad_norm": 24.30746078491211,
|
| 82 |
+
"learning_rate": 1.1727748691099476e-05,
|
| 83 |
+
"loss": 3.7977,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.309305373525557,
|
| 88 |
+
"grad_norm": 46.72258377075195,
|
| 89 |
+
"learning_rate": 1.3036649214659686e-05,
|
| 90 |
+
"loss": 3.4673,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.4403669724770642,
|
| 95 |
+
"grad_norm": 16.41703224182129,
|
| 96 |
+
"learning_rate": 1.4345549738219897e-05,
|
| 97 |
+
"loss": 3.0033,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.5714285714285714,
|
| 102 |
+
"grad_norm": 113.51791381835938,
|
| 103 |
+
"learning_rate": 1.5654450261780107e-05,
|
| 104 |
+
"loss": 2.6431,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7024901703800785,
|
| 109 |
+
"grad_norm": 41.03617858886719,
|
| 110 |
+
"learning_rate": 1.6963350785340316e-05,
|
| 111 |
+
"loss": 2.4171,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.833551769331586,
|
| 116 |
+
"grad_norm": 40.75381851196289,
|
| 117 |
+
"learning_rate": 1.8272251308900526e-05,
|
| 118 |
+
"loss": 2.3819,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 1.964613368283093,
|
| 123 |
+
"grad_norm": 37.86177062988281,
|
| 124 |
+
"learning_rate": 1.9581151832460736e-05,
|
| 125 |
+
"loss": 2.2975,
|
| 126 |
+
"step": 375
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"epoch": 2.0,
|
| 130 |
+
"eval_accuracy": 0.8236301369863014,
|
| 131 |
+
"eval_loss": 0.5043326616287231,
|
| 132 |
+
"eval_macro_f1": 0.8232426802524113,
|
| 133 |
+
"eval_runtime": 1.1943,
|
| 134 |
+
"eval_samples_per_second": 488.98,
|
| 135 |
+
"eval_steps_per_second": 30.98,
|
| 136 |
+
"eval_weighted_f1": 0.8233629823810918,
|
| 137 |
+
"step": 382
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.094364351245085,
|
| 141 |
+
"grad_norm": 41.09190368652344,
|
| 142 |
+
"learning_rate": 1.9901105293775453e-05,
|
| 143 |
+
"loss": 1.7503,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.2254259501965925,
|
| 148 |
+
"grad_norm": 59.534854888916016,
|
| 149 |
+
"learning_rate": 1.9755671902268764e-05,
|
| 150 |
+
"loss": 2.0186,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.3564875491480994,
|
| 155 |
+
"grad_norm": 42.519718170166016,
|
| 156 |
+
"learning_rate": 1.961023851076207e-05,
|
| 157 |
+
"loss": 1.7058,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.487549148099607,
|
| 162 |
+
"grad_norm": 65.17015838623047,
|
| 163 |
+
"learning_rate": 1.9464805119255382e-05,
|
| 164 |
+
"loss": 1.8268,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.618610747051114,
|
| 169 |
+
"grad_norm": 42.40889358520508,
|
| 170 |
+
"learning_rate": 1.9319371727748693e-05,
|
| 171 |
+
"loss": 1.6248,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.749672346002621,
|
| 176 |
+
"grad_norm": 59.507144927978516,
|
| 177 |
+
"learning_rate": 1.9173938336242e-05,
|
| 178 |
+
"loss": 1.5336,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 2.8807339449541285,
|
| 183 |
+
"grad_norm": 56.277626037597656,
|
| 184 |
+
"learning_rate": 1.9028504944735315e-05,
|
| 185 |
+
"loss": 1.7832,
|
| 186 |
+
"step": 550
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 3.0,
|
| 190 |
+
"eval_accuracy": 0.8287671232876712,
|
| 191 |
+
"eval_loss": 0.4839979112148285,
|
| 192 |
+
"eval_macro_f1": 0.8277774662495601,
|
| 193 |
+
"eval_runtime": 1.1573,
|
| 194 |
+
"eval_samples_per_second": 504.644,
|
| 195 |
+
"eval_steps_per_second": 31.972,
|
| 196 |
+
"eval_weighted_f1": 0.8279726033495205,
|
| 197 |
+
"step": 573
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.0104849279161208,
|
| 201 |
+
"grad_norm": 16.086835861206055,
|
| 202 |
+
"learning_rate": 1.8883071553228622e-05,
|
| 203 |
+
"loss": 1.5576,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.1415465268676277,
|
| 208 |
+
"grad_norm": 38.5900764465332,
|
| 209 |
+
"learning_rate": 1.8737638161721933e-05,
|
| 210 |
+
"loss": 1.2188,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.272608125819135,
|
| 215 |
+
"grad_norm": 62.46556091308594,
|
| 216 |
+
"learning_rate": 1.8592204770215244e-05,
|
| 217 |
+
"loss": 1.1246,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.4036697247706424,
|
| 222 |
+
"grad_norm": 33.22957229614258,
|
| 223 |
+
"learning_rate": 1.844677137870855e-05,
|
| 224 |
+
"loss": 1.1218,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.5347313237221494,
|
| 229 |
+
"grad_norm": 51.590599060058594,
|
| 230 |
+
"learning_rate": 1.8301337987201865e-05,
|
| 231 |
+
"loss": 1.2173,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.6657929226736568,
|
| 236 |
+
"grad_norm": 60.49343490600586,
|
| 237 |
+
"learning_rate": 1.8155904595695173e-05,
|
| 238 |
+
"loss": 1.1069,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.7968545216251637,
|
| 243 |
+
"grad_norm": 13.992277145385742,
|
| 244 |
+
"learning_rate": 1.8010471204188483e-05,
|
| 245 |
+
"loss": 1.2053,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 3.927916120576671,
|
| 250 |
+
"grad_norm": 25.929916381835938,
|
| 251 |
+
"learning_rate": 1.7865037812681794e-05,
|
| 252 |
+
"loss": 1.2599,
|
| 253 |
+
"step": 750
|
| 254 |
+
},
|
| 255 |
+
{
|
| 256 |
+
"epoch": 4.0,
|
| 257 |
+
"eval_accuracy": 0.833904109589041,
|
| 258 |
+
"eval_loss": 0.46194958686828613,
|
| 259 |
+
"eval_macro_f1": 0.8338684891422288,
|
| 260 |
+
"eval_runtime": 1.1895,
|
| 261 |
+
"eval_samples_per_second": 490.968,
|
| 262 |
+
"eval_steps_per_second": 31.106,
|
| 263 |
+
"eval_weighted_f1": 0.8339768721491453,
|
| 264 |
+
"step": 764
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.057667103538663,
|
| 268 |
+
"grad_norm": 28.464736938476562,
|
| 269 |
+
"learning_rate": 1.77196044211751e-05,
|
| 270 |
+
"loss": 0.9678,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.18872870249017,
|
| 275 |
+
"grad_norm": 10.684721946716309,
|
| 276 |
+
"learning_rate": 1.7574171029668416e-05,
|
| 277 |
+
"loss": 0.7347,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.319790301441677,
|
| 282 |
+
"grad_norm": 28.156612396240234,
|
| 283 |
+
"learning_rate": 1.7428737638161723e-05,
|
| 284 |
+
"loss": 0.6759,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.450851900393185,
|
| 289 |
+
"grad_norm": 44.93278503417969,
|
| 290 |
+
"learning_rate": 1.7283304246655034e-05,
|
| 291 |
+
"loss": 0.7375,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.581913499344692,
|
| 296 |
+
"grad_norm": 65.99188232421875,
|
| 297 |
+
"learning_rate": 1.7137870855148345e-05,
|
| 298 |
+
"loss": 0.784,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.712975098296199,
|
| 303 |
+
"grad_norm": 121.46563720703125,
|
| 304 |
+
"learning_rate": 1.6992437463641652e-05,
|
| 305 |
+
"loss": 0.7848,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 4.844036697247707,
|
| 310 |
+
"grad_norm": 15.833542823791504,
|
| 311 |
+
"learning_rate": 1.6847004072134963e-05,
|
| 312 |
+
"loss": 0.7481,
|
| 313 |
+
"step": 925
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 4.975098296199214,
|
| 317 |
+
"grad_norm": 29.14629554748535,
|
| 318 |
+
"learning_rate": 1.6701570680628274e-05,
|
| 319 |
+
"loss": 0.8473,
|
| 320 |
+
"step": 950
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 5.0,
|
| 324 |
+
"eval_accuracy": 0.8253424657534246,
|
| 325 |
+
"eval_loss": 0.5701684951782227,
|
| 326 |
+
"eval_macro_f1": 0.8249473143027406,
|
| 327 |
+
"eval_runtime": 1.1368,
|
| 328 |
+
"eval_samples_per_second": 513.704,
|
| 329 |
+
"eval_steps_per_second": 32.546,
|
| 330 |
+
"eval_weighted_f1": 0.8252094165190567,
|
| 331 |
+
"step": 955
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.104849279161206,
|
| 335 |
+
"grad_norm": 38.70903015136719,
|
| 336 |
+
"learning_rate": 1.6556137289121585e-05,
|
| 337 |
+
"loss": 0.5724,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.235910878112713,
|
| 342 |
+
"grad_norm": 70.35674285888672,
|
| 343 |
+
"learning_rate": 1.6410703897614895e-05,
|
| 344 |
+
"loss": 0.5921,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.36697247706422,
|
| 349 |
+
"grad_norm": 0.7167691588401794,
|
| 350 |
+
"learning_rate": 1.6265270506108203e-05,
|
| 351 |
+
"loss": 0.471,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.498034076015728,
|
| 356 |
+
"grad_norm": 113.53590393066406,
|
| 357 |
+
"learning_rate": 1.6119837114601514e-05,
|
| 358 |
+
"loss": 0.5086,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.6290956749672345,
|
| 363 |
+
"grad_norm": 106.8655014038086,
|
| 364 |
+
"learning_rate": 1.5974403723094824e-05,
|
| 365 |
+
"loss": 0.5397,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 5.760157273918741,
|
| 370 |
+
"grad_norm": 60.988162994384766,
|
| 371 |
+
"learning_rate": 1.5828970331588135e-05,
|
| 372 |
+
"loss": 0.6833,
|
| 373 |
+
"step": 1100
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 5.891218872870249,
|
| 377 |
+
"grad_norm": 59.26218795776367,
|
| 378 |
+
"learning_rate": 1.5683536940081443e-05,
|
| 379 |
+
"loss": 0.6606,
|
| 380 |
+
"step": 1125
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 6.0,
|
| 384 |
+
"eval_accuracy": 0.8424657534246576,
|
| 385 |
+
"eval_loss": 0.6837798953056335,
|
| 386 |
+
"eval_macro_f1": 0.8437205023046266,
|
| 387 |
+
"eval_runtime": 1.1415,
|
| 388 |
+
"eval_samples_per_second": 511.606,
|
| 389 |
+
"eval_steps_per_second": 32.413,
|
| 390 |
+
"eval_weighted_f1": 0.8435703502320667,
|
| 391 |
+
"step": 1146
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 6.0209698558322415,
|
| 395 |
+
"grad_norm": 68.30610656738281,
|
| 396 |
+
"learning_rate": 1.5538103548574753e-05,
|
| 397 |
+
"loss": 0.6209,
|
| 398 |
+
"step": 1150
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.1520314547837485,
|
| 402 |
+
"grad_norm": 11.186919212341309,
|
| 403 |
+
"learning_rate": 1.5392670157068064e-05,
|
| 404 |
+
"loss": 0.4205,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.283093053735255,
|
| 409 |
+
"grad_norm": 0.615466296672821,
|
| 410 |
+
"learning_rate": 1.5247236765561375e-05,
|
| 411 |
+
"loss": 0.3474,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.414154652686763,
|
| 416 |
+
"grad_norm": 15.087629318237305,
|
| 417 |
+
"learning_rate": 1.5101803374054684e-05,
|
| 418 |
+
"loss": 0.3894,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.54521625163827,
|
| 423 |
+
"grad_norm": 20.47812843322754,
|
| 424 |
+
"learning_rate": 1.4956369982547993e-05,
|
| 425 |
+
"loss": 0.4567,
|
| 426 |
+
"step": 1250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 6.676277850589777,
|
| 430 |
+
"grad_norm": 6.125617980957031,
|
| 431 |
+
"learning_rate": 1.4810936591041306e-05,
|
| 432 |
+
"loss": 0.378,
|
| 433 |
+
"step": 1275
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 6.807339449541285,
|
| 437 |
+
"grad_norm": 112.74131774902344,
|
| 438 |
+
"learning_rate": 1.4665503199534615e-05,
|
| 439 |
+
"loss": 0.5217,
|
| 440 |
+
"step": 1300
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 6.938401048492792,
|
| 444 |
+
"grad_norm": 0.35432302951812744,
|
| 445 |
+
"learning_rate": 1.4520069808027924e-05,
|
| 446 |
+
"loss": 0.4555,
|
| 447 |
+
"step": 1325
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 7.0,
|
| 451 |
+
"eval_accuracy": 0.827054794520548,
|
| 452 |
+
"eval_loss": 0.7681780457496643,
|
| 453 |
+
"eval_macro_f1": 0.8265861011647528,
|
| 454 |
+
"eval_runtime": 1.1576,
|
| 455 |
+
"eval_samples_per_second": 504.498,
|
| 456 |
+
"eval_steps_per_second": 31.963,
|
| 457 |
+
"eval_weighted_f1": 0.8266571310596437,
|
| 458 |
+
"step": 1337
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 7.068152031454784,
|
| 462 |
+
"grad_norm": 123.20954132080078,
|
| 463 |
+
"learning_rate": 1.4374636416521235e-05,
|
| 464 |
+
"loss": 0.474,
|
| 465 |
+
"step": 1350
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 7.199213630406291,
|
| 469 |
+
"grad_norm": 297.6741943359375,
|
| 470 |
+
"learning_rate": 1.4229203025014544e-05,
|
| 471 |
+
"loss": 0.3764,
|
| 472 |
+
"step": 1375
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 7.330275229357798,
|
| 476 |
+
"grad_norm": 50.92286682128906,
|
| 477 |
+
"learning_rate": 1.4083769633507855e-05,
|
| 478 |
+
"loss": 0.2421,
|
| 479 |
+
"step": 1400
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 7.461336828309306,
|
| 483 |
+
"grad_norm": 366.05426025390625,
|
| 484 |
+
"learning_rate": 1.3938336242001166e-05,
|
| 485 |
+
"loss": 0.4125,
|
| 486 |
+
"step": 1425
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 7.592398427260813,
|
| 490 |
+
"grad_norm": 0.3513495922088623,
|
| 491 |
+
"learning_rate": 1.3792902850494475e-05,
|
| 492 |
+
"loss": 0.3016,
|
| 493 |
+
"step": 1450
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 7.72346002621232,
|
| 497 |
+
"grad_norm": 27.299068450927734,
|
| 498 |
+
"learning_rate": 1.3647469458987785e-05,
|
| 499 |
+
"loss": 0.5213,
|
| 500 |
+
"step": 1475
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 7.854521625163827,
|
| 504 |
+
"grad_norm": 36.74614334106445,
|
| 505 |
+
"learning_rate": 1.3502036067481095e-05,
|
| 506 |
+
"loss": 0.4195,
|
| 507 |
+
"step": 1500
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 7.985583224115334,
|
| 511 |
+
"grad_norm": 219.2817840576172,
|
| 512 |
+
"learning_rate": 1.3356602675974404e-05,
|
| 513 |
+
"loss": 0.3384,
|
| 514 |
+
"step": 1525
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 8.0,
|
| 518 |
+
"eval_accuracy": 0.833904109589041,
|
| 519 |
+
"eval_loss": 0.8284252882003784,
|
| 520 |
+
"eval_macro_f1": 0.8342311242203179,
|
| 521 |
+
"eval_runtime": 1.1829,
|
| 522 |
+
"eval_samples_per_second": 493.705,
|
| 523 |
+
"eval_steps_per_second": 31.279,
|
| 524 |
+
"eval_weighted_f1": 0.8341418803790681,
|
| 525 |
+
"step": 1528
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 8.115334207077327,
|
| 529 |
+
"grad_norm": 0.2374507486820221,
|
| 530 |
+
"learning_rate": 1.3211169284467716e-05,
|
| 531 |
+
"loss": 0.2521,
|
| 532 |
+
"step": 1550
|
| 533 |
+
},
|
| 534 |
+
{
|
| 535 |
+
"epoch": 8.246395806028833,
|
| 536 |
+
"grad_norm": 15.147400856018066,
|
| 537 |
+
"learning_rate": 1.3065735892961025e-05,
|
| 538 |
+
"loss": 0.2342,
|
| 539 |
+
"step": 1575
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 8.37745740498034,
|
| 543 |
+
"grad_norm": 0.26705312728881836,
|
| 544 |
+
"learning_rate": 1.2920302501454336e-05,
|
| 545 |
+
"loss": 0.3186,
|
| 546 |
+
"step": 1600
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 8.508519003931848,
|
| 550 |
+
"grad_norm": 25.72291374206543,
|
| 551 |
+
"learning_rate": 1.2774869109947645e-05,
|
| 552 |
+
"loss": 0.2126,
|
| 553 |
+
"step": 1625
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 8.639580602883354,
|
| 557 |
+
"grad_norm": 0.2880823314189911,
|
| 558 |
+
"learning_rate": 1.2629435718440954e-05,
|
| 559 |
+
"loss": 0.1812,
|
| 560 |
+
"step": 1650
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 8.770642201834862,
|
| 564 |
+
"grad_norm": 2.384773015975952,
|
| 565 |
+
"learning_rate": 1.2484002326934265e-05,
|
| 566 |
+
"loss": 0.3413,
|
| 567 |
+
"step": 1675
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 8.90170380078637,
|
| 571 |
+
"grad_norm": 4.403160572052002,
|
| 572 |
+
"learning_rate": 1.2338568935427576e-05,
|
| 573 |
+
"loss": 0.1985,
|
| 574 |
+
"step": 1700
|
| 575 |
+
},
|
| 576 |
+
{
|
| 577 |
+
"epoch": 9.0,
|
| 578 |
+
"eval_accuracy": 0.8407534246575342,
|
| 579 |
+
"eval_loss": 0.8359805345535278,
|
| 580 |
+
"eval_macro_f1": 0.8403922473790963,
|
| 581 |
+
"eval_runtime": 1.1421,
|
| 582 |
+
"eval_samples_per_second": 511.338,
|
| 583 |
+
"eval_steps_per_second": 32.396,
|
| 584 |
+
"eval_weighted_f1": 0.8404904459864052,
|
| 585 |
+
"step": 1719
|
| 586 |
+
}
|
| 587 |
+
],
|
| 588 |
+
"logging_steps": 25,
|
| 589 |
+
"max_steps": 3820,
|
| 590 |
+
"num_input_tokens_seen": 0,
|
| 591 |
+
"num_train_epochs": 20,
|
| 592 |
+
"save_steps": 500,
|
| 593 |
+
"stateful_callbacks": {
|
| 594 |
+
"EarlyStoppingCallback": {
|
| 595 |
+
"args": {
|
| 596 |
+
"early_stopping_patience": 3,
|
| 597 |
+
"early_stopping_threshold": 0.0
|
| 598 |
+
},
|
| 599 |
+
"attributes": {
|
| 600 |
+
"early_stopping_patience_counter": 3
|
| 601 |
+
}
|
| 602 |
+
},
|
| 603 |
+
"TrainerControl": {
|
| 604 |
+
"args": {
|
| 605 |
+
"should_epoch_stop": false,
|
| 606 |
+
"should_evaluate": false,
|
| 607 |
+
"should_log": false,
|
| 608 |
+
"should_save": true,
|
| 609 |
+
"should_training_stop": true
|
| 610 |
+
},
|
| 611 |
+
"attributes": {}
|
| 612 |
+
}
|
| 613 |
+
},
|
| 614 |
+
"total_flos": 0.0,
|
| 615 |
+
"train_batch_size": 8,
|
| 616 |
+
"trial_name": null,
|
| 617 |
+
"trial_params": null
|
| 618 |
+
}
|
ood-vifactcheck-phobert-topic-domain-01-s43-20260814T133200Z-bdd2c308/trainer/checkpoint-1719/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-02-s43-20260814T134559Z-bdd2c308/trainer/checkpoint-1092/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-topic-domain-02-s43-20260814T134559Z-bdd2c308/trainer/checkpoint-1092/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-02-s43-20260814T134559Z-bdd2c308/trainer/checkpoint-1092/special_tokens_map.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": "<mask>",
|
| 6 |
+
"pad_token": "<pad>",
|
| 7 |
+
"sep_token": "</s>",
|
| 8 |
+
"unk_token": "<unk>"
|
| 9 |
+
}
|
ood-vifactcheck-phobert-topic-domain-02-s43-20260814T134559Z-bdd2c308/trainer/checkpoint-1092/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"64000": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 256,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "PhobertTokenizer",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-phobert-topic-domain-02-s43-20260814T134559Z-bdd2c308/trainer/checkpoint-1092/trainer_state.json
ADDED
|
@@ -0,0 +1,410 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1092,
|
| 3 |
+
"best_metric": 0.8548617051611064,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-topic-domain-02-s43-20260814T134559Z-bdd2c308/trainer/checkpoint-1092",
|
| 5 |
+
"epoch": 6.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1092,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.13793103448275862,
|
| 14 |
+
"grad_norm": 6.502751350402832,
|
| 15 |
+
"learning_rate": 1.3186813186813187e-06,
|
| 16 |
+
"loss": 4.4294,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.27586206896551724,
|
| 21 |
+
"grad_norm": 6.533897399902344,
|
| 22 |
+
"learning_rate": 2.6923076923076923e-06,
|
| 23 |
+
"loss": 4.4498,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.41379310344827586,
|
| 28 |
+
"grad_norm": 6.485318660736084,
|
| 29 |
+
"learning_rate": 4.065934065934066e-06,
|
| 30 |
+
"loss": 4.4152,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5517241379310345,
|
| 35 |
+
"grad_norm": 7.700263023376465,
|
| 36 |
+
"learning_rate": 5.43956043956044e-06,
|
| 37 |
+
"loss": 4.3639,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6896551724137931,
|
| 42 |
+
"grad_norm": 9.55636215209961,
|
| 43 |
+
"learning_rate": 6.813186813186814e-06,
|
| 44 |
+
"loss": 4.3391,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8275862068965517,
|
| 49 |
+
"grad_norm": 11.730230331420898,
|
| 50 |
+
"learning_rate": 8.186813186813188e-06,
|
| 51 |
+
"loss": 4.2742,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9655172413793104,
|
| 56 |
+
"grad_norm": 10.071218490600586,
|
| 57 |
+
"learning_rate": 9.560439560439562e-06,
|
| 58 |
+
"loss": 4.2261,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.510934393638171,
|
| 64 |
+
"eval_loss": 1.0159550905227661,
|
| 65 |
+
"eval_macro_f1": 0.49256508563077905,
|
| 66 |
+
"eval_runtime": 1.1,
|
| 67 |
+
"eval_samples_per_second": 457.273,
|
| 68 |
+
"eval_steps_per_second": 29.091,
|
| 69 |
+
"eval_weighted_f1": 0.4937639432603952,
|
| 70 |
+
"step": 182
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0993103448275863,
|
| 74 |
+
"grad_norm": 17.43917465209961,
|
| 75 |
+
"learning_rate": 1.0934065934065933e-05,
|
| 76 |
+
"loss": 3.8656,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.237241379310345,
|
| 81 |
+
"grad_norm": 22.632905960083008,
|
| 82 |
+
"learning_rate": 1.230769230769231e-05,
|
| 83 |
+
"loss": 3.6757,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.3751724137931034,
|
| 88 |
+
"grad_norm": 27.157556533813477,
|
| 89 |
+
"learning_rate": 1.3681318681318683e-05,
|
| 90 |
+
"loss": 3.3253,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.513103448275862,
|
| 95 |
+
"grad_norm": 35.907798767089844,
|
| 96 |
+
"learning_rate": 1.5054945054945056e-05,
|
| 97 |
+
"loss": 2.9085,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.6510344827586207,
|
| 102 |
+
"grad_norm": 35.90727996826172,
|
| 103 |
+
"learning_rate": 1.642857142857143e-05,
|
| 104 |
+
"loss": 2.7074,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7889655172413792,
|
| 109 |
+
"grad_norm": 33.298580169677734,
|
| 110 |
+
"learning_rate": 1.78021978021978e-05,
|
| 111 |
+
"loss": 2.2439,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.926896551724138,
|
| 116 |
+
"grad_norm": 32.35009002685547,
|
| 117 |
+
"learning_rate": 1.9175824175824178e-05,
|
| 118 |
+
"loss": 2.1258,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.805168986083499,
|
| 124 |
+
"eval_loss": 0.49421149492263794,
|
| 125 |
+
"eval_macro_f1": 0.80480474458109,
|
| 126 |
+
"eval_runtime": 1.0389,
|
| 127 |
+
"eval_samples_per_second": 484.152,
|
| 128 |
+
"eval_steps_per_second": 30.801,
|
| 129 |
+
"eval_weighted_f1": 0.8051478160810519,
|
| 130 |
+
"step": 364
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.0606896551724136,
|
| 134 |
+
"grad_norm": 26.5976619720459,
|
| 135 |
+
"learning_rate": 1.993894993894994e-05,
|
| 136 |
+
"loss": 2.0294,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.1986206896551725,
|
| 141 |
+
"grad_norm": 60.08498764038086,
|
| 142 |
+
"learning_rate": 1.9786324786324788e-05,
|
| 143 |
+
"loss": 1.6727,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.336551724137931,
|
| 148 |
+
"grad_norm": 45.483428955078125,
|
| 149 |
+
"learning_rate": 1.9633699633699634e-05,
|
| 150 |
+
"loss": 1.7497,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.47448275862069,
|
| 155 |
+
"grad_norm": 56.38203430175781,
|
| 156 |
+
"learning_rate": 1.948107448107448e-05,
|
| 157 |
+
"loss": 1.6546,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.6124137931034483,
|
| 162 |
+
"grad_norm": 43.6146125793457,
|
| 163 |
+
"learning_rate": 1.932844932844933e-05,
|
| 164 |
+
"loss": 1.6076,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.750344827586207,
|
| 169 |
+
"grad_norm": 104.45264434814453,
|
| 170 |
+
"learning_rate": 1.9175824175824178e-05,
|
| 171 |
+
"loss": 1.665,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.8882758620689657,
|
| 176 |
+
"grad_norm": 37.07889175415039,
|
| 177 |
+
"learning_rate": 1.9023199023199025e-05,
|
| 178 |
+
"loss": 1.6418,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 3.0,
|
| 183 |
+
"eval_accuracy": 0.852882703777336,
|
| 184 |
+
"eval_loss": 0.4100320041179657,
|
| 185 |
+
"eval_macro_f1": 0.8531422835626169,
|
| 186 |
+
"eval_runtime": 1.2145,
|
| 187 |
+
"eval_samples_per_second": 414.168,
|
| 188 |
+
"eval_steps_per_second": 26.349,
|
| 189 |
+
"eval_weighted_f1": 0.8531716659680295,
|
| 190 |
+
"step": 546
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.0220689655172412,
|
| 194 |
+
"grad_norm": 30.45008087158203,
|
| 195 |
+
"learning_rate": 1.887057387057387e-05,
|
| 196 |
+
"loss": 1.4073,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.16,
|
| 201 |
+
"grad_norm": 31.528188705444336,
|
| 202 |
+
"learning_rate": 1.8717948717948718e-05,
|
| 203 |
+
"loss": 1.1455,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.2979310344827586,
|
| 208 |
+
"grad_norm": 47.32204818725586,
|
| 209 |
+
"learning_rate": 1.8565323565323568e-05,
|
| 210 |
+
"loss": 1.1364,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.435862068965517,
|
| 215 |
+
"grad_norm": 62.04407501220703,
|
| 216 |
+
"learning_rate": 1.8412698412698415e-05,
|
| 217 |
+
"loss": 1.1754,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.573793103448276,
|
| 222 |
+
"grad_norm": 61.582069396972656,
|
| 223 |
+
"learning_rate": 1.826007326007326e-05,
|
| 224 |
+
"loss": 1.1587,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.7117241379310344,
|
| 229 |
+
"grad_norm": 78.23368072509766,
|
| 230 |
+
"learning_rate": 1.8107448107448108e-05,
|
| 231 |
+
"loss": 1.1347,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.849655172413793,
|
| 236 |
+
"grad_norm": 80.21173095703125,
|
| 237 |
+
"learning_rate": 1.7954822954822958e-05,
|
| 238 |
+
"loss": 1.1009,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.9875862068965517,
|
| 243 |
+
"grad_norm": 37.001888275146484,
|
| 244 |
+
"learning_rate": 1.78021978021978e-05,
|
| 245 |
+
"loss": 0.9838,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 4.0,
|
| 250 |
+
"eval_accuracy": 0.8508946322067594,
|
| 251 |
+
"eval_loss": 0.4349372386932373,
|
| 252 |
+
"eval_macro_f1": 0.8503660235798499,
|
| 253 |
+
"eval_runtime": 1.2065,
|
| 254 |
+
"eval_samples_per_second": 416.902,
|
| 255 |
+
"eval_steps_per_second": 26.523,
|
| 256 |
+
"eval_weighted_f1": 0.8506039358703087,
|
| 257 |
+
"step": 728
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.121379310344827,
|
| 261 |
+
"grad_norm": 8.690960884094238,
|
| 262 |
+
"learning_rate": 1.764957264957265e-05,
|
| 263 |
+
"loss": 0.6978,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.259310344827586,
|
| 268 |
+
"grad_norm": 55.83857727050781,
|
| 269 |
+
"learning_rate": 1.74969474969475e-05,
|
| 270 |
+
"loss": 0.9204,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.397241379310345,
|
| 275 |
+
"grad_norm": 49.880088806152344,
|
| 276 |
+
"learning_rate": 1.7344322344322345e-05,
|
| 277 |
+
"loss": 0.8488,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.535172413793103,
|
| 282 |
+
"grad_norm": 30.68686866760254,
|
| 283 |
+
"learning_rate": 1.7191697191697195e-05,
|
| 284 |
+
"loss": 0.765,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.673103448275862,
|
| 289 |
+
"grad_norm": 173.29953002929688,
|
| 290 |
+
"learning_rate": 1.703907203907204e-05,
|
| 291 |
+
"loss": 0.6741,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.811034482758621,
|
| 296 |
+
"grad_norm": 30.304973602294922,
|
| 297 |
+
"learning_rate": 1.688644688644689e-05,
|
| 298 |
+
"loss": 0.7912,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.94896551724138,
|
| 303 |
+
"grad_norm": 19.50191307067871,
|
| 304 |
+
"learning_rate": 1.6733821733821735e-05,
|
| 305 |
+
"loss": 0.7726,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 5.0,
|
| 310 |
+
"eval_accuracy": 0.8548707753479126,
|
| 311 |
+
"eval_loss": 0.4566419720649719,
|
| 312 |
+
"eval_macro_f1": 0.8532414420320026,
|
| 313 |
+
"eval_runtime": 1.2038,
|
| 314 |
+
"eval_samples_per_second": 417.843,
|
| 315 |
+
"eval_steps_per_second": 26.582,
|
| 316 |
+
"eval_weighted_f1": 0.8533275301022799,
|
| 317 |
+
"step": 910
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.082758620689655,
|
| 321 |
+
"grad_norm": 26.890073776245117,
|
| 322 |
+
"learning_rate": 1.6581196581196585e-05,
|
| 323 |
+
"loss": 0.4669,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.220689655172414,
|
| 328 |
+
"grad_norm": 110.97564697265625,
|
| 329 |
+
"learning_rate": 1.642857142857143e-05,
|
| 330 |
+
"loss": 0.4888,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.358620689655172,
|
| 335 |
+
"grad_norm": 62.88267135620117,
|
| 336 |
+
"learning_rate": 1.627594627594628e-05,
|
| 337 |
+
"loss": 0.6254,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.496551724137931,
|
| 342 |
+
"grad_norm": 2.0096068382263184,
|
| 343 |
+
"learning_rate": 1.6123321123321125e-05,
|
| 344 |
+
"loss": 0.5119,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.63448275862069,
|
| 349 |
+
"grad_norm": 18.092445373535156,
|
| 350 |
+
"learning_rate": 1.5970695970695972e-05,
|
| 351 |
+
"loss": 0.5095,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.772413793103448,
|
| 356 |
+
"grad_norm": 27.10129165649414,
|
| 357 |
+
"learning_rate": 1.581807081807082e-05,
|
| 358 |
+
"loss": 0.4641,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.910344827586207,
|
| 363 |
+
"grad_norm": 66.32721710205078,
|
| 364 |
+
"learning_rate": 1.5665445665445665e-05,
|
| 365 |
+
"loss": 0.7483,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 6.0,
|
| 370 |
+
"eval_accuracy": 0.8548707753479126,
|
| 371 |
+
"eval_loss": 0.5105668306350708,
|
| 372 |
+
"eval_macro_f1": 0.8548617051611064,
|
| 373 |
+
"eval_runtime": 1.1636,
|
| 374 |
+
"eval_samples_per_second": 432.262,
|
| 375 |
+
"eval_steps_per_second": 27.5,
|
| 376 |
+
"eval_weighted_f1": 0.8548737515029583,
|
| 377 |
+
"step": 1092
|
| 378 |
+
}
|
| 379 |
+
],
|
| 380 |
+
"logging_steps": 25,
|
| 381 |
+
"max_steps": 3640,
|
| 382 |
+
"num_input_tokens_seen": 0,
|
| 383 |
+
"num_train_epochs": 20,
|
| 384 |
+
"save_steps": 500,
|
| 385 |
+
"stateful_callbacks": {
|
| 386 |
+
"EarlyStoppingCallback": {
|
| 387 |
+
"args": {
|
| 388 |
+
"early_stopping_patience": 3,
|
| 389 |
+
"early_stopping_threshold": 0.0
|
| 390 |
+
},
|
| 391 |
+
"attributes": {
|
| 392 |
+
"early_stopping_patience_counter": 0
|
| 393 |
+
}
|
| 394 |
+
},
|
| 395 |
+
"TrainerControl": {
|
| 396 |
+
"args": {
|
| 397 |
+
"should_epoch_stop": false,
|
| 398 |
+
"should_evaluate": false,
|
| 399 |
+
"should_log": false,
|
| 400 |
+
"should_save": true,
|
| 401 |
+
"should_training_stop": false
|
| 402 |
+
},
|
| 403 |
+
"attributes": {}
|
| 404 |
+
}
|
| 405 |
+
},
|
| 406 |
+
"total_flos": 0.0,
|
| 407 |
+
"train_batch_size": 8,
|
| 408 |
+
"trial_name": null,
|
| 409 |
+
"trial_params": null
|
| 410 |
+
}
|
ood-vifactcheck-phobert-topic-domain-02-s43-20260814T134559Z-bdd2c308/trainer/checkpoint-1092/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-02-s43-20260814T134559Z-bdd2c308/trainer/checkpoint-1638/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-topic-domain-02-s43-20260814T134559Z-bdd2c308/trainer/checkpoint-1638/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-02-s43-20260814T134559Z-bdd2c308/trainer/checkpoint-1638/special_tokens_map.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": "<mask>",
|
| 6 |
+
"pad_token": "<pad>",
|
| 7 |
+
"sep_token": "</s>",
|
| 8 |
+
"unk_token": "<unk>"
|
| 9 |
+
}
|
ood-vifactcheck-phobert-topic-domain-02-s43-20260814T134559Z-bdd2c308/trainer/checkpoint-1638/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"64000": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 256,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "PhobertTokenizer",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-phobert-topic-domain-02-s43-20260814T134559Z-bdd2c308/trainer/checkpoint-1638/trainer_state.json
ADDED
|
@@ -0,0 +1,597 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1092,
|
| 3 |
+
"best_metric": 0.8548617051611064,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-topic-domain-02-s43-20260814T134559Z-bdd2c308/trainer/checkpoint-1092",
|
| 5 |
+
"epoch": 9.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1638,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.13793103448275862,
|
| 14 |
+
"grad_norm": 6.502751350402832,
|
| 15 |
+
"learning_rate": 1.3186813186813187e-06,
|
| 16 |
+
"loss": 4.4294,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.27586206896551724,
|
| 21 |
+
"grad_norm": 6.533897399902344,
|
| 22 |
+
"learning_rate": 2.6923076923076923e-06,
|
| 23 |
+
"loss": 4.4498,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.41379310344827586,
|
| 28 |
+
"grad_norm": 6.485318660736084,
|
| 29 |
+
"learning_rate": 4.065934065934066e-06,
|
| 30 |
+
"loss": 4.4152,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5517241379310345,
|
| 35 |
+
"grad_norm": 7.700263023376465,
|
| 36 |
+
"learning_rate": 5.43956043956044e-06,
|
| 37 |
+
"loss": 4.3639,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6896551724137931,
|
| 42 |
+
"grad_norm": 9.55636215209961,
|
| 43 |
+
"learning_rate": 6.813186813186814e-06,
|
| 44 |
+
"loss": 4.3391,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8275862068965517,
|
| 49 |
+
"grad_norm": 11.730230331420898,
|
| 50 |
+
"learning_rate": 8.186813186813188e-06,
|
| 51 |
+
"loss": 4.2742,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9655172413793104,
|
| 56 |
+
"grad_norm": 10.071218490600586,
|
| 57 |
+
"learning_rate": 9.560439560439562e-06,
|
| 58 |
+
"loss": 4.2261,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.510934393638171,
|
| 64 |
+
"eval_loss": 1.0159550905227661,
|
| 65 |
+
"eval_macro_f1": 0.49256508563077905,
|
| 66 |
+
"eval_runtime": 1.1,
|
| 67 |
+
"eval_samples_per_second": 457.273,
|
| 68 |
+
"eval_steps_per_second": 29.091,
|
| 69 |
+
"eval_weighted_f1": 0.4937639432603952,
|
| 70 |
+
"step": 182
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0993103448275863,
|
| 74 |
+
"grad_norm": 17.43917465209961,
|
| 75 |
+
"learning_rate": 1.0934065934065933e-05,
|
| 76 |
+
"loss": 3.8656,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.237241379310345,
|
| 81 |
+
"grad_norm": 22.632905960083008,
|
| 82 |
+
"learning_rate": 1.230769230769231e-05,
|
| 83 |
+
"loss": 3.6757,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.3751724137931034,
|
| 88 |
+
"grad_norm": 27.157556533813477,
|
| 89 |
+
"learning_rate": 1.3681318681318683e-05,
|
| 90 |
+
"loss": 3.3253,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.513103448275862,
|
| 95 |
+
"grad_norm": 35.907798767089844,
|
| 96 |
+
"learning_rate": 1.5054945054945056e-05,
|
| 97 |
+
"loss": 2.9085,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.6510344827586207,
|
| 102 |
+
"grad_norm": 35.90727996826172,
|
| 103 |
+
"learning_rate": 1.642857142857143e-05,
|
| 104 |
+
"loss": 2.7074,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7889655172413792,
|
| 109 |
+
"grad_norm": 33.298580169677734,
|
| 110 |
+
"learning_rate": 1.78021978021978e-05,
|
| 111 |
+
"loss": 2.2439,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.926896551724138,
|
| 116 |
+
"grad_norm": 32.35009002685547,
|
| 117 |
+
"learning_rate": 1.9175824175824178e-05,
|
| 118 |
+
"loss": 2.1258,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.805168986083499,
|
| 124 |
+
"eval_loss": 0.49421149492263794,
|
| 125 |
+
"eval_macro_f1": 0.80480474458109,
|
| 126 |
+
"eval_runtime": 1.0389,
|
| 127 |
+
"eval_samples_per_second": 484.152,
|
| 128 |
+
"eval_steps_per_second": 30.801,
|
| 129 |
+
"eval_weighted_f1": 0.8051478160810519,
|
| 130 |
+
"step": 364
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.0606896551724136,
|
| 134 |
+
"grad_norm": 26.5976619720459,
|
| 135 |
+
"learning_rate": 1.993894993894994e-05,
|
| 136 |
+
"loss": 2.0294,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.1986206896551725,
|
| 141 |
+
"grad_norm": 60.08498764038086,
|
| 142 |
+
"learning_rate": 1.9786324786324788e-05,
|
| 143 |
+
"loss": 1.6727,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.336551724137931,
|
| 148 |
+
"grad_norm": 45.483428955078125,
|
| 149 |
+
"learning_rate": 1.9633699633699634e-05,
|
| 150 |
+
"loss": 1.7497,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.47448275862069,
|
| 155 |
+
"grad_norm": 56.38203430175781,
|
| 156 |
+
"learning_rate": 1.948107448107448e-05,
|
| 157 |
+
"loss": 1.6546,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.6124137931034483,
|
| 162 |
+
"grad_norm": 43.6146125793457,
|
| 163 |
+
"learning_rate": 1.932844932844933e-05,
|
| 164 |
+
"loss": 1.6076,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.750344827586207,
|
| 169 |
+
"grad_norm": 104.45264434814453,
|
| 170 |
+
"learning_rate": 1.9175824175824178e-05,
|
| 171 |
+
"loss": 1.665,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.8882758620689657,
|
| 176 |
+
"grad_norm": 37.07889175415039,
|
| 177 |
+
"learning_rate": 1.9023199023199025e-05,
|
| 178 |
+
"loss": 1.6418,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 3.0,
|
| 183 |
+
"eval_accuracy": 0.852882703777336,
|
| 184 |
+
"eval_loss": 0.4100320041179657,
|
| 185 |
+
"eval_macro_f1": 0.8531422835626169,
|
| 186 |
+
"eval_runtime": 1.2145,
|
| 187 |
+
"eval_samples_per_second": 414.168,
|
| 188 |
+
"eval_steps_per_second": 26.349,
|
| 189 |
+
"eval_weighted_f1": 0.8531716659680295,
|
| 190 |
+
"step": 546
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.0220689655172412,
|
| 194 |
+
"grad_norm": 30.45008087158203,
|
| 195 |
+
"learning_rate": 1.887057387057387e-05,
|
| 196 |
+
"loss": 1.4073,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.16,
|
| 201 |
+
"grad_norm": 31.528188705444336,
|
| 202 |
+
"learning_rate": 1.8717948717948718e-05,
|
| 203 |
+
"loss": 1.1455,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.2979310344827586,
|
| 208 |
+
"grad_norm": 47.32204818725586,
|
| 209 |
+
"learning_rate": 1.8565323565323568e-05,
|
| 210 |
+
"loss": 1.1364,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.435862068965517,
|
| 215 |
+
"grad_norm": 62.04407501220703,
|
| 216 |
+
"learning_rate": 1.8412698412698415e-05,
|
| 217 |
+
"loss": 1.1754,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.573793103448276,
|
| 222 |
+
"grad_norm": 61.582069396972656,
|
| 223 |
+
"learning_rate": 1.826007326007326e-05,
|
| 224 |
+
"loss": 1.1587,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.7117241379310344,
|
| 229 |
+
"grad_norm": 78.23368072509766,
|
| 230 |
+
"learning_rate": 1.8107448107448108e-05,
|
| 231 |
+
"loss": 1.1347,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.849655172413793,
|
| 236 |
+
"grad_norm": 80.21173095703125,
|
| 237 |
+
"learning_rate": 1.7954822954822958e-05,
|
| 238 |
+
"loss": 1.1009,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.9875862068965517,
|
| 243 |
+
"grad_norm": 37.001888275146484,
|
| 244 |
+
"learning_rate": 1.78021978021978e-05,
|
| 245 |
+
"loss": 0.9838,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 4.0,
|
| 250 |
+
"eval_accuracy": 0.8508946322067594,
|
| 251 |
+
"eval_loss": 0.4349372386932373,
|
| 252 |
+
"eval_macro_f1": 0.8503660235798499,
|
| 253 |
+
"eval_runtime": 1.2065,
|
| 254 |
+
"eval_samples_per_second": 416.902,
|
| 255 |
+
"eval_steps_per_second": 26.523,
|
| 256 |
+
"eval_weighted_f1": 0.8506039358703087,
|
| 257 |
+
"step": 728
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.121379310344827,
|
| 261 |
+
"grad_norm": 8.690960884094238,
|
| 262 |
+
"learning_rate": 1.764957264957265e-05,
|
| 263 |
+
"loss": 0.6978,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.259310344827586,
|
| 268 |
+
"grad_norm": 55.83857727050781,
|
| 269 |
+
"learning_rate": 1.74969474969475e-05,
|
| 270 |
+
"loss": 0.9204,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.397241379310345,
|
| 275 |
+
"grad_norm": 49.880088806152344,
|
| 276 |
+
"learning_rate": 1.7344322344322345e-05,
|
| 277 |
+
"loss": 0.8488,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.535172413793103,
|
| 282 |
+
"grad_norm": 30.68686866760254,
|
| 283 |
+
"learning_rate": 1.7191697191697195e-05,
|
| 284 |
+
"loss": 0.765,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.673103448275862,
|
| 289 |
+
"grad_norm": 173.29953002929688,
|
| 290 |
+
"learning_rate": 1.703907203907204e-05,
|
| 291 |
+
"loss": 0.6741,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.811034482758621,
|
| 296 |
+
"grad_norm": 30.304973602294922,
|
| 297 |
+
"learning_rate": 1.688644688644689e-05,
|
| 298 |
+
"loss": 0.7912,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.94896551724138,
|
| 303 |
+
"grad_norm": 19.50191307067871,
|
| 304 |
+
"learning_rate": 1.6733821733821735e-05,
|
| 305 |
+
"loss": 0.7726,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 5.0,
|
| 310 |
+
"eval_accuracy": 0.8548707753479126,
|
| 311 |
+
"eval_loss": 0.4566419720649719,
|
| 312 |
+
"eval_macro_f1": 0.8532414420320026,
|
| 313 |
+
"eval_runtime": 1.2038,
|
| 314 |
+
"eval_samples_per_second": 417.843,
|
| 315 |
+
"eval_steps_per_second": 26.582,
|
| 316 |
+
"eval_weighted_f1": 0.8533275301022799,
|
| 317 |
+
"step": 910
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.082758620689655,
|
| 321 |
+
"grad_norm": 26.890073776245117,
|
| 322 |
+
"learning_rate": 1.6581196581196585e-05,
|
| 323 |
+
"loss": 0.4669,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.220689655172414,
|
| 328 |
+
"grad_norm": 110.97564697265625,
|
| 329 |
+
"learning_rate": 1.642857142857143e-05,
|
| 330 |
+
"loss": 0.4888,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.358620689655172,
|
| 335 |
+
"grad_norm": 62.88267135620117,
|
| 336 |
+
"learning_rate": 1.627594627594628e-05,
|
| 337 |
+
"loss": 0.6254,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.496551724137931,
|
| 342 |
+
"grad_norm": 2.0096068382263184,
|
| 343 |
+
"learning_rate": 1.6123321123321125e-05,
|
| 344 |
+
"loss": 0.5119,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.63448275862069,
|
| 349 |
+
"grad_norm": 18.092445373535156,
|
| 350 |
+
"learning_rate": 1.5970695970695972e-05,
|
| 351 |
+
"loss": 0.5095,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.772413793103448,
|
| 356 |
+
"grad_norm": 27.10129165649414,
|
| 357 |
+
"learning_rate": 1.581807081807082e-05,
|
| 358 |
+
"loss": 0.4641,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.910344827586207,
|
| 363 |
+
"grad_norm": 66.32721710205078,
|
| 364 |
+
"learning_rate": 1.5665445665445665e-05,
|
| 365 |
+
"loss": 0.7483,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 6.0,
|
| 370 |
+
"eval_accuracy": 0.8548707753479126,
|
| 371 |
+
"eval_loss": 0.5105668306350708,
|
| 372 |
+
"eval_macro_f1": 0.8548617051611064,
|
| 373 |
+
"eval_runtime": 1.1636,
|
| 374 |
+
"eval_samples_per_second": 432.262,
|
| 375 |
+
"eval_steps_per_second": 27.5,
|
| 376 |
+
"eval_weighted_f1": 0.8548737515029583,
|
| 377 |
+
"step": 1092
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 6.0441379310344825,
|
| 381 |
+
"grad_norm": 74.94464874267578,
|
| 382 |
+
"learning_rate": 1.5512820512820516e-05,
|
| 383 |
+
"loss": 0.4408,
|
| 384 |
+
"step": 1100
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 6.182068965517241,
|
| 388 |
+
"grad_norm": 32.69307327270508,
|
| 389 |
+
"learning_rate": 1.5360195360195362e-05,
|
| 390 |
+
"loss": 0.437,
|
| 391 |
+
"step": 1125
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 6.32,
|
| 395 |
+
"grad_norm": 4.236227512359619,
|
| 396 |
+
"learning_rate": 1.5207570207570207e-05,
|
| 397 |
+
"loss": 0.2534,
|
| 398 |
+
"step": 1150
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.457931034482758,
|
| 402 |
+
"grad_norm": 0.9062155485153198,
|
| 403 |
+
"learning_rate": 1.5054945054945056e-05,
|
| 404 |
+
"loss": 0.3229,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.595862068965517,
|
| 409 |
+
"grad_norm": 9.682204246520996,
|
| 410 |
+
"learning_rate": 1.4902319902319904e-05,
|
| 411 |
+
"loss": 0.31,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.733793103448276,
|
| 416 |
+
"grad_norm": 106.4243392944336,
|
| 417 |
+
"learning_rate": 1.4749694749694752e-05,
|
| 418 |
+
"loss": 0.6466,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.871724137931034,
|
| 423 |
+
"grad_norm": 79.56591033935547,
|
| 424 |
+
"learning_rate": 1.4597069597069597e-05,
|
| 425 |
+
"loss": 0.6479,
|
| 426 |
+
"step": 1250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 7.0,
|
| 430 |
+
"eval_accuracy": 0.8469184890656064,
|
| 431 |
+
"eval_loss": 0.6670135855674744,
|
| 432 |
+
"eval_macro_f1": 0.8469041498524125,
|
| 433 |
+
"eval_runtime": 1.2602,
|
| 434 |
+
"eval_samples_per_second": 399.145,
|
| 435 |
+
"eval_steps_per_second": 25.393,
|
| 436 |
+
"eval_weighted_f1": 0.8468221393613329,
|
| 437 |
+
"step": 1274
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 7.0055172413793105,
|
| 441 |
+
"grad_norm": 403.63482666015625,
|
| 442 |
+
"learning_rate": 1.4444444444444446e-05,
|
| 443 |
+
"loss": 0.4999,
|
| 444 |
+
"step": 1275
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 7.1434482758620685,
|
| 448 |
+
"grad_norm": 27.240480422973633,
|
| 449 |
+
"learning_rate": 1.4291819291819293e-05,
|
| 450 |
+
"loss": 0.4513,
|
| 451 |
+
"step": 1300
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 7.281379310344827,
|
| 455 |
+
"grad_norm": 1.986925721168518,
|
| 456 |
+
"learning_rate": 1.4139194139194141e-05,
|
| 457 |
+
"loss": 0.4293,
|
| 458 |
+
"step": 1325
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 7.419310344827586,
|
| 462 |
+
"grad_norm": 0.271933913230896,
|
| 463 |
+
"learning_rate": 1.3986568986568988e-05,
|
| 464 |
+
"loss": 0.2505,
|
| 465 |
+
"step": 1350
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 7.557241379310344,
|
| 469 |
+
"grad_norm": 0.33632025122642517,
|
| 470 |
+
"learning_rate": 1.3833943833943834e-05,
|
| 471 |
+
"loss": 0.2655,
|
| 472 |
+
"step": 1375
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 7.695172413793103,
|
| 476 |
+
"grad_norm": 69.18501281738281,
|
| 477 |
+
"learning_rate": 1.3681318681318683e-05,
|
| 478 |
+
"loss": 0.2353,
|
| 479 |
+
"step": 1400
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 7.833103448275862,
|
| 483 |
+
"grad_norm": 32.5136604309082,
|
| 484 |
+
"learning_rate": 1.3528693528693531e-05,
|
| 485 |
+
"loss": 0.3636,
|
| 486 |
+
"step": 1425
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 7.971034482758621,
|
| 490 |
+
"grad_norm": 0.32109498977661133,
|
| 491 |
+
"learning_rate": 1.3376068376068376e-05,
|
| 492 |
+
"loss": 0.3245,
|
| 493 |
+
"step": 1450
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 8.0,
|
| 497 |
+
"eval_accuracy": 0.8508946322067594,
|
| 498 |
+
"eval_loss": 0.7320056557655334,
|
| 499 |
+
"eval_macro_f1": 0.8498263972070234,
|
| 500 |
+
"eval_runtime": 1.1761,
|
| 501 |
+
"eval_samples_per_second": 427.692,
|
| 502 |
+
"eval_steps_per_second": 27.209,
|
| 503 |
+
"eval_weighted_f1": 0.849861939421347,
|
| 504 |
+
"step": 1456
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 8.104827586206897,
|
| 508 |
+
"grad_norm": 176.09683227539062,
|
| 509 |
+
"learning_rate": 1.3223443223443224e-05,
|
| 510 |
+
"loss": 0.4312,
|
| 511 |
+
"step": 1475
|
| 512 |
+
},
|
| 513 |
+
{
|
| 514 |
+
"epoch": 8.242758620689655,
|
| 515 |
+
"grad_norm": 2.4117465019226074,
|
| 516 |
+
"learning_rate": 1.3070818070818073e-05,
|
| 517 |
+
"loss": 0.3072,
|
| 518 |
+
"step": 1500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 8.380689655172414,
|
| 522 |
+
"grad_norm": 1.4015716314315796,
|
| 523 |
+
"learning_rate": 1.291819291819292e-05,
|
| 524 |
+
"loss": 0.3011,
|
| 525 |
+
"step": 1525
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 8.518620689655172,
|
| 529 |
+
"grad_norm": 1.6123242378234863,
|
| 530 |
+
"learning_rate": 1.2765567765567766e-05,
|
| 531 |
+
"loss": 0.3385,
|
| 532 |
+
"step": 1550
|
| 533 |
+
},
|
| 534 |
+
{
|
| 535 |
+
"epoch": 8.65655172413793,
|
| 536 |
+
"grad_norm": 27.9930419921875,
|
| 537 |
+
"learning_rate": 1.2612942612942613e-05,
|
| 538 |
+
"loss": 0.1829,
|
| 539 |
+
"step": 1575
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 8.79448275862069,
|
| 543 |
+
"grad_norm": 135.72134399414062,
|
| 544 |
+
"learning_rate": 1.2460317460317461e-05,
|
| 545 |
+
"loss": 0.3654,
|
| 546 |
+
"step": 1600
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 8.932413793103448,
|
| 550 |
+
"grad_norm": 0.1962214708328247,
|
| 551 |
+
"learning_rate": 1.230769230769231e-05,
|
| 552 |
+
"loss": 0.3426,
|
| 553 |
+
"step": 1625
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 9.0,
|
| 557 |
+
"eval_accuracy": 0.8449304174950298,
|
| 558 |
+
"eval_loss": 0.7963999509811401,
|
| 559 |
+
"eval_macro_f1": 0.8415480270703521,
|
| 560 |
+
"eval_runtime": 1.1543,
|
| 561 |
+
"eval_samples_per_second": 435.766,
|
| 562 |
+
"eval_steps_per_second": 27.723,
|
| 563 |
+
"eval_weighted_f1": 0.8419117687081752,
|
| 564 |
+
"step": 1638
|
| 565 |
+
}
|
| 566 |
+
],
|
| 567 |
+
"logging_steps": 25,
|
| 568 |
+
"max_steps": 3640,
|
| 569 |
+
"num_input_tokens_seen": 0,
|
| 570 |
+
"num_train_epochs": 20,
|
| 571 |
+
"save_steps": 500,
|
| 572 |
+
"stateful_callbacks": {
|
| 573 |
+
"EarlyStoppingCallback": {
|
| 574 |
+
"args": {
|
| 575 |
+
"early_stopping_patience": 3,
|
| 576 |
+
"early_stopping_threshold": 0.0
|
| 577 |
+
},
|
| 578 |
+
"attributes": {
|
| 579 |
+
"early_stopping_patience_counter": 3
|
| 580 |
+
}
|
| 581 |
+
},
|
| 582 |
+
"TrainerControl": {
|
| 583 |
+
"args": {
|
| 584 |
+
"should_epoch_stop": false,
|
| 585 |
+
"should_evaluate": false,
|
| 586 |
+
"should_log": false,
|
| 587 |
+
"should_save": true,
|
| 588 |
+
"should_training_stop": true
|
| 589 |
+
},
|
| 590 |
+
"attributes": {}
|
| 591 |
+
}
|
| 592 |
+
},
|
| 593 |
+
"total_flos": 0.0,
|
| 594 |
+
"train_batch_size": 8,
|
| 595 |
+
"trial_name": null,
|
| 596 |
+
"trial_params": null
|
| 597 |
+
}
|
ood-vifactcheck-phobert-topic-domain-02-s43-20260814T134559Z-bdd2c308/trainer/checkpoint-1638/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-03-s43-20260814T135955Z-bdd2c308/trainer/checkpoint-1755/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-topic-domain-03-s43-20260814T135955Z-bdd2c308/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-s43-20260814T135955Z-bdd2c308/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-s43-20260814T135955Z-bdd2c308/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-s43-20260814T135955Z-bdd2c308/trainer/checkpoint-1755/trainer_state.json
ADDED
|
@@ -0,0 +1,632 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1755,
|
| 3 |
+
"best_metric": 0.8630274914204891,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-topic-domain-03-s43-20260814T135955Z-bdd2c308/trainer/checkpoint-1755",
|
| 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": 9.472962379455566,
|
| 15 |
+
"learning_rate": 1.230769230769231e-06,
|
| 16 |
+
"loss": 4.4379,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.25673940949935814,
|
| 21 |
+
"grad_norm": 8.352157592773438,
|
| 22 |
+
"learning_rate": 2.512820512820513e-06,
|
| 23 |
+
"loss": 4.4427,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.3851091142490372,
|
| 28 |
+
"grad_norm": 8.57148551940918,
|
| 29 |
+
"learning_rate": 3.794871794871795e-06,
|
| 30 |
+
"loss": 4.4149,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5134788189987163,
|
| 35 |
+
"grad_norm": 7.510636329650879,
|
| 36 |
+
"learning_rate": 5.076923076923077e-06,
|
| 37 |
+
"loss": 4.3645,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6418485237483954,
|
| 42 |
+
"grad_norm": 10.156326293945312,
|
| 43 |
+
"learning_rate": 6.358974358974359e-06,
|
| 44 |
+
"loss": 4.3624,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.7702182284980744,
|
| 49 |
+
"grad_norm": 7.181392669677734,
|
| 50 |
+
"learning_rate": 7.641025641025641e-06,
|
| 51 |
+
"loss": 4.3311,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.8985879332477535,
|
| 56 |
+
"grad_norm": 10.214773178100586,
|
| 57 |
+
"learning_rate": 8.923076923076925e-06,
|
| 58 |
+
"loss": 4.251,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.5346020761245674,
|
| 64 |
+
"eval_loss": 0.9750317931175232,
|
| 65 |
+
"eval_macro_f1": 0.4885035020153532,
|
| 66 |
+
"eval_runtime": 1.2373,
|
| 67 |
+
"eval_samples_per_second": 467.15,
|
| 68 |
+
"eval_steps_per_second": 29.904,
|
| 69 |
+
"eval_weighted_f1": 0.4916625203693561,
|
| 70 |
+
"step": 195
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0256739409499358,
|
| 74 |
+
"grad_norm": 22.54351234436035,
|
| 75 |
+
"learning_rate": 1.0205128205128205e-05,
|
| 76 |
+
"loss": 4.0429,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.1540436456996148,
|
| 81 |
+
"grad_norm": 20.747333526611328,
|
| 82 |
+
"learning_rate": 1.1487179487179487e-05,
|
| 83 |
+
"loss": 3.8301,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.282413350449294,
|
| 88 |
+
"grad_norm": 28.735864639282227,
|
| 89 |
+
"learning_rate": 1.2769230769230769e-05,
|
| 90 |
+
"loss": 3.5419,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.410783055198973,
|
| 95 |
+
"grad_norm": 36.93158721923828,
|
| 96 |
+
"learning_rate": 1.405128205128205e-05,
|
| 97 |
+
"loss": 3.2036,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.539152759948652,
|
| 102 |
+
"grad_norm": 54.45057678222656,
|
| 103 |
+
"learning_rate": 1.5333333333333334e-05,
|
| 104 |
+
"loss": 2.4884,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.6675224646983313,
|
| 109 |
+
"grad_norm": 49.76697540283203,
|
| 110 |
+
"learning_rate": 1.6615384615384618e-05,
|
| 111 |
+
"loss": 2.6189,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.7958921694480101,
|
| 116 |
+
"grad_norm": 50.385623931884766,
|
| 117 |
+
"learning_rate": 1.78974358974359e-05,
|
| 118 |
+
"loss": 2.2054,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 1.9242618741976893,
|
| 123 |
+
"grad_norm": 44.956642150878906,
|
| 124 |
+
"learning_rate": 1.9179487179487182e-05,
|
| 125 |
+
"loss": 2.1112,
|
| 126 |
+
"step": 375
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"epoch": 2.0,
|
| 130 |
+
"eval_accuracy": 0.8200692041522492,
|
| 131 |
+
"eval_loss": 0.469224750995636,
|
| 132 |
+
"eval_macro_f1": 0.8198326682862765,
|
| 133 |
+
"eval_runtime": 1.2303,
|
| 134 |
+
"eval_samples_per_second": 469.792,
|
| 135 |
+
"eval_steps_per_second": 30.073,
|
| 136 |
+
"eval_weighted_f1": 0.8199733551426199,
|
| 137 |
+
"step": 390
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.0513478818998716,
|
| 141 |
+
"grad_norm": 95.71858978271484,
|
| 142 |
+
"learning_rate": 1.994871794871795e-05,
|
| 143 |
+
"loss": 2.0482,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.179717586649551,
|
| 148 |
+
"grad_norm": 46.47174072265625,
|
| 149 |
+
"learning_rate": 1.9806267806267806e-05,
|
| 150 |
+
"loss": 1.715,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.3080872913992296,
|
| 155 |
+
"grad_norm": 48.007205963134766,
|
| 156 |
+
"learning_rate": 1.9663817663817667e-05,
|
| 157 |
+
"loss": 1.6391,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.436456996148909,
|
| 162 |
+
"grad_norm": 50.561866760253906,
|
| 163 |
+
"learning_rate": 1.952136752136752e-05,
|
| 164 |
+
"loss": 1.7053,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.564826700898588,
|
| 169 |
+
"grad_norm": 23.58455467224121,
|
| 170 |
+
"learning_rate": 1.9378917378917382e-05,
|
| 171 |
+
"loss": 1.5795,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.693196405648267,
|
| 176 |
+
"grad_norm": 17.744104385375977,
|
| 177 |
+
"learning_rate": 1.923646723646724e-05,
|
| 178 |
+
"loss": 1.7004,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 2.821566110397946,
|
| 183 |
+
"grad_norm": 79.87342834472656,
|
| 184 |
+
"learning_rate": 1.9094017094017097e-05,
|
| 185 |
+
"loss": 1.6293,
|
| 186 |
+
"step": 550
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 2.949935815147625,
|
| 190 |
+
"grad_norm": 42.42940139770508,
|
| 191 |
+
"learning_rate": 1.8951566951566954e-05,
|
| 192 |
+
"loss": 1.4183,
|
| 193 |
+
"step": 575
|
| 194 |
+
},
|
| 195 |
+
{
|
| 196 |
+
"epoch": 3.0,
|
| 197 |
+
"eval_accuracy": 0.842560553633218,
|
| 198 |
+
"eval_loss": 0.438308447599411,
|
| 199 |
+
"eval_macro_f1": 0.8416056571104035,
|
| 200 |
+
"eval_runtime": 1.2573,
|
| 201 |
+
"eval_samples_per_second": 459.725,
|
| 202 |
+
"eval_steps_per_second": 29.429,
|
| 203 |
+
"eval_weighted_f1": 0.8418449527499576,
|
| 204 |
+
"step": 585
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.0770218228498076,
|
| 208 |
+
"grad_norm": 69.78426361083984,
|
| 209 |
+
"learning_rate": 1.880911680911681e-05,
|
| 210 |
+
"loss": 1.3933,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.2053915275994864,
|
| 215 |
+
"grad_norm": 47.075809478759766,
|
| 216 |
+
"learning_rate": 1.866666666666667e-05,
|
| 217 |
+
"loss": 0.9495,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.3337612323491657,
|
| 222 |
+
"grad_norm": 35.8579216003418,
|
| 223 |
+
"learning_rate": 1.8524216524216526e-05,
|
| 224 |
+
"loss": 0.9142,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.462130937098845,
|
| 229 |
+
"grad_norm": 20.10189437866211,
|
| 230 |
+
"learning_rate": 1.8381766381766383e-05,
|
| 231 |
+
"loss": 1.0526,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.5905006418485237,
|
| 236 |
+
"grad_norm": 17.218551635742188,
|
| 237 |
+
"learning_rate": 1.823931623931624e-05,
|
| 238 |
+
"loss": 1.0184,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.718870346598203,
|
| 243 |
+
"grad_norm": 41.16714859008789,
|
| 244 |
+
"learning_rate": 1.8096866096866098e-05,
|
| 245 |
+
"loss": 0.9362,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 3.8472400513478817,
|
| 250 |
+
"grad_norm": 59.61590576171875,
|
| 251 |
+
"learning_rate": 1.7954415954415955e-05,
|
| 252 |
+
"loss": 1.1327,
|
| 253 |
+
"step": 750
|
| 254 |
+
},
|
| 255 |
+
{
|
| 256 |
+
"epoch": 3.975609756097561,
|
| 257 |
+
"grad_norm": 32.190032958984375,
|
| 258 |
+
"learning_rate": 1.7811965811965813e-05,
|
| 259 |
+
"loss": 1.2118,
|
| 260 |
+
"step": 775
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 4.0,
|
| 264 |
+
"eval_accuracy": 0.8460207612456747,
|
| 265 |
+
"eval_loss": 0.479248046875,
|
| 266 |
+
"eval_macro_f1": 0.8461434530745976,
|
| 267 |
+
"eval_runtime": 1.2085,
|
| 268 |
+
"eval_samples_per_second": 478.271,
|
| 269 |
+
"eval_steps_per_second": 30.616,
|
| 270 |
+
"eval_weighted_f1": 0.8462817187889256,
|
| 271 |
+
"step": 780
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.102695763799743,
|
| 275 |
+
"grad_norm": 7.671431064605713,
|
| 276 |
+
"learning_rate": 1.766951566951567e-05,
|
| 277 |
+
"loss": 0.7384,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.2310654685494224,
|
| 282 |
+
"grad_norm": 39.89632034301758,
|
| 283 |
+
"learning_rate": 1.7527065527065528e-05,
|
| 284 |
+
"loss": 0.7164,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.359435173299102,
|
| 289 |
+
"grad_norm": 44.15858840942383,
|
| 290 |
+
"learning_rate": 1.7384615384615385e-05,
|
| 291 |
+
"loss": 0.9587,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.487804878048781,
|
| 296 |
+
"grad_norm": 90.62368774414062,
|
| 297 |
+
"learning_rate": 1.7242165242165242e-05,
|
| 298 |
+
"loss": 0.6554,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.616174582798459,
|
| 303 |
+
"grad_norm": 1.3872368335723877,
|
| 304 |
+
"learning_rate": 1.7099715099715103e-05,
|
| 305 |
+
"loss": 0.681,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 4.7445442875481385,
|
| 310 |
+
"grad_norm": 6.195230484008789,
|
| 311 |
+
"learning_rate": 1.6957264957264957e-05,
|
| 312 |
+
"loss": 0.6918,
|
| 313 |
+
"step": 925
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 4.872913992297818,
|
| 317 |
+
"grad_norm": 147.07801818847656,
|
| 318 |
+
"learning_rate": 1.6814814814814818e-05,
|
| 319 |
+
"loss": 0.5796,
|
| 320 |
+
"step": 950
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 5.0,
|
| 324 |
+
"grad_norm": 14.160751342773438,
|
| 325 |
+
"learning_rate": 1.6672364672364675e-05,
|
| 326 |
+
"loss": 0.8333,
|
| 327 |
+
"step": 975
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 5.0,
|
| 331 |
+
"eval_accuracy": 0.8391003460207612,
|
| 332 |
+
"eval_loss": 0.5641457438468933,
|
| 333 |
+
"eval_macro_f1": 0.8396643546127365,
|
| 334 |
+
"eval_runtime": 1.2612,
|
| 335 |
+
"eval_samples_per_second": 458.311,
|
| 336 |
+
"eval_steps_per_second": 29.338,
|
| 337 |
+
"eval_weighted_f1": 0.8397169728153938,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.128369704749679,
|
| 342 |
+
"grad_norm": 35.67943572998047,
|
| 343 |
+
"learning_rate": 1.652991452991453e-05,
|
| 344 |
+
"loss": 0.5692,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.2567394094993585,
|
| 349 |
+
"grad_norm": 2.249913215637207,
|
| 350 |
+
"learning_rate": 1.638746438746439e-05,
|
| 351 |
+
"loss": 0.4787,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.385109114249037,
|
| 356 |
+
"grad_norm": 48.28496170043945,
|
| 357 |
+
"learning_rate": 1.6245014245014247e-05,
|
| 358 |
+
"loss": 0.4725,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.513478818998716,
|
| 363 |
+
"grad_norm": 2.8437349796295166,
|
| 364 |
+
"learning_rate": 1.6102564102564105e-05,
|
| 365 |
+
"loss": 0.5654,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 5.641848523748395,
|
| 370 |
+
"grad_norm": 40.56161117553711,
|
| 371 |
+
"learning_rate": 1.5960113960113962e-05,
|
| 372 |
+
"loss": 0.6465,
|
| 373 |
+
"step": 1100
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 5.7702182284980745,
|
| 377 |
+
"grad_norm": 7.292684078216553,
|
| 378 |
+
"learning_rate": 1.581766381766382e-05,
|
| 379 |
+
"loss": 0.6451,
|
| 380 |
+
"step": 1125
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 5.898587933247754,
|
| 384 |
+
"grad_norm": 459.4928283691406,
|
| 385 |
+
"learning_rate": 1.5675213675213677e-05,
|
| 386 |
+
"loss": 0.4834,
|
| 387 |
+
"step": 1150
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 6.0,
|
| 391 |
+
"eval_accuracy": 0.8529411764705882,
|
| 392 |
+
"eval_loss": 0.5885111093521118,
|
| 393 |
+
"eval_macro_f1": 0.8524204646306593,
|
| 394 |
+
"eval_runtime": 1.3233,
|
| 395 |
+
"eval_samples_per_second": 436.776,
|
| 396 |
+
"eval_steps_per_second": 27.96,
|
| 397 |
+
"eval_weighted_f1": 0.8525879456553224,
|
| 398 |
+
"step": 1170
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.025673940949936,
|
| 402 |
+
"grad_norm": 24.552349090576172,
|
| 403 |
+
"learning_rate": 1.5532763532763534e-05,
|
| 404 |
+
"loss": 0.5777,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.154043645699615,
|
| 409 |
+
"grad_norm": 66.84564971923828,
|
| 410 |
+
"learning_rate": 1.539031339031339e-05,
|
| 411 |
+
"loss": 0.3109,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.282413350449294,
|
| 416 |
+
"grad_norm": 19.659093856811523,
|
| 417 |
+
"learning_rate": 1.5247863247863249e-05,
|
| 418 |
+
"loss": 0.5152,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.410783055198973,
|
| 423 |
+
"grad_norm": 7.469658851623535,
|
| 424 |
+
"learning_rate": 1.5105413105413106e-05,
|
| 425 |
+
"loss": 0.3627,
|
| 426 |
+
"step": 1250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 6.539152759948652,
|
| 430 |
+
"grad_norm": 144.7767791748047,
|
| 431 |
+
"learning_rate": 1.4962962962962964e-05,
|
| 432 |
+
"loss": 0.4557,
|
| 433 |
+
"step": 1275
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 6.667522464698331,
|
| 437 |
+
"grad_norm": 1.6089030504226685,
|
| 438 |
+
"learning_rate": 1.4820512820512823e-05,
|
| 439 |
+
"loss": 0.3658,
|
| 440 |
+
"step": 1300
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 6.7958921694480106,
|
| 444 |
+
"grad_norm": 65.19385528564453,
|
| 445 |
+
"learning_rate": 1.4678062678062678e-05,
|
| 446 |
+
"loss": 0.3657,
|
| 447 |
+
"step": 1325
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 6.92426187419769,
|
| 451 |
+
"grad_norm": 6.4792866706848145,
|
| 452 |
+
"learning_rate": 1.4535612535612537e-05,
|
| 453 |
+
"loss": 0.4717,
|
| 454 |
+
"step": 1350
|
| 455 |
+
},
|
| 456 |
+
{
|
| 457 |
+
"epoch": 7.0,
|
| 458 |
+
"eval_accuracy": 0.8442906574394463,
|
| 459 |
+
"eval_loss": 0.7064892053604126,
|
| 460 |
+
"eval_macro_f1": 0.8432133689559431,
|
| 461 |
+
"eval_runtime": 1.2693,
|
| 462 |
+
"eval_samples_per_second": 455.358,
|
| 463 |
+
"eval_steps_per_second": 29.149,
|
| 464 |
+
"eval_weighted_f1": 0.8435507614033771,
|
| 465 |
+
"step": 1365
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 7.051347881899872,
|
| 469 |
+
"grad_norm": 34.33955764770508,
|
| 470 |
+
"learning_rate": 1.4393162393162395e-05,
|
| 471 |
+
"loss": 0.3895,
|
| 472 |
+
"step": 1375
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 7.17971758664955,
|
| 476 |
+
"grad_norm": 60.092464447021484,
|
| 477 |
+
"learning_rate": 1.425071225071225e-05,
|
| 478 |
+
"loss": 0.2466,
|
| 479 |
+
"step": 1400
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 7.30808729139923,
|
| 483 |
+
"grad_norm": 43.983917236328125,
|
| 484 |
+
"learning_rate": 1.410826210826211e-05,
|
| 485 |
+
"loss": 0.3693,
|
| 486 |
+
"step": 1425
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 7.436456996148909,
|
| 490 |
+
"grad_norm": 6.01005220413208,
|
| 491 |
+
"learning_rate": 1.3965811965811967e-05,
|
| 492 |
+
"loss": 0.3067,
|
| 493 |
+
"step": 1450
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 7.564826700898588,
|
| 497 |
+
"grad_norm": 0.34186023473739624,
|
| 498 |
+
"learning_rate": 1.3823361823361824e-05,
|
| 499 |
+
"loss": 0.4126,
|
| 500 |
+
"step": 1475
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 7.693196405648267,
|
| 504 |
+
"grad_norm": 15.058670997619629,
|
| 505 |
+
"learning_rate": 1.3680911680911682e-05,
|
| 506 |
+
"loss": 0.4612,
|
| 507 |
+
"step": 1500
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 7.821566110397946,
|
| 511 |
+
"grad_norm": 0.2911962866783142,
|
| 512 |
+
"learning_rate": 1.353846153846154e-05,
|
| 513 |
+
"loss": 0.4724,
|
| 514 |
+
"step": 1525
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 7.949935815147625,
|
| 518 |
+
"grad_norm": 3.1831634044647217,
|
| 519 |
+
"learning_rate": 1.3396011396011396e-05,
|
| 520 |
+
"loss": 0.2389,
|
| 521 |
+
"step": 1550
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 8.0,
|
| 525 |
+
"eval_accuracy": 0.8598615916955017,
|
| 526 |
+
"eval_loss": 0.6812170743942261,
|
| 527 |
+
"eval_macro_f1": 0.8590503728747493,
|
| 528 |
+
"eval_runtime": 1.2245,
|
| 529 |
+
"eval_samples_per_second": 472.016,
|
| 530 |
+
"eval_steps_per_second": 30.216,
|
| 531 |
+
"eval_weighted_f1": 0.8592997656967396,
|
| 532 |
+
"step": 1560
|
| 533 |
+
},
|
| 534 |
+
{
|
| 535 |
+
"epoch": 8.077021822849808,
|
| 536 |
+
"grad_norm": 0.26451513171195984,
|
| 537 |
+
"learning_rate": 1.3253561253561255e-05,
|
| 538 |
+
"loss": 0.2438,
|
| 539 |
+
"step": 1575
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 8.205391527599486,
|
| 543 |
+
"grad_norm": 0.2473522275686264,
|
| 544 |
+
"learning_rate": 1.3111111111111113e-05,
|
| 545 |
+
"loss": 0.2306,
|
| 546 |
+
"step": 1600
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 8.333761232349165,
|
| 550 |
+
"grad_norm": 0.45871150493621826,
|
| 551 |
+
"learning_rate": 1.2968660968660968e-05,
|
| 552 |
+
"loss": 0.2556,
|
| 553 |
+
"step": 1625
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 8.462130937098845,
|
| 557 |
+
"grad_norm": 80.5791015625,
|
| 558 |
+
"learning_rate": 1.2826210826210828e-05,
|
| 559 |
+
"loss": 0.2882,
|
| 560 |
+
"step": 1650
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 8.590500641848523,
|
| 564 |
+
"grad_norm": 1.3554444313049316,
|
| 565 |
+
"learning_rate": 1.2683760683760685e-05,
|
| 566 |
+
"loss": 0.1983,
|
| 567 |
+
"step": 1675
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 8.718870346598203,
|
| 571 |
+
"grad_norm": 0.2086154818534851,
|
| 572 |
+
"learning_rate": 1.2541310541310542e-05,
|
| 573 |
+
"loss": 0.2711,
|
| 574 |
+
"step": 1700
|
| 575 |
+
},
|
| 576 |
+
{
|
| 577 |
+
"epoch": 8.847240051347882,
|
| 578 |
+
"grad_norm": 0.23722316324710846,
|
| 579 |
+
"learning_rate": 1.23988603988604e-05,
|
| 580 |
+
"loss": 0.3298,
|
| 581 |
+
"step": 1725
|
| 582 |
+
},
|
| 583 |
+
{
|
| 584 |
+
"epoch": 8.975609756097562,
|
| 585 |
+
"grad_norm": 0.2185768038034439,
|
| 586 |
+
"learning_rate": 1.2256410256410259e-05,
|
| 587 |
+
"loss": 0.2352,
|
| 588 |
+
"step": 1750
|
| 589 |
+
},
|
| 590 |
+
{
|
| 591 |
+
"epoch": 9.0,
|
| 592 |
+
"eval_accuracy": 0.8633217993079585,
|
| 593 |
+
"eval_loss": 0.7458310127258301,
|
| 594 |
+
"eval_macro_f1": 0.8630274914204891,
|
| 595 |
+
"eval_runtime": 1.248,
|
| 596 |
+
"eval_samples_per_second": 463.139,
|
| 597 |
+
"eval_steps_per_second": 29.647,
|
| 598 |
+
"eval_weighted_f1": 0.8629193015186526,
|
| 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": 0
|
| 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": false
|
| 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-s43-20260814T135955Z-bdd2c308/trainer/checkpoint-1755/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-03-s43-20260814T135955Z-bdd2c308/trainer/checkpoint-2340/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-topic-domain-03-s43-20260814T135955Z-bdd2c308/trainer/checkpoint-2340/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-03-s43-20260814T135955Z-bdd2c308/trainer/checkpoint-2340/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-s43-20260814T135955Z-bdd2c308/trainer/checkpoint-2340/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-s43-20260814T135955Z-bdd2c308/trainer/checkpoint-2340/trainer_state.json
ADDED
|
@@ -0,0 +1,826 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1755,
|
| 3 |
+
"best_metric": 0.8630274914204891,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-topic-domain-03-s43-20260814T135955Z-bdd2c308/trainer/checkpoint-1755",
|
| 5 |
+
"epoch": 12.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 2340,
|
| 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": 9.472962379455566,
|
| 15 |
+
"learning_rate": 1.230769230769231e-06,
|
| 16 |
+
"loss": 4.4379,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.25673940949935814,
|
| 21 |
+
"grad_norm": 8.352157592773438,
|
| 22 |
+
"learning_rate": 2.512820512820513e-06,
|
| 23 |
+
"loss": 4.4427,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.3851091142490372,
|
| 28 |
+
"grad_norm": 8.57148551940918,
|
| 29 |
+
"learning_rate": 3.794871794871795e-06,
|
| 30 |
+
"loss": 4.4149,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5134788189987163,
|
| 35 |
+
"grad_norm": 7.510636329650879,
|
| 36 |
+
"learning_rate": 5.076923076923077e-06,
|
| 37 |
+
"loss": 4.3645,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6418485237483954,
|
| 42 |
+
"grad_norm": 10.156326293945312,
|
| 43 |
+
"learning_rate": 6.358974358974359e-06,
|
| 44 |
+
"loss": 4.3624,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.7702182284980744,
|
| 49 |
+
"grad_norm": 7.181392669677734,
|
| 50 |
+
"learning_rate": 7.641025641025641e-06,
|
| 51 |
+
"loss": 4.3311,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.8985879332477535,
|
| 56 |
+
"grad_norm": 10.214773178100586,
|
| 57 |
+
"learning_rate": 8.923076923076925e-06,
|
| 58 |
+
"loss": 4.251,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.5346020761245674,
|
| 64 |
+
"eval_loss": 0.9750317931175232,
|
| 65 |
+
"eval_macro_f1": 0.4885035020153532,
|
| 66 |
+
"eval_runtime": 1.2373,
|
| 67 |
+
"eval_samples_per_second": 467.15,
|
| 68 |
+
"eval_steps_per_second": 29.904,
|
| 69 |
+
"eval_weighted_f1": 0.4916625203693561,
|
| 70 |
+
"step": 195
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0256739409499358,
|
| 74 |
+
"grad_norm": 22.54351234436035,
|
| 75 |
+
"learning_rate": 1.0205128205128205e-05,
|
| 76 |
+
"loss": 4.0429,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.1540436456996148,
|
| 81 |
+
"grad_norm": 20.747333526611328,
|
| 82 |
+
"learning_rate": 1.1487179487179487e-05,
|
| 83 |
+
"loss": 3.8301,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.282413350449294,
|
| 88 |
+
"grad_norm": 28.735864639282227,
|
| 89 |
+
"learning_rate": 1.2769230769230769e-05,
|
| 90 |
+
"loss": 3.5419,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.410783055198973,
|
| 95 |
+
"grad_norm": 36.93158721923828,
|
| 96 |
+
"learning_rate": 1.405128205128205e-05,
|
| 97 |
+
"loss": 3.2036,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.539152759948652,
|
| 102 |
+
"grad_norm": 54.45057678222656,
|
| 103 |
+
"learning_rate": 1.5333333333333334e-05,
|
| 104 |
+
"loss": 2.4884,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.6675224646983313,
|
| 109 |
+
"grad_norm": 49.76697540283203,
|
| 110 |
+
"learning_rate": 1.6615384615384618e-05,
|
| 111 |
+
"loss": 2.6189,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.7958921694480101,
|
| 116 |
+
"grad_norm": 50.385623931884766,
|
| 117 |
+
"learning_rate": 1.78974358974359e-05,
|
| 118 |
+
"loss": 2.2054,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 1.9242618741976893,
|
| 123 |
+
"grad_norm": 44.956642150878906,
|
| 124 |
+
"learning_rate": 1.9179487179487182e-05,
|
| 125 |
+
"loss": 2.1112,
|
| 126 |
+
"step": 375
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"epoch": 2.0,
|
| 130 |
+
"eval_accuracy": 0.8200692041522492,
|
| 131 |
+
"eval_loss": 0.469224750995636,
|
| 132 |
+
"eval_macro_f1": 0.8198326682862765,
|
| 133 |
+
"eval_runtime": 1.2303,
|
| 134 |
+
"eval_samples_per_second": 469.792,
|
| 135 |
+
"eval_steps_per_second": 30.073,
|
| 136 |
+
"eval_weighted_f1": 0.8199733551426199,
|
| 137 |
+
"step": 390
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.0513478818998716,
|
| 141 |
+
"grad_norm": 95.71858978271484,
|
| 142 |
+
"learning_rate": 1.994871794871795e-05,
|
| 143 |
+
"loss": 2.0482,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.179717586649551,
|
| 148 |
+
"grad_norm": 46.47174072265625,
|
| 149 |
+
"learning_rate": 1.9806267806267806e-05,
|
| 150 |
+
"loss": 1.715,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.3080872913992296,
|
| 155 |
+
"grad_norm": 48.007205963134766,
|
| 156 |
+
"learning_rate": 1.9663817663817667e-05,
|
| 157 |
+
"loss": 1.6391,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.436456996148909,
|
| 162 |
+
"grad_norm": 50.561866760253906,
|
| 163 |
+
"learning_rate": 1.952136752136752e-05,
|
| 164 |
+
"loss": 1.7053,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.564826700898588,
|
| 169 |
+
"grad_norm": 23.58455467224121,
|
| 170 |
+
"learning_rate": 1.9378917378917382e-05,
|
| 171 |
+
"loss": 1.5795,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.693196405648267,
|
| 176 |
+
"grad_norm": 17.744104385375977,
|
| 177 |
+
"learning_rate": 1.923646723646724e-05,
|
| 178 |
+
"loss": 1.7004,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 2.821566110397946,
|
| 183 |
+
"grad_norm": 79.87342834472656,
|
| 184 |
+
"learning_rate": 1.9094017094017097e-05,
|
| 185 |
+
"loss": 1.6293,
|
| 186 |
+
"step": 550
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 2.949935815147625,
|
| 190 |
+
"grad_norm": 42.42940139770508,
|
| 191 |
+
"learning_rate": 1.8951566951566954e-05,
|
| 192 |
+
"loss": 1.4183,
|
| 193 |
+
"step": 575
|
| 194 |
+
},
|
| 195 |
+
{
|
| 196 |
+
"epoch": 3.0,
|
| 197 |
+
"eval_accuracy": 0.842560553633218,
|
| 198 |
+
"eval_loss": 0.438308447599411,
|
| 199 |
+
"eval_macro_f1": 0.8416056571104035,
|
| 200 |
+
"eval_runtime": 1.2573,
|
| 201 |
+
"eval_samples_per_second": 459.725,
|
| 202 |
+
"eval_steps_per_second": 29.429,
|
| 203 |
+
"eval_weighted_f1": 0.8418449527499576,
|
| 204 |
+
"step": 585
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.0770218228498076,
|
| 208 |
+
"grad_norm": 69.78426361083984,
|
| 209 |
+
"learning_rate": 1.880911680911681e-05,
|
| 210 |
+
"loss": 1.3933,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.2053915275994864,
|
| 215 |
+
"grad_norm": 47.075809478759766,
|
| 216 |
+
"learning_rate": 1.866666666666667e-05,
|
| 217 |
+
"loss": 0.9495,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.3337612323491657,
|
| 222 |
+
"grad_norm": 35.8579216003418,
|
| 223 |
+
"learning_rate": 1.8524216524216526e-05,
|
| 224 |
+
"loss": 0.9142,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.462130937098845,
|
| 229 |
+
"grad_norm": 20.10189437866211,
|
| 230 |
+
"learning_rate": 1.8381766381766383e-05,
|
| 231 |
+
"loss": 1.0526,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.5905006418485237,
|
| 236 |
+
"grad_norm": 17.218551635742188,
|
| 237 |
+
"learning_rate": 1.823931623931624e-05,
|
| 238 |
+
"loss": 1.0184,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.718870346598203,
|
| 243 |
+
"grad_norm": 41.16714859008789,
|
| 244 |
+
"learning_rate": 1.8096866096866098e-05,
|
| 245 |
+
"loss": 0.9362,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 3.8472400513478817,
|
| 250 |
+
"grad_norm": 59.61590576171875,
|
| 251 |
+
"learning_rate": 1.7954415954415955e-05,
|
| 252 |
+
"loss": 1.1327,
|
| 253 |
+
"step": 750
|
| 254 |
+
},
|
| 255 |
+
{
|
| 256 |
+
"epoch": 3.975609756097561,
|
| 257 |
+
"grad_norm": 32.190032958984375,
|
| 258 |
+
"learning_rate": 1.7811965811965813e-05,
|
| 259 |
+
"loss": 1.2118,
|
| 260 |
+
"step": 775
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 4.0,
|
| 264 |
+
"eval_accuracy": 0.8460207612456747,
|
| 265 |
+
"eval_loss": 0.479248046875,
|
| 266 |
+
"eval_macro_f1": 0.8461434530745976,
|
| 267 |
+
"eval_runtime": 1.2085,
|
| 268 |
+
"eval_samples_per_second": 478.271,
|
| 269 |
+
"eval_steps_per_second": 30.616,
|
| 270 |
+
"eval_weighted_f1": 0.8462817187889256,
|
| 271 |
+
"step": 780
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.102695763799743,
|
| 275 |
+
"grad_norm": 7.671431064605713,
|
| 276 |
+
"learning_rate": 1.766951566951567e-05,
|
| 277 |
+
"loss": 0.7384,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.2310654685494224,
|
| 282 |
+
"grad_norm": 39.89632034301758,
|
| 283 |
+
"learning_rate": 1.7527065527065528e-05,
|
| 284 |
+
"loss": 0.7164,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.359435173299102,
|
| 289 |
+
"grad_norm": 44.15858840942383,
|
| 290 |
+
"learning_rate": 1.7384615384615385e-05,
|
| 291 |
+
"loss": 0.9587,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.487804878048781,
|
| 296 |
+
"grad_norm": 90.62368774414062,
|
| 297 |
+
"learning_rate": 1.7242165242165242e-05,
|
| 298 |
+
"loss": 0.6554,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.616174582798459,
|
| 303 |
+
"grad_norm": 1.3872368335723877,
|
| 304 |
+
"learning_rate": 1.7099715099715103e-05,
|
| 305 |
+
"loss": 0.681,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 4.7445442875481385,
|
| 310 |
+
"grad_norm": 6.195230484008789,
|
| 311 |
+
"learning_rate": 1.6957264957264957e-05,
|
| 312 |
+
"loss": 0.6918,
|
| 313 |
+
"step": 925
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 4.872913992297818,
|
| 317 |
+
"grad_norm": 147.07801818847656,
|
| 318 |
+
"learning_rate": 1.6814814814814818e-05,
|
| 319 |
+
"loss": 0.5796,
|
| 320 |
+
"step": 950
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 5.0,
|
| 324 |
+
"grad_norm": 14.160751342773438,
|
| 325 |
+
"learning_rate": 1.6672364672364675e-05,
|
| 326 |
+
"loss": 0.8333,
|
| 327 |
+
"step": 975
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 5.0,
|
| 331 |
+
"eval_accuracy": 0.8391003460207612,
|
| 332 |
+
"eval_loss": 0.5641457438468933,
|
| 333 |
+
"eval_macro_f1": 0.8396643546127365,
|
| 334 |
+
"eval_runtime": 1.2612,
|
| 335 |
+
"eval_samples_per_second": 458.311,
|
| 336 |
+
"eval_steps_per_second": 29.338,
|
| 337 |
+
"eval_weighted_f1": 0.8397169728153938,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.128369704749679,
|
| 342 |
+
"grad_norm": 35.67943572998047,
|
| 343 |
+
"learning_rate": 1.652991452991453e-05,
|
| 344 |
+
"loss": 0.5692,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.2567394094993585,
|
| 349 |
+
"grad_norm": 2.249913215637207,
|
| 350 |
+
"learning_rate": 1.638746438746439e-05,
|
| 351 |
+
"loss": 0.4787,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.385109114249037,
|
| 356 |
+
"grad_norm": 48.28496170043945,
|
| 357 |
+
"learning_rate": 1.6245014245014247e-05,
|
| 358 |
+
"loss": 0.4725,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.513478818998716,
|
| 363 |
+
"grad_norm": 2.8437349796295166,
|
| 364 |
+
"learning_rate": 1.6102564102564105e-05,
|
| 365 |
+
"loss": 0.5654,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 5.641848523748395,
|
| 370 |
+
"grad_norm": 40.56161117553711,
|
| 371 |
+
"learning_rate": 1.5960113960113962e-05,
|
| 372 |
+
"loss": 0.6465,
|
| 373 |
+
"step": 1100
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 5.7702182284980745,
|
| 377 |
+
"grad_norm": 7.292684078216553,
|
| 378 |
+
"learning_rate": 1.581766381766382e-05,
|
| 379 |
+
"loss": 0.6451,
|
| 380 |
+
"step": 1125
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 5.898587933247754,
|
| 384 |
+
"grad_norm": 459.4928283691406,
|
| 385 |
+
"learning_rate": 1.5675213675213677e-05,
|
| 386 |
+
"loss": 0.4834,
|
| 387 |
+
"step": 1150
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 6.0,
|
| 391 |
+
"eval_accuracy": 0.8529411764705882,
|
| 392 |
+
"eval_loss": 0.5885111093521118,
|
| 393 |
+
"eval_macro_f1": 0.8524204646306593,
|
| 394 |
+
"eval_runtime": 1.3233,
|
| 395 |
+
"eval_samples_per_second": 436.776,
|
| 396 |
+
"eval_steps_per_second": 27.96,
|
| 397 |
+
"eval_weighted_f1": 0.8525879456553224,
|
| 398 |
+
"step": 1170
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.025673940949936,
|
| 402 |
+
"grad_norm": 24.552349090576172,
|
| 403 |
+
"learning_rate": 1.5532763532763534e-05,
|
| 404 |
+
"loss": 0.5777,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.154043645699615,
|
| 409 |
+
"grad_norm": 66.84564971923828,
|
| 410 |
+
"learning_rate": 1.539031339031339e-05,
|
| 411 |
+
"loss": 0.3109,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.282413350449294,
|
| 416 |
+
"grad_norm": 19.659093856811523,
|
| 417 |
+
"learning_rate": 1.5247863247863249e-05,
|
| 418 |
+
"loss": 0.5152,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.410783055198973,
|
| 423 |
+
"grad_norm": 7.469658851623535,
|
| 424 |
+
"learning_rate": 1.5105413105413106e-05,
|
| 425 |
+
"loss": 0.3627,
|
| 426 |
+
"step": 1250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 6.539152759948652,
|
| 430 |
+
"grad_norm": 144.7767791748047,
|
| 431 |
+
"learning_rate": 1.4962962962962964e-05,
|
| 432 |
+
"loss": 0.4557,
|
| 433 |
+
"step": 1275
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 6.667522464698331,
|
| 437 |
+
"grad_norm": 1.6089030504226685,
|
| 438 |
+
"learning_rate": 1.4820512820512823e-05,
|
| 439 |
+
"loss": 0.3658,
|
| 440 |
+
"step": 1300
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 6.7958921694480106,
|
| 444 |
+
"grad_norm": 65.19385528564453,
|
| 445 |
+
"learning_rate": 1.4678062678062678e-05,
|
| 446 |
+
"loss": 0.3657,
|
| 447 |
+
"step": 1325
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 6.92426187419769,
|
| 451 |
+
"grad_norm": 6.4792866706848145,
|
| 452 |
+
"learning_rate": 1.4535612535612537e-05,
|
| 453 |
+
"loss": 0.4717,
|
| 454 |
+
"step": 1350
|
| 455 |
+
},
|
| 456 |
+
{
|
| 457 |
+
"epoch": 7.0,
|
| 458 |
+
"eval_accuracy": 0.8442906574394463,
|
| 459 |
+
"eval_loss": 0.7064892053604126,
|
| 460 |
+
"eval_macro_f1": 0.8432133689559431,
|
| 461 |
+
"eval_runtime": 1.2693,
|
| 462 |
+
"eval_samples_per_second": 455.358,
|
| 463 |
+
"eval_steps_per_second": 29.149,
|
| 464 |
+
"eval_weighted_f1": 0.8435507614033771,
|
| 465 |
+
"step": 1365
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 7.051347881899872,
|
| 469 |
+
"grad_norm": 34.33955764770508,
|
| 470 |
+
"learning_rate": 1.4393162393162395e-05,
|
| 471 |
+
"loss": 0.3895,
|
| 472 |
+
"step": 1375
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 7.17971758664955,
|
| 476 |
+
"grad_norm": 60.092464447021484,
|
| 477 |
+
"learning_rate": 1.425071225071225e-05,
|
| 478 |
+
"loss": 0.2466,
|
| 479 |
+
"step": 1400
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 7.30808729139923,
|
| 483 |
+
"grad_norm": 43.983917236328125,
|
| 484 |
+
"learning_rate": 1.410826210826211e-05,
|
| 485 |
+
"loss": 0.3693,
|
| 486 |
+
"step": 1425
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 7.436456996148909,
|
| 490 |
+
"grad_norm": 6.01005220413208,
|
| 491 |
+
"learning_rate": 1.3965811965811967e-05,
|
| 492 |
+
"loss": 0.3067,
|
| 493 |
+
"step": 1450
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 7.564826700898588,
|
| 497 |
+
"grad_norm": 0.34186023473739624,
|
| 498 |
+
"learning_rate": 1.3823361823361824e-05,
|
| 499 |
+
"loss": 0.4126,
|
| 500 |
+
"step": 1475
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 7.693196405648267,
|
| 504 |
+
"grad_norm": 15.058670997619629,
|
| 505 |
+
"learning_rate": 1.3680911680911682e-05,
|
| 506 |
+
"loss": 0.4612,
|
| 507 |
+
"step": 1500
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 7.821566110397946,
|
| 511 |
+
"grad_norm": 0.2911962866783142,
|
| 512 |
+
"learning_rate": 1.353846153846154e-05,
|
| 513 |
+
"loss": 0.4724,
|
| 514 |
+
"step": 1525
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 7.949935815147625,
|
| 518 |
+
"grad_norm": 3.1831634044647217,
|
| 519 |
+
"learning_rate": 1.3396011396011396e-05,
|
| 520 |
+
"loss": 0.2389,
|
| 521 |
+
"step": 1550
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 8.0,
|
| 525 |
+
"eval_accuracy": 0.8598615916955017,
|
| 526 |
+
"eval_loss": 0.6812170743942261,
|
| 527 |
+
"eval_macro_f1": 0.8590503728747493,
|
| 528 |
+
"eval_runtime": 1.2245,
|
| 529 |
+
"eval_samples_per_second": 472.016,
|
| 530 |
+
"eval_steps_per_second": 30.216,
|
| 531 |
+
"eval_weighted_f1": 0.8592997656967396,
|
| 532 |
+
"step": 1560
|
| 533 |
+
},
|
| 534 |
+
{
|
| 535 |
+
"epoch": 8.077021822849808,
|
| 536 |
+
"grad_norm": 0.26451513171195984,
|
| 537 |
+
"learning_rate": 1.3253561253561255e-05,
|
| 538 |
+
"loss": 0.2438,
|
| 539 |
+
"step": 1575
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 8.205391527599486,
|
| 543 |
+
"grad_norm": 0.2473522275686264,
|
| 544 |
+
"learning_rate": 1.3111111111111113e-05,
|
| 545 |
+
"loss": 0.2306,
|
| 546 |
+
"step": 1600
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 8.333761232349165,
|
| 550 |
+
"grad_norm": 0.45871150493621826,
|
| 551 |
+
"learning_rate": 1.2968660968660968e-05,
|
| 552 |
+
"loss": 0.2556,
|
| 553 |
+
"step": 1625
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 8.462130937098845,
|
| 557 |
+
"grad_norm": 80.5791015625,
|
| 558 |
+
"learning_rate": 1.2826210826210828e-05,
|
| 559 |
+
"loss": 0.2882,
|
| 560 |
+
"step": 1650
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 8.590500641848523,
|
| 564 |
+
"grad_norm": 1.3554444313049316,
|
| 565 |
+
"learning_rate": 1.2683760683760685e-05,
|
| 566 |
+
"loss": 0.1983,
|
| 567 |
+
"step": 1675
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 8.718870346598203,
|
| 571 |
+
"grad_norm": 0.2086154818534851,
|
| 572 |
+
"learning_rate": 1.2541310541310542e-05,
|
| 573 |
+
"loss": 0.2711,
|
| 574 |
+
"step": 1700
|
| 575 |
+
},
|
| 576 |
+
{
|
| 577 |
+
"epoch": 8.847240051347882,
|
| 578 |
+
"grad_norm": 0.23722316324710846,
|
| 579 |
+
"learning_rate": 1.23988603988604e-05,
|
| 580 |
+
"loss": 0.3298,
|
| 581 |
+
"step": 1725
|
| 582 |
+
},
|
| 583 |
+
{
|
| 584 |
+
"epoch": 8.975609756097562,
|
| 585 |
+
"grad_norm": 0.2185768038034439,
|
| 586 |
+
"learning_rate": 1.2256410256410259e-05,
|
| 587 |
+
"loss": 0.2352,
|
| 588 |
+
"step": 1750
|
| 589 |
+
},
|
| 590 |
+
{
|
| 591 |
+
"epoch": 9.0,
|
| 592 |
+
"eval_accuracy": 0.8633217993079585,
|
| 593 |
+
"eval_loss": 0.7458310127258301,
|
| 594 |
+
"eval_macro_f1": 0.8630274914204891,
|
| 595 |
+
"eval_runtime": 1.248,
|
| 596 |
+
"eval_samples_per_second": 463.139,
|
| 597 |
+
"eval_steps_per_second": 29.647,
|
| 598 |
+
"eval_weighted_f1": 0.8629193015186526,
|
| 599 |
+
"step": 1755
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 9.102695763799744,
|
| 603 |
+
"grad_norm": 20.82085418701172,
|
| 604 |
+
"learning_rate": 1.2113960113960114e-05,
|
| 605 |
+
"loss": 0.2223,
|
| 606 |
+
"step": 1775
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"epoch": 9.231065468549422,
|
| 610 |
+
"grad_norm": 70.62593078613281,
|
| 611 |
+
"learning_rate": 1.1971509971509972e-05,
|
| 612 |
+
"loss": 0.2843,
|
| 613 |
+
"step": 1800
|
| 614 |
+
},
|
| 615 |
+
{
|
| 616 |
+
"epoch": 9.3594351732991,
|
| 617 |
+
"grad_norm": 0.17582455277442932,
|
| 618 |
+
"learning_rate": 1.182905982905983e-05,
|
| 619 |
+
"loss": 0.1697,
|
| 620 |
+
"step": 1825
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 9.487804878048781,
|
| 624 |
+
"grad_norm": 0.3045012056827545,
|
| 625 |
+
"learning_rate": 1.1686609686609687e-05,
|
| 626 |
+
"loss": 0.0505,
|
| 627 |
+
"step": 1850
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 9.61617458279846,
|
| 631 |
+
"grad_norm": 0.15618158876895905,
|
| 632 |
+
"learning_rate": 1.1544159544159546e-05,
|
| 633 |
+
"loss": 0.1075,
|
| 634 |
+
"step": 1875
|
| 635 |
+
},
|
| 636 |
+
{
|
| 637 |
+
"epoch": 9.74454428754814,
|
| 638 |
+
"grad_norm": 3.3294765949249268,
|
| 639 |
+
"learning_rate": 1.1401709401709403e-05,
|
| 640 |
+
"loss": 0.2725,
|
| 641 |
+
"step": 1900
|
| 642 |
+
},
|
| 643 |
+
{
|
| 644 |
+
"epoch": 9.872913992297818,
|
| 645 |
+
"grad_norm": 0.13897669315338135,
|
| 646 |
+
"learning_rate": 1.125925925925926e-05,
|
| 647 |
+
"loss": 0.1761,
|
| 648 |
+
"step": 1925
|
| 649 |
+
},
|
| 650 |
+
{
|
| 651 |
+
"epoch": 10.0,
|
| 652 |
+
"grad_norm": 1.7988786697387695,
|
| 653 |
+
"learning_rate": 1.1116809116809118e-05,
|
| 654 |
+
"loss": 0.296,
|
| 655 |
+
"step": 1950
|
| 656 |
+
},
|
| 657 |
+
{
|
| 658 |
+
"epoch": 10.0,
|
| 659 |
+
"eval_accuracy": 0.856401384083045,
|
| 660 |
+
"eval_loss": 0.7808553576469421,
|
| 661 |
+
"eval_macro_f1": 0.856336060433711,
|
| 662 |
+
"eval_runtime": 1.236,
|
| 663 |
+
"eval_samples_per_second": 467.655,
|
| 664 |
+
"eval_steps_per_second": 29.936,
|
| 665 |
+
"eval_weighted_f1": 0.8562520548625401,
|
| 666 |
+
"step": 1950
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"epoch": 10.128369704749678,
|
| 670 |
+
"grad_norm": 67.68605041503906,
|
| 671 |
+
"learning_rate": 1.0974358974358977e-05,
|
| 672 |
+
"loss": 0.2254,
|
| 673 |
+
"step": 1975
|
| 674 |
+
},
|
| 675 |
+
{
|
| 676 |
+
"epoch": 10.256739409499358,
|
| 677 |
+
"grad_norm": 0.1412421315908432,
|
| 678 |
+
"learning_rate": 1.0831908831908832e-05,
|
| 679 |
+
"loss": 0.1184,
|
| 680 |
+
"step": 2000
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 10.385109114249037,
|
| 684 |
+
"grad_norm": 0.14181852340698242,
|
| 685 |
+
"learning_rate": 1.068945868945869e-05,
|
| 686 |
+
"loss": 0.2071,
|
| 687 |
+
"step": 2025
|
| 688 |
+
},
|
| 689 |
+
{
|
| 690 |
+
"epoch": 10.513478818998717,
|
| 691 |
+
"grad_norm": 0.12591703236103058,
|
| 692 |
+
"learning_rate": 1.0547008547008549e-05,
|
| 693 |
+
"loss": 0.1894,
|
| 694 |
+
"step": 2050
|
| 695 |
+
},
|
| 696 |
+
{
|
| 697 |
+
"epoch": 10.641848523748395,
|
| 698 |
+
"grad_norm": 0.3328363299369812,
|
| 699 |
+
"learning_rate": 1.0404558404558405e-05,
|
| 700 |
+
"loss": 0.2026,
|
| 701 |
+
"step": 2075
|
| 702 |
+
},
|
| 703 |
+
{
|
| 704 |
+
"epoch": 10.770218228498074,
|
| 705 |
+
"grad_norm": 6.01046085357666,
|
| 706 |
+
"learning_rate": 1.0262108262108264e-05,
|
| 707 |
+
"loss": 0.2305,
|
| 708 |
+
"step": 2100
|
| 709 |
+
},
|
| 710 |
+
{
|
| 711 |
+
"epoch": 10.898587933247754,
|
| 712 |
+
"grad_norm": 0.44954463839530945,
|
| 713 |
+
"learning_rate": 1.011965811965812e-05,
|
| 714 |
+
"loss": 0.2382,
|
| 715 |
+
"step": 2125
|
| 716 |
+
},
|
| 717 |
+
{
|
| 718 |
+
"epoch": 11.0,
|
| 719 |
+
"eval_accuracy": 0.8615916955017301,
|
| 720 |
+
"eval_loss": 0.8130559921264648,
|
| 721 |
+
"eval_macro_f1": 0.8615002739259378,
|
| 722 |
+
"eval_runtime": 1.2722,
|
| 723 |
+
"eval_samples_per_second": 454.334,
|
| 724 |
+
"eval_steps_per_second": 29.084,
|
| 725 |
+
"eval_weighted_f1": 0.8615352470040494,
|
| 726 |
+
"step": 2145
|
| 727 |
+
},
|
| 728 |
+
{
|
| 729 |
+
"epoch": 11.025673940949936,
|
| 730 |
+
"grad_norm": 7.375920295715332,
|
| 731 |
+
"learning_rate": 9.977207977207978e-06,
|
| 732 |
+
"loss": 0.1576,
|
| 733 |
+
"step": 2150
|
| 734 |
+
},
|
| 735 |
+
{
|
| 736 |
+
"epoch": 11.154043645699614,
|
| 737 |
+
"grad_norm": 479.59307861328125,
|
| 738 |
+
"learning_rate": 9.834757834757836e-06,
|
| 739 |
+
"loss": 0.0695,
|
| 740 |
+
"step": 2175
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 11.282413350449294,
|
| 744 |
+
"grad_norm": 0.11192168295383453,
|
| 745 |
+
"learning_rate": 9.692307692307693e-06,
|
| 746 |
+
"loss": 0.1549,
|
| 747 |
+
"step": 2200
|
| 748 |
+
},
|
| 749 |
+
{
|
| 750 |
+
"epoch": 11.410783055198973,
|
| 751 |
+
"grad_norm": 0.117418572306633,
|
| 752 |
+
"learning_rate": 9.54985754985755e-06,
|
| 753 |
+
"loss": 0.1265,
|
| 754 |
+
"step": 2225
|
| 755 |
+
},
|
| 756 |
+
{
|
| 757 |
+
"epoch": 11.539152759948653,
|
| 758 |
+
"grad_norm": 2.1237947940826416,
|
| 759 |
+
"learning_rate": 9.407407407407408e-06,
|
| 760 |
+
"loss": 0.0706,
|
| 761 |
+
"step": 2250
|
| 762 |
+
},
|
| 763 |
+
{
|
| 764 |
+
"epoch": 11.667522464698331,
|
| 765 |
+
"grad_norm": 0.23376493155956268,
|
| 766 |
+
"learning_rate": 9.264957264957267e-06,
|
| 767 |
+
"loss": 0.2288,
|
| 768 |
+
"step": 2275
|
| 769 |
+
},
|
| 770 |
+
{
|
| 771 |
+
"epoch": 11.79589216944801,
|
| 772 |
+
"grad_norm": 1.69961416721344,
|
| 773 |
+
"learning_rate": 9.122507122507124e-06,
|
| 774 |
+
"loss": 0.1979,
|
| 775 |
+
"step": 2300
|
| 776 |
+
},
|
| 777 |
+
{
|
| 778 |
+
"epoch": 11.92426187419769,
|
| 779 |
+
"grad_norm": 0.1018761619925499,
|
| 780 |
+
"learning_rate": 8.98005698005698e-06,
|
| 781 |
+
"loss": 0.0724,
|
| 782 |
+
"step": 2325
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 12.0,
|
| 786 |
+
"eval_accuracy": 0.8442906574394463,
|
| 787 |
+
"eval_loss": 0.892682671546936,
|
| 788 |
+
"eval_macro_f1": 0.844903444968808,
|
| 789 |
+
"eval_runtime": 1.2478,
|
| 790 |
+
"eval_samples_per_second": 463.218,
|
| 791 |
+
"eval_steps_per_second": 29.652,
|
| 792 |
+
"eval_weighted_f1": 0.8451140400879661,
|
| 793 |
+
"step": 2340
|
| 794 |
+
}
|
| 795 |
+
],
|
| 796 |
+
"logging_steps": 25,
|
| 797 |
+
"max_steps": 3900,
|
| 798 |
+
"num_input_tokens_seen": 0,
|
| 799 |
+
"num_train_epochs": 20,
|
| 800 |
+
"save_steps": 500,
|
| 801 |
+
"stateful_callbacks": {
|
| 802 |
+
"EarlyStoppingCallback": {
|
| 803 |
+
"args": {
|
| 804 |
+
"early_stopping_patience": 3,
|
| 805 |
+
"early_stopping_threshold": 0.0
|
| 806 |
+
},
|
| 807 |
+
"attributes": {
|
| 808 |
+
"early_stopping_patience_counter": 3
|
| 809 |
+
}
|
| 810 |
+
},
|
| 811 |
+
"TrainerControl": {
|
| 812 |
+
"args": {
|
| 813 |
+
"should_epoch_stop": false,
|
| 814 |
+
"should_evaluate": false,
|
| 815 |
+
"should_log": false,
|
| 816 |
+
"should_save": true,
|
| 817 |
+
"should_training_stop": true
|
| 818 |
+
},
|
| 819 |
+
"attributes": {}
|
| 820 |
+
}
|
| 821 |
+
},
|
| 822 |
+
"total_flos": 0.0,
|
| 823 |
+
"train_batch_size": 8,
|
| 824 |
+
"trial_name": null,
|
| 825 |
+
"trial_params": null
|
| 826 |
+
}
|
ood-vifactcheck-phobert-topic-domain-03-s43-20260814T135955Z-bdd2c308/trainer/checkpoint-2340/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-04-s43-20260814T141531Z-bdd2c308/trainer/checkpoint-1880/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-topic-domain-04-s43-20260814T141531Z-bdd2c308/trainer/checkpoint-1880/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|