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-bamibert-source-source-00-s44-20260829T014846Z-496d6871/trainer/checkpoint-1274/special_tokens_map.json +51 -0
- ood-vifactcheck-bamibert-source-source-00-s44-20260829T014846Z-496d6871/trainer/checkpoint-1274/tokenizer.json +0 -0
- ood-vifactcheck-bamibert-source-source-00-s44-20260829T014846Z-496d6871/trainer/checkpoint-1274/tokenizer_config.json +55 -0
- ood-vifactcheck-bamibert-source-source-00-s44-20260829T014846Z-496d6871/trainer/checkpoint-1274/trainer_state.json +470 -0
- ood-vifactcheck-bamibert-source-source-00-s44-20260829T014846Z-496d6871/trainer/checkpoint-1820/special_tokens_map.json +51 -0
- ood-vifactcheck-bamibert-source-source-00-s44-20260829T014846Z-496d6871/trainer/checkpoint-1820/tokenizer.json +0 -0
- ood-vifactcheck-bamibert-source-source-00-s44-20260829T014846Z-496d6871/trainer/checkpoint-1820/tokenizer_config.json +55 -0
- ood-vifactcheck-bamibert-source-source-00-s44-20260829T014846Z-496d6871/trainer/checkpoint-1820/trainer_state.json +657 -0
- ood-vifactcheck-bamibert-source-source-01-s44-20260829T015815Z-496d6871/trainer/checkpoint-1026/special_tokens_map.json +51 -0
- ood-vifactcheck-bamibert-source-source-01-s44-20260829T015815Z-496d6871/trainer/checkpoint-1026/tokenizer.json +0 -0
- ood-vifactcheck-bamibert-source-source-01-s44-20260829T015815Z-496d6871/trainer/checkpoint-1026/tokenizer_config.json +55 -0
- ood-vifactcheck-bamibert-source-source-01-s44-20260829T015815Z-496d6871/trainer/checkpoint-1026/trainer_state.json +396 -0
- ood-vifactcheck-bamibert-source-source-01-s44-20260829T015815Z-496d6871/trainer/checkpoint-1539/special_tokens_map.json +51 -0
- ood-vifactcheck-bamibert-source-source-01-s44-20260829T015815Z-496d6871/trainer/checkpoint-1539/tokenizer.json +0 -0
- ood-vifactcheck-bamibert-source-source-01-s44-20260829T015815Z-496d6871/trainer/checkpoint-1539/tokenizer_config.json +55 -0
- ood-vifactcheck-bamibert-source-source-01-s44-20260829T015815Z-496d6871/trainer/checkpoint-1539/trainer_state.json +569 -0
- ood-vifactcheck-bamibert-source-source-02-s44-20260829T020641Z-496d6871/trainer/checkpoint-1086/special_tokens_map.json +51 -0
- ood-vifactcheck-bamibert-source-source-02-s44-20260829T020641Z-496d6871/trainer/checkpoint-1086/tokenizer.json +0 -0
- ood-vifactcheck-bamibert-source-source-02-s44-20260829T020641Z-496d6871/trainer/checkpoint-1086/tokenizer_config.json +55 -0
- ood-vifactcheck-bamibert-source-source-02-s44-20260829T020641Z-496d6871/trainer/checkpoint-1086/trainer_state.json +410 -0
- ood-vifactcheck-bamibert-source-source-02-s44-20260829T020641Z-496d6871/trainer/checkpoint-1629/special_tokens_map.json +51 -0
- ood-vifactcheck-bamibert-source-source-02-s44-20260829T020641Z-496d6871/trainer/checkpoint-1629/tokenizer.json +0 -0
- ood-vifactcheck-bamibert-source-source-02-s44-20260829T020641Z-496d6871/trainer/checkpoint-1629/tokenizer_config.json +55 -0
- ood-vifactcheck-bamibert-source-source-02-s44-20260829T020641Z-496d6871/trainer/checkpoint-1629/trainer_state.json +597 -0
- ood-vifactcheck-bamibert-source-source-03-s44-20260829T021529Z-496d6871/trainer/checkpoint-1281/special_tokens_map.json +51 -0
- ood-vifactcheck-bamibert-source-source-03-s44-20260829T021529Z-496d6871/trainer/checkpoint-1281/tokenizer.json +0 -0
- ood-vifactcheck-bamibert-source-source-03-s44-20260829T021529Z-496d6871/trainer/checkpoint-1281/tokenizer_config.json +55 -0
- ood-vifactcheck-bamibert-source-source-03-s44-20260829T021529Z-496d6871/trainer/checkpoint-1281/trainer_state.json +477 -0
- ood-vifactcheck-bamibert-source-source-03-s44-20260829T021529Z-496d6871/trainer/checkpoint-732/special_tokens_map.json +51 -0
- ood-vifactcheck-bamibert-source-source-03-s44-20260829T021529Z-496d6871/trainer/checkpoint-732/tokenizer.json +0 -0
- ood-vifactcheck-bamibert-source-source-03-s44-20260829T021529Z-496d6871/trainer/checkpoint-732/tokenizer_config.json +55 -0
- ood-vifactcheck-bamibert-source-source-03-s44-20260829T021529Z-496d6871/trainer/checkpoint-732/trainer_state.json +290 -0
- ood-vifactcheck-bamibert-source-source-04-s44-20260829T022221Z-496d6871/trainer/checkpoint-1232/special_tokens_map.json +51 -0
- ood-vifactcheck-bamibert-source-source-04-s44-20260829T022221Z-496d6871/trainer/checkpoint-1232/tokenizer.json +0 -0
- ood-vifactcheck-bamibert-source-source-04-s44-20260829T022221Z-496d6871/trainer/checkpoint-1232/tokenizer_config.json +55 -0
- ood-vifactcheck-bamibert-source-source-04-s44-20260829T022221Z-496d6871/trainer/checkpoint-1232/trainer_state.json +463 -0
- ood-vifactcheck-bamibert-source-source-04-s44-20260829T022221Z-496d6871/trainer/checkpoint-704/special_tokens_map.json +51 -0
- ood-vifactcheck-bamibert-source-source-04-s44-20260829T022221Z-496d6871/trainer/checkpoint-704/tokenizer.json +0 -0
- ood-vifactcheck-bamibert-source-source-04-s44-20260829T022221Z-496d6871/trainer/checkpoint-704/tokenizer_config.json +55 -0
- ood-vifactcheck-bamibert-source-source-04-s44-20260829T022221Z-496d6871/trainer/checkpoint-704/trainer_state.json +283 -0
- ood-vifactcheck-bamibert-source-source-05-s44-20260829T022906Z-496d6871/trainer/checkpoint-1620/special_tokens_map.json +51 -0
- ood-vifactcheck-bamibert-source-source-05-s44-20260829T022906Z-496d6871/trainer/checkpoint-1620/tokenizer.json +0 -0
- ood-vifactcheck-bamibert-source-source-05-s44-20260829T022906Z-496d6871/trainer/checkpoint-1620/tokenizer_config.json +55 -0
- ood-vifactcheck-bamibert-source-source-05-s44-20260829T022906Z-496d6871/trainer/checkpoint-1620/trainer_state.json +590 -0
- ood-vifactcheck-bamibert-source-source-05-s44-20260829T022906Z-496d6871/trainer/checkpoint-2160/special_tokens_map.json +51 -0
- ood-vifactcheck-bamibert-source-source-05-s44-20260829T022906Z-496d6871/trainer/checkpoint-2160/tokenizer.json +0 -0
- ood-vifactcheck-bamibert-source-source-05-s44-20260829T022906Z-496d6871/trainer/checkpoint-2160/tokenizer_config.json +55 -0
- ood-vifactcheck-bamibert-source-source-05-s44-20260829T022906Z-496d6871/trainer/checkpoint-2160/trainer_state.json +777 -0
- ood-vifactcheck-bamibert-source-source-06-s44-20260829T024021Z-496d6871/trainer/checkpoint-1480/special_tokens_map.json +51 -0
- ood-vifactcheck-bamibert-source-source-06-s44-20260829T024021Z-496d6871/trainer/checkpoint-1480/tokenizer.json +0 -0
ood-vifactcheck-bamibert-source-source-00-s44-20260829T014846Z-496d6871/trainer/checkpoint-1274/special_tokens_map.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "<s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "</s>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"mask_token": {
|
| 24 |
+
"content": "<mask>",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "<pad>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "</s>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "<unk>",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": false,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
+
}
|
ood-vifactcheck-bamibert-source-source-00-s44-20260829T014846Z-496d6871/trainer/checkpoint-1274/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-bamibert-source-source-00-s44-20260829T014846Z-496d6871/trainer/checkpoint-1274/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"20480": {
|
| 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": 2048,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "XLMRobertaTokenizerFast",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-bamibert-source-source-00-s44-20260829T014846Z-496d6871/trainer/checkpoint-1274/trainer_state.json
ADDED
|
@@ -0,0 +1,470 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1274,
|
| 3 |
+
"best_metric": 0.7969752183912441,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-bamibert-source-source-00-s44-20260829T014846Z-496d6871/trainer/checkpoint-1274",
|
| 5 |
+
"epoch": 7.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1274,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.13793103448275862,
|
| 14 |
+
"grad_norm": 54.01082992553711,
|
| 15 |
+
"learning_rate": 1.3186813186813187e-06,
|
| 16 |
+
"loss": 4.4196,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.27586206896551724,
|
| 21 |
+
"grad_norm": 12.204716682434082,
|
| 22 |
+
"learning_rate": 2.6923076923076923e-06,
|
| 23 |
+
"loss": 4.427,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.41379310344827586,
|
| 28 |
+
"grad_norm": 12.715564727783203,
|
| 29 |
+
"learning_rate": 4.065934065934066e-06,
|
| 30 |
+
"loss": 4.4222,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5517241379310345,
|
| 35 |
+
"grad_norm": 11.498578071594238,
|
| 36 |
+
"learning_rate": 5.43956043956044e-06,
|
| 37 |
+
"loss": 4.4058,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6896551724137931,
|
| 42 |
+
"grad_norm": 42.35261917114258,
|
| 43 |
+
"learning_rate": 6.813186813186814e-06,
|
| 44 |
+
"loss": 4.3856,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8275862068965517,
|
| 49 |
+
"grad_norm": 9.582566261291504,
|
| 50 |
+
"learning_rate": 8.186813186813188e-06,
|
| 51 |
+
"loss": 4.3638,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9655172413793104,
|
| 56 |
+
"grad_norm": 11.560235023498535,
|
| 57 |
+
"learning_rate": 9.560439560439562e-06,
|
| 58 |
+
"loss": 4.3152,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.48325358851674644,
|
| 64 |
+
"eval_loss": 1.0334211587905884,
|
| 65 |
+
"eval_macro_f1": 0.45638882168786105,
|
| 66 |
+
"eval_runtime": 1.1988,
|
| 67 |
+
"eval_samples_per_second": 523.007,
|
| 68 |
+
"eval_steps_per_second": 33.366,
|
| 69 |
+
"eval_weighted_f1": 0.459631819839607,
|
| 70 |
+
"step": 182
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0993103448275863,
|
| 74 |
+
"grad_norm": 14.288512229919434,
|
| 75 |
+
"learning_rate": 1.0934065934065933e-05,
|
| 76 |
+
"loss": 4.0449,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.237241379310345,
|
| 81 |
+
"grad_norm": 15.983325958251953,
|
| 82 |
+
"learning_rate": 1.230769230769231e-05,
|
| 83 |
+
"loss": 4.0666,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.3751724137931034,
|
| 88 |
+
"grad_norm": 42.29781723022461,
|
| 89 |
+
"learning_rate": 1.3681318681318683e-05,
|
| 90 |
+
"loss": 3.6554,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.513103448275862,
|
| 95 |
+
"grad_norm": 23.497636795043945,
|
| 96 |
+
"learning_rate": 1.5054945054945056e-05,
|
| 97 |
+
"loss": 3.1472,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.6510344827586207,
|
| 102 |
+
"grad_norm": 43.16012191772461,
|
| 103 |
+
"learning_rate": 1.642857142857143e-05,
|
| 104 |
+
"loss": 2.7675,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7889655172413792,
|
| 109 |
+
"grad_norm": 39.7609977722168,
|
| 110 |
+
"learning_rate": 1.78021978021978e-05,
|
| 111 |
+
"loss": 2.84,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.926896551724138,
|
| 116 |
+
"grad_norm": 38.72256851196289,
|
| 117 |
+
"learning_rate": 1.9175824175824178e-05,
|
| 118 |
+
"loss": 2.7312,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.7272727272727273,
|
| 124 |
+
"eval_loss": 0.6552956104278564,
|
| 125 |
+
"eval_macro_f1": 0.7262046708473037,
|
| 126 |
+
"eval_runtime": 1.1274,
|
| 127 |
+
"eval_samples_per_second": 556.136,
|
| 128 |
+
"eval_steps_per_second": 35.479,
|
| 129 |
+
"eval_weighted_f1": 0.7259729375405137,
|
| 130 |
+
"step": 364
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.0606896551724136,
|
| 134 |
+
"grad_norm": 27.361074447631836,
|
| 135 |
+
"learning_rate": 1.993894993894994e-05,
|
| 136 |
+
"loss": 2.2093,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.1986206896551725,
|
| 141 |
+
"grad_norm": 63.280757904052734,
|
| 142 |
+
"learning_rate": 1.9786324786324788e-05,
|
| 143 |
+
"loss": 2.0278,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.336551724137931,
|
| 148 |
+
"grad_norm": 54.477439880371094,
|
| 149 |
+
"learning_rate": 1.9633699633699634e-05,
|
| 150 |
+
"loss": 1.8783,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.47448275862069,
|
| 155 |
+
"grad_norm": 36.70521926879883,
|
| 156 |
+
"learning_rate": 1.948107448107448e-05,
|
| 157 |
+
"loss": 1.8676,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.6124137931034483,
|
| 162 |
+
"grad_norm": 36.292030334472656,
|
| 163 |
+
"learning_rate": 1.932844932844933e-05,
|
| 164 |
+
"loss": 1.7396,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.750344827586207,
|
| 169 |
+
"grad_norm": 36.98339080810547,
|
| 170 |
+
"learning_rate": 1.9175824175824178e-05,
|
| 171 |
+
"loss": 1.9373,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.8882758620689657,
|
| 176 |
+
"grad_norm": 41.03074264526367,
|
| 177 |
+
"learning_rate": 1.9023199023199025e-05,
|
| 178 |
+
"loss": 1.8564,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 3.0,
|
| 183 |
+
"eval_accuracy": 0.7862838915470495,
|
| 184 |
+
"eval_loss": 0.5142162442207336,
|
| 185 |
+
"eval_macro_f1": 0.7851252524733177,
|
| 186 |
+
"eval_runtime": 1.1526,
|
| 187 |
+
"eval_samples_per_second": 544.006,
|
| 188 |
+
"eval_steps_per_second": 34.705,
|
| 189 |
+
"eval_weighted_f1": 0.7849732656354655,
|
| 190 |
+
"step": 546
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.0220689655172412,
|
| 194 |
+
"grad_norm": 59.45548629760742,
|
| 195 |
+
"learning_rate": 1.887057387057387e-05,
|
| 196 |
+
"loss": 1.8189,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.16,
|
| 201 |
+
"grad_norm": 56.83470153808594,
|
| 202 |
+
"learning_rate": 1.8717948717948718e-05,
|
| 203 |
+
"loss": 1.3462,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.2979310344827586,
|
| 208 |
+
"grad_norm": 37.6110725402832,
|
| 209 |
+
"learning_rate": 1.8565323565323568e-05,
|
| 210 |
+
"loss": 1.3509,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.435862068965517,
|
| 215 |
+
"grad_norm": 86.88182067871094,
|
| 216 |
+
"learning_rate": 1.8412698412698415e-05,
|
| 217 |
+
"loss": 1.2755,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.573793103448276,
|
| 222 |
+
"grad_norm": 51.880916595458984,
|
| 223 |
+
"learning_rate": 1.826007326007326e-05,
|
| 224 |
+
"loss": 1.2467,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.7117241379310344,
|
| 229 |
+
"grad_norm": 53.63358688354492,
|
| 230 |
+
"learning_rate": 1.8107448107448108e-05,
|
| 231 |
+
"loss": 1.3301,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.849655172413793,
|
| 236 |
+
"grad_norm": 41.2372932434082,
|
| 237 |
+
"learning_rate": 1.7954822954822958e-05,
|
| 238 |
+
"loss": 1.3166,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.9875862068965517,
|
| 243 |
+
"grad_norm": 50.010047912597656,
|
| 244 |
+
"learning_rate": 1.78021978021978e-05,
|
| 245 |
+
"loss": 1.2189,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 4.0,
|
| 250 |
+
"eval_accuracy": 0.7942583732057417,
|
| 251 |
+
"eval_loss": 0.6195688247680664,
|
| 252 |
+
"eval_macro_f1": 0.7930461477520301,
|
| 253 |
+
"eval_runtime": 1.3727,
|
| 254 |
+
"eval_samples_per_second": 456.769,
|
| 255 |
+
"eval_steps_per_second": 29.14,
|
| 256 |
+
"eval_weighted_f1": 0.7933870082758345,
|
| 257 |
+
"step": 728
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.121379310344827,
|
| 261 |
+
"grad_norm": 32.44444274902344,
|
| 262 |
+
"learning_rate": 1.764957264957265e-05,
|
| 263 |
+
"loss": 0.678,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.259310344827586,
|
| 268 |
+
"grad_norm": 56.89309310913086,
|
| 269 |
+
"learning_rate": 1.74969474969475e-05,
|
| 270 |
+
"loss": 0.7734,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.397241379310345,
|
| 275 |
+
"grad_norm": 8.098810195922852,
|
| 276 |
+
"learning_rate": 1.7344322344322345e-05,
|
| 277 |
+
"loss": 0.7491,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.535172413793103,
|
| 282 |
+
"grad_norm": 94.43305969238281,
|
| 283 |
+
"learning_rate": 1.7191697191697195e-05,
|
| 284 |
+
"loss": 1.0379,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.673103448275862,
|
| 289 |
+
"grad_norm": 47.52196502685547,
|
| 290 |
+
"learning_rate": 1.703907203907204e-05,
|
| 291 |
+
"loss": 0.9656,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.811034482758621,
|
| 296 |
+
"grad_norm": 40.76150131225586,
|
| 297 |
+
"learning_rate": 1.688644688644689e-05,
|
| 298 |
+
"loss": 0.9777,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.94896551724138,
|
| 303 |
+
"grad_norm": 54.57170867919922,
|
| 304 |
+
"learning_rate": 1.6733821733821735e-05,
|
| 305 |
+
"loss": 1.0308,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 5.0,
|
| 310 |
+
"eval_accuracy": 0.7799043062200957,
|
| 311 |
+
"eval_loss": 0.6780532598495483,
|
| 312 |
+
"eval_macro_f1": 0.7803155865479537,
|
| 313 |
+
"eval_runtime": 1.1274,
|
| 314 |
+
"eval_samples_per_second": 556.143,
|
| 315 |
+
"eval_steps_per_second": 35.48,
|
| 316 |
+
"eval_weighted_f1": 0.7800841058298832,
|
| 317 |
+
"step": 910
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.082758620689655,
|
| 321 |
+
"grad_norm": 82.09213256835938,
|
| 322 |
+
"learning_rate": 1.6581196581196585e-05,
|
| 323 |
+
"loss": 0.6402,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.220689655172414,
|
| 328 |
+
"grad_norm": 7.258779525756836,
|
| 329 |
+
"learning_rate": 1.642857142857143e-05,
|
| 330 |
+
"loss": 0.6106,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.358620689655172,
|
| 335 |
+
"grad_norm": 111.25167083740234,
|
| 336 |
+
"learning_rate": 1.627594627594628e-05,
|
| 337 |
+
"loss": 0.5954,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.496551724137931,
|
| 342 |
+
"grad_norm": 68.25827026367188,
|
| 343 |
+
"learning_rate": 1.6123321123321125e-05,
|
| 344 |
+
"loss": 0.5107,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.63448275862069,
|
| 349 |
+
"grad_norm": 30.255788803100586,
|
| 350 |
+
"learning_rate": 1.5970695970695972e-05,
|
| 351 |
+
"loss": 0.6666,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.772413793103448,
|
| 356 |
+
"grad_norm": 109.78109741210938,
|
| 357 |
+
"learning_rate": 1.581807081807082e-05,
|
| 358 |
+
"loss": 0.6979,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.910344827586207,
|
| 363 |
+
"grad_norm": 80.31668853759766,
|
| 364 |
+
"learning_rate": 1.5665445665445665e-05,
|
| 365 |
+
"loss": 0.632,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 6.0,
|
| 370 |
+
"eval_accuracy": 0.79585326953748,
|
| 371 |
+
"eval_loss": 0.7992599606513977,
|
| 372 |
+
"eval_macro_f1": 0.7954922584251038,
|
| 373 |
+
"eval_runtime": 1.2083,
|
| 374 |
+
"eval_samples_per_second": 518.904,
|
| 375 |
+
"eval_steps_per_second": 33.104,
|
| 376 |
+
"eval_weighted_f1": 0.7957097552967219,
|
| 377 |
+
"step": 1092
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 6.0441379310344825,
|
| 381 |
+
"grad_norm": 42.77943420410156,
|
| 382 |
+
"learning_rate": 1.5512820512820516e-05,
|
| 383 |
+
"loss": 0.4553,
|
| 384 |
+
"step": 1100
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 6.182068965517241,
|
| 388 |
+
"grad_norm": 63.570701599121094,
|
| 389 |
+
"learning_rate": 1.5360195360195362e-05,
|
| 390 |
+
"loss": 0.4478,
|
| 391 |
+
"step": 1125
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 6.32,
|
| 395 |
+
"grad_norm": 102.54994201660156,
|
| 396 |
+
"learning_rate": 1.5207570207570207e-05,
|
| 397 |
+
"loss": 0.6502,
|
| 398 |
+
"step": 1150
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.457931034482758,
|
| 402 |
+
"grad_norm": 49.475460052490234,
|
| 403 |
+
"learning_rate": 1.5054945054945056e-05,
|
| 404 |
+
"loss": 0.5545,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.595862068965517,
|
| 409 |
+
"grad_norm": 141.17816162109375,
|
| 410 |
+
"learning_rate": 1.4902319902319904e-05,
|
| 411 |
+
"loss": 0.4004,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.733793103448276,
|
| 416 |
+
"grad_norm": 91.87596893310547,
|
| 417 |
+
"learning_rate": 1.4749694749694752e-05,
|
| 418 |
+
"loss": 0.5273,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.871724137931034,
|
| 423 |
+
"grad_norm": 78.8913345336914,
|
| 424 |
+
"learning_rate": 1.4597069597069597e-05,
|
| 425 |
+
"loss": 0.4044,
|
| 426 |
+
"step": 1250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 7.0,
|
| 430 |
+
"eval_accuracy": 0.7990430622009569,
|
| 431 |
+
"eval_loss": 0.8133569359779358,
|
| 432 |
+
"eval_macro_f1": 0.7969752183912441,
|
| 433 |
+
"eval_runtime": 1.1915,
|
| 434 |
+
"eval_samples_per_second": 526.245,
|
| 435 |
+
"eval_steps_per_second": 33.572,
|
| 436 |
+
"eval_weighted_f1": 0.7972947353658276,
|
| 437 |
+
"step": 1274
|
| 438 |
+
}
|
| 439 |
+
],
|
| 440 |
+
"logging_steps": 25,
|
| 441 |
+
"max_steps": 3640,
|
| 442 |
+
"num_input_tokens_seen": 0,
|
| 443 |
+
"num_train_epochs": 20,
|
| 444 |
+
"save_steps": 500,
|
| 445 |
+
"stateful_callbacks": {
|
| 446 |
+
"EarlyStoppingCallback": {
|
| 447 |
+
"args": {
|
| 448 |
+
"early_stopping_patience": 3,
|
| 449 |
+
"early_stopping_threshold": 0.0
|
| 450 |
+
},
|
| 451 |
+
"attributes": {
|
| 452 |
+
"early_stopping_patience_counter": 0
|
| 453 |
+
}
|
| 454 |
+
},
|
| 455 |
+
"TrainerControl": {
|
| 456 |
+
"args": {
|
| 457 |
+
"should_epoch_stop": false,
|
| 458 |
+
"should_evaluate": false,
|
| 459 |
+
"should_log": false,
|
| 460 |
+
"should_save": true,
|
| 461 |
+
"should_training_stop": false
|
| 462 |
+
},
|
| 463 |
+
"attributes": {}
|
| 464 |
+
}
|
| 465 |
+
},
|
| 466 |
+
"total_flos": 0.0,
|
| 467 |
+
"train_batch_size": 8,
|
| 468 |
+
"trial_name": null,
|
| 469 |
+
"trial_params": null
|
| 470 |
+
}
|
ood-vifactcheck-bamibert-source-source-00-s44-20260829T014846Z-496d6871/trainer/checkpoint-1820/special_tokens_map.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "<s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "</s>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"mask_token": {
|
| 24 |
+
"content": "<mask>",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "<pad>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "</s>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "<unk>",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": false,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
+
}
|
ood-vifactcheck-bamibert-source-source-00-s44-20260829T014846Z-496d6871/trainer/checkpoint-1820/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-bamibert-source-source-00-s44-20260829T014846Z-496d6871/trainer/checkpoint-1820/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 |
+
"20480": {
|
| 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": 2048,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "XLMRobertaTokenizerFast",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-bamibert-source-source-00-s44-20260829T014846Z-496d6871/trainer/checkpoint-1820/trainer_state.json
ADDED
|
@@ -0,0 +1,657 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1274,
|
| 3 |
+
"best_metric": 0.7969752183912441,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-bamibert-source-source-00-s44-20260829T014846Z-496d6871/trainer/checkpoint-1274",
|
| 5 |
+
"epoch": 10.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1820,
|
| 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": 54.01082992553711,
|
| 15 |
+
"learning_rate": 1.3186813186813187e-06,
|
| 16 |
+
"loss": 4.4196,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.27586206896551724,
|
| 21 |
+
"grad_norm": 12.204716682434082,
|
| 22 |
+
"learning_rate": 2.6923076923076923e-06,
|
| 23 |
+
"loss": 4.427,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.41379310344827586,
|
| 28 |
+
"grad_norm": 12.715564727783203,
|
| 29 |
+
"learning_rate": 4.065934065934066e-06,
|
| 30 |
+
"loss": 4.4222,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5517241379310345,
|
| 35 |
+
"grad_norm": 11.498578071594238,
|
| 36 |
+
"learning_rate": 5.43956043956044e-06,
|
| 37 |
+
"loss": 4.4058,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6896551724137931,
|
| 42 |
+
"grad_norm": 42.35261917114258,
|
| 43 |
+
"learning_rate": 6.813186813186814e-06,
|
| 44 |
+
"loss": 4.3856,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8275862068965517,
|
| 49 |
+
"grad_norm": 9.582566261291504,
|
| 50 |
+
"learning_rate": 8.186813186813188e-06,
|
| 51 |
+
"loss": 4.3638,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9655172413793104,
|
| 56 |
+
"grad_norm": 11.560235023498535,
|
| 57 |
+
"learning_rate": 9.560439560439562e-06,
|
| 58 |
+
"loss": 4.3152,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.48325358851674644,
|
| 64 |
+
"eval_loss": 1.0334211587905884,
|
| 65 |
+
"eval_macro_f1": 0.45638882168786105,
|
| 66 |
+
"eval_runtime": 1.1988,
|
| 67 |
+
"eval_samples_per_second": 523.007,
|
| 68 |
+
"eval_steps_per_second": 33.366,
|
| 69 |
+
"eval_weighted_f1": 0.459631819839607,
|
| 70 |
+
"step": 182
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0993103448275863,
|
| 74 |
+
"grad_norm": 14.288512229919434,
|
| 75 |
+
"learning_rate": 1.0934065934065933e-05,
|
| 76 |
+
"loss": 4.0449,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.237241379310345,
|
| 81 |
+
"grad_norm": 15.983325958251953,
|
| 82 |
+
"learning_rate": 1.230769230769231e-05,
|
| 83 |
+
"loss": 4.0666,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.3751724137931034,
|
| 88 |
+
"grad_norm": 42.29781723022461,
|
| 89 |
+
"learning_rate": 1.3681318681318683e-05,
|
| 90 |
+
"loss": 3.6554,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.513103448275862,
|
| 95 |
+
"grad_norm": 23.497636795043945,
|
| 96 |
+
"learning_rate": 1.5054945054945056e-05,
|
| 97 |
+
"loss": 3.1472,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.6510344827586207,
|
| 102 |
+
"grad_norm": 43.16012191772461,
|
| 103 |
+
"learning_rate": 1.642857142857143e-05,
|
| 104 |
+
"loss": 2.7675,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7889655172413792,
|
| 109 |
+
"grad_norm": 39.7609977722168,
|
| 110 |
+
"learning_rate": 1.78021978021978e-05,
|
| 111 |
+
"loss": 2.84,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.926896551724138,
|
| 116 |
+
"grad_norm": 38.72256851196289,
|
| 117 |
+
"learning_rate": 1.9175824175824178e-05,
|
| 118 |
+
"loss": 2.7312,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.7272727272727273,
|
| 124 |
+
"eval_loss": 0.6552956104278564,
|
| 125 |
+
"eval_macro_f1": 0.7262046708473037,
|
| 126 |
+
"eval_runtime": 1.1274,
|
| 127 |
+
"eval_samples_per_second": 556.136,
|
| 128 |
+
"eval_steps_per_second": 35.479,
|
| 129 |
+
"eval_weighted_f1": 0.7259729375405137,
|
| 130 |
+
"step": 364
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.0606896551724136,
|
| 134 |
+
"grad_norm": 27.361074447631836,
|
| 135 |
+
"learning_rate": 1.993894993894994e-05,
|
| 136 |
+
"loss": 2.2093,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.1986206896551725,
|
| 141 |
+
"grad_norm": 63.280757904052734,
|
| 142 |
+
"learning_rate": 1.9786324786324788e-05,
|
| 143 |
+
"loss": 2.0278,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.336551724137931,
|
| 148 |
+
"grad_norm": 54.477439880371094,
|
| 149 |
+
"learning_rate": 1.9633699633699634e-05,
|
| 150 |
+
"loss": 1.8783,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.47448275862069,
|
| 155 |
+
"grad_norm": 36.70521926879883,
|
| 156 |
+
"learning_rate": 1.948107448107448e-05,
|
| 157 |
+
"loss": 1.8676,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.6124137931034483,
|
| 162 |
+
"grad_norm": 36.292030334472656,
|
| 163 |
+
"learning_rate": 1.932844932844933e-05,
|
| 164 |
+
"loss": 1.7396,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.750344827586207,
|
| 169 |
+
"grad_norm": 36.98339080810547,
|
| 170 |
+
"learning_rate": 1.9175824175824178e-05,
|
| 171 |
+
"loss": 1.9373,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.8882758620689657,
|
| 176 |
+
"grad_norm": 41.03074264526367,
|
| 177 |
+
"learning_rate": 1.9023199023199025e-05,
|
| 178 |
+
"loss": 1.8564,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 3.0,
|
| 183 |
+
"eval_accuracy": 0.7862838915470495,
|
| 184 |
+
"eval_loss": 0.5142162442207336,
|
| 185 |
+
"eval_macro_f1": 0.7851252524733177,
|
| 186 |
+
"eval_runtime": 1.1526,
|
| 187 |
+
"eval_samples_per_second": 544.006,
|
| 188 |
+
"eval_steps_per_second": 34.705,
|
| 189 |
+
"eval_weighted_f1": 0.7849732656354655,
|
| 190 |
+
"step": 546
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.0220689655172412,
|
| 194 |
+
"grad_norm": 59.45548629760742,
|
| 195 |
+
"learning_rate": 1.887057387057387e-05,
|
| 196 |
+
"loss": 1.8189,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.16,
|
| 201 |
+
"grad_norm": 56.83470153808594,
|
| 202 |
+
"learning_rate": 1.8717948717948718e-05,
|
| 203 |
+
"loss": 1.3462,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.2979310344827586,
|
| 208 |
+
"grad_norm": 37.6110725402832,
|
| 209 |
+
"learning_rate": 1.8565323565323568e-05,
|
| 210 |
+
"loss": 1.3509,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.435862068965517,
|
| 215 |
+
"grad_norm": 86.88182067871094,
|
| 216 |
+
"learning_rate": 1.8412698412698415e-05,
|
| 217 |
+
"loss": 1.2755,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.573793103448276,
|
| 222 |
+
"grad_norm": 51.880916595458984,
|
| 223 |
+
"learning_rate": 1.826007326007326e-05,
|
| 224 |
+
"loss": 1.2467,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.7117241379310344,
|
| 229 |
+
"grad_norm": 53.63358688354492,
|
| 230 |
+
"learning_rate": 1.8107448107448108e-05,
|
| 231 |
+
"loss": 1.3301,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.849655172413793,
|
| 236 |
+
"grad_norm": 41.2372932434082,
|
| 237 |
+
"learning_rate": 1.7954822954822958e-05,
|
| 238 |
+
"loss": 1.3166,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.9875862068965517,
|
| 243 |
+
"grad_norm": 50.010047912597656,
|
| 244 |
+
"learning_rate": 1.78021978021978e-05,
|
| 245 |
+
"loss": 1.2189,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 4.0,
|
| 250 |
+
"eval_accuracy": 0.7942583732057417,
|
| 251 |
+
"eval_loss": 0.6195688247680664,
|
| 252 |
+
"eval_macro_f1": 0.7930461477520301,
|
| 253 |
+
"eval_runtime": 1.3727,
|
| 254 |
+
"eval_samples_per_second": 456.769,
|
| 255 |
+
"eval_steps_per_second": 29.14,
|
| 256 |
+
"eval_weighted_f1": 0.7933870082758345,
|
| 257 |
+
"step": 728
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.121379310344827,
|
| 261 |
+
"grad_norm": 32.44444274902344,
|
| 262 |
+
"learning_rate": 1.764957264957265e-05,
|
| 263 |
+
"loss": 0.678,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.259310344827586,
|
| 268 |
+
"grad_norm": 56.89309310913086,
|
| 269 |
+
"learning_rate": 1.74969474969475e-05,
|
| 270 |
+
"loss": 0.7734,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.397241379310345,
|
| 275 |
+
"grad_norm": 8.098810195922852,
|
| 276 |
+
"learning_rate": 1.7344322344322345e-05,
|
| 277 |
+
"loss": 0.7491,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.535172413793103,
|
| 282 |
+
"grad_norm": 94.43305969238281,
|
| 283 |
+
"learning_rate": 1.7191697191697195e-05,
|
| 284 |
+
"loss": 1.0379,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.673103448275862,
|
| 289 |
+
"grad_norm": 47.52196502685547,
|
| 290 |
+
"learning_rate": 1.703907203907204e-05,
|
| 291 |
+
"loss": 0.9656,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.811034482758621,
|
| 296 |
+
"grad_norm": 40.76150131225586,
|
| 297 |
+
"learning_rate": 1.688644688644689e-05,
|
| 298 |
+
"loss": 0.9777,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.94896551724138,
|
| 303 |
+
"grad_norm": 54.57170867919922,
|
| 304 |
+
"learning_rate": 1.6733821733821735e-05,
|
| 305 |
+
"loss": 1.0308,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 5.0,
|
| 310 |
+
"eval_accuracy": 0.7799043062200957,
|
| 311 |
+
"eval_loss": 0.6780532598495483,
|
| 312 |
+
"eval_macro_f1": 0.7803155865479537,
|
| 313 |
+
"eval_runtime": 1.1274,
|
| 314 |
+
"eval_samples_per_second": 556.143,
|
| 315 |
+
"eval_steps_per_second": 35.48,
|
| 316 |
+
"eval_weighted_f1": 0.7800841058298832,
|
| 317 |
+
"step": 910
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.082758620689655,
|
| 321 |
+
"grad_norm": 82.09213256835938,
|
| 322 |
+
"learning_rate": 1.6581196581196585e-05,
|
| 323 |
+
"loss": 0.6402,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.220689655172414,
|
| 328 |
+
"grad_norm": 7.258779525756836,
|
| 329 |
+
"learning_rate": 1.642857142857143e-05,
|
| 330 |
+
"loss": 0.6106,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.358620689655172,
|
| 335 |
+
"grad_norm": 111.25167083740234,
|
| 336 |
+
"learning_rate": 1.627594627594628e-05,
|
| 337 |
+
"loss": 0.5954,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.496551724137931,
|
| 342 |
+
"grad_norm": 68.25827026367188,
|
| 343 |
+
"learning_rate": 1.6123321123321125e-05,
|
| 344 |
+
"loss": 0.5107,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.63448275862069,
|
| 349 |
+
"grad_norm": 30.255788803100586,
|
| 350 |
+
"learning_rate": 1.5970695970695972e-05,
|
| 351 |
+
"loss": 0.6666,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.772413793103448,
|
| 356 |
+
"grad_norm": 109.78109741210938,
|
| 357 |
+
"learning_rate": 1.581807081807082e-05,
|
| 358 |
+
"loss": 0.6979,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.910344827586207,
|
| 363 |
+
"grad_norm": 80.31668853759766,
|
| 364 |
+
"learning_rate": 1.5665445665445665e-05,
|
| 365 |
+
"loss": 0.632,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 6.0,
|
| 370 |
+
"eval_accuracy": 0.79585326953748,
|
| 371 |
+
"eval_loss": 0.7992599606513977,
|
| 372 |
+
"eval_macro_f1": 0.7954922584251038,
|
| 373 |
+
"eval_runtime": 1.2083,
|
| 374 |
+
"eval_samples_per_second": 518.904,
|
| 375 |
+
"eval_steps_per_second": 33.104,
|
| 376 |
+
"eval_weighted_f1": 0.7957097552967219,
|
| 377 |
+
"step": 1092
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 6.0441379310344825,
|
| 381 |
+
"grad_norm": 42.77943420410156,
|
| 382 |
+
"learning_rate": 1.5512820512820516e-05,
|
| 383 |
+
"loss": 0.4553,
|
| 384 |
+
"step": 1100
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 6.182068965517241,
|
| 388 |
+
"grad_norm": 63.570701599121094,
|
| 389 |
+
"learning_rate": 1.5360195360195362e-05,
|
| 390 |
+
"loss": 0.4478,
|
| 391 |
+
"step": 1125
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 6.32,
|
| 395 |
+
"grad_norm": 102.54994201660156,
|
| 396 |
+
"learning_rate": 1.5207570207570207e-05,
|
| 397 |
+
"loss": 0.6502,
|
| 398 |
+
"step": 1150
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.457931034482758,
|
| 402 |
+
"grad_norm": 49.475460052490234,
|
| 403 |
+
"learning_rate": 1.5054945054945056e-05,
|
| 404 |
+
"loss": 0.5545,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.595862068965517,
|
| 409 |
+
"grad_norm": 141.17816162109375,
|
| 410 |
+
"learning_rate": 1.4902319902319904e-05,
|
| 411 |
+
"loss": 0.4004,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.733793103448276,
|
| 416 |
+
"grad_norm": 91.87596893310547,
|
| 417 |
+
"learning_rate": 1.4749694749694752e-05,
|
| 418 |
+
"loss": 0.5273,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.871724137931034,
|
| 423 |
+
"grad_norm": 78.8913345336914,
|
| 424 |
+
"learning_rate": 1.4597069597069597e-05,
|
| 425 |
+
"loss": 0.4044,
|
| 426 |
+
"step": 1250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 7.0,
|
| 430 |
+
"eval_accuracy": 0.7990430622009569,
|
| 431 |
+
"eval_loss": 0.8133569359779358,
|
| 432 |
+
"eval_macro_f1": 0.7969752183912441,
|
| 433 |
+
"eval_runtime": 1.1915,
|
| 434 |
+
"eval_samples_per_second": 526.245,
|
| 435 |
+
"eval_steps_per_second": 33.572,
|
| 436 |
+
"eval_weighted_f1": 0.7972947353658276,
|
| 437 |
+
"step": 1274
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 7.0055172413793105,
|
| 441 |
+
"grad_norm": 17.523698806762695,
|
| 442 |
+
"learning_rate": 1.4444444444444446e-05,
|
| 443 |
+
"loss": 0.3918,
|
| 444 |
+
"step": 1275
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 7.1434482758620685,
|
| 448 |
+
"grad_norm": 69.68782043457031,
|
| 449 |
+
"learning_rate": 1.4291819291819293e-05,
|
| 450 |
+
"loss": 0.2294,
|
| 451 |
+
"step": 1300
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 7.281379310344827,
|
| 455 |
+
"grad_norm": 79.98806762695312,
|
| 456 |
+
"learning_rate": 1.4139194139194141e-05,
|
| 457 |
+
"loss": 0.4722,
|
| 458 |
+
"step": 1325
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 7.419310344827586,
|
| 462 |
+
"grad_norm": 226.70718383789062,
|
| 463 |
+
"learning_rate": 1.3986568986568988e-05,
|
| 464 |
+
"loss": 0.2555,
|
| 465 |
+
"step": 1350
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 7.557241379310344,
|
| 469 |
+
"grad_norm": 94.90642547607422,
|
| 470 |
+
"learning_rate": 1.3833943833943834e-05,
|
| 471 |
+
"loss": 0.274,
|
| 472 |
+
"step": 1375
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 7.695172413793103,
|
| 476 |
+
"grad_norm": 5.394479274749756,
|
| 477 |
+
"learning_rate": 1.3681318681318683e-05,
|
| 478 |
+
"loss": 0.3056,
|
| 479 |
+
"step": 1400
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 7.833103448275862,
|
| 483 |
+
"grad_norm": 130.3593292236328,
|
| 484 |
+
"learning_rate": 1.3528693528693531e-05,
|
| 485 |
+
"loss": 0.2706,
|
| 486 |
+
"step": 1425
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 7.971034482758621,
|
| 490 |
+
"grad_norm": 147.6232147216797,
|
| 491 |
+
"learning_rate": 1.3376068376068376e-05,
|
| 492 |
+
"loss": 0.6034,
|
| 493 |
+
"step": 1450
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 8.0,
|
| 497 |
+
"eval_accuracy": 0.7942583732057417,
|
| 498 |
+
"eval_loss": 1.1467808485031128,
|
| 499 |
+
"eval_macro_f1": 0.7945346813974496,
|
| 500 |
+
"eval_runtime": 1.1398,
|
| 501 |
+
"eval_samples_per_second": 550.102,
|
| 502 |
+
"eval_steps_per_second": 35.094,
|
| 503 |
+
"eval_weighted_f1": 0.7950505714413084,
|
| 504 |
+
"step": 1456
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 8.104827586206897,
|
| 508 |
+
"grad_norm": 0.9876725077629089,
|
| 509 |
+
"learning_rate": 1.3223443223443224e-05,
|
| 510 |
+
"loss": 0.4463,
|
| 511 |
+
"step": 1475
|
| 512 |
+
},
|
| 513 |
+
{
|
| 514 |
+
"epoch": 8.242758620689655,
|
| 515 |
+
"grad_norm": 0.5366671085357666,
|
| 516 |
+
"learning_rate": 1.3070818070818073e-05,
|
| 517 |
+
"loss": 0.1693,
|
| 518 |
+
"step": 1500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 8.380689655172414,
|
| 522 |
+
"grad_norm": 115.87263488769531,
|
| 523 |
+
"learning_rate": 1.291819291819292e-05,
|
| 524 |
+
"loss": 0.2498,
|
| 525 |
+
"step": 1525
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 8.518620689655172,
|
| 529 |
+
"grad_norm": 3.285022020339966,
|
| 530 |
+
"learning_rate": 1.2765567765567766e-05,
|
| 531 |
+
"loss": 0.2392,
|
| 532 |
+
"step": 1550
|
| 533 |
+
},
|
| 534 |
+
{
|
| 535 |
+
"epoch": 8.65655172413793,
|
| 536 |
+
"grad_norm": 11.375731468200684,
|
| 537 |
+
"learning_rate": 1.2612942612942613e-05,
|
| 538 |
+
"loss": 0.0957,
|
| 539 |
+
"step": 1575
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 8.79448275862069,
|
| 543 |
+
"grad_norm": 127.38224792480469,
|
| 544 |
+
"learning_rate": 1.2460317460317461e-05,
|
| 545 |
+
"loss": 0.3758,
|
| 546 |
+
"step": 1600
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 8.932413793103448,
|
| 550 |
+
"grad_norm": 2.712312698364258,
|
| 551 |
+
"learning_rate": 1.230769230769231e-05,
|
| 552 |
+
"loss": 0.2001,
|
| 553 |
+
"step": 1625
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 9.0,
|
| 557 |
+
"eval_accuracy": 0.7783094098883573,
|
| 558 |
+
"eval_loss": 1.3673937320709229,
|
| 559 |
+
"eval_macro_f1": 0.77898279803167,
|
| 560 |
+
"eval_runtime": 1.2133,
|
| 561 |
+
"eval_samples_per_second": 516.771,
|
| 562 |
+
"eval_steps_per_second": 32.968,
|
| 563 |
+
"eval_weighted_f1": 0.7787296368698932,
|
| 564 |
+
"step": 1638
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"epoch": 9.066206896551725,
|
| 568 |
+
"grad_norm": 2.5123658180236816,
|
| 569 |
+
"learning_rate": 1.2155067155067155e-05,
|
| 570 |
+
"loss": 0.2097,
|
| 571 |
+
"step": 1650
|
| 572 |
+
},
|
| 573 |
+
{
|
| 574 |
+
"epoch": 9.204137931034483,
|
| 575 |
+
"grad_norm": 0.15499205887317657,
|
| 576 |
+
"learning_rate": 1.2002442002442003e-05,
|
| 577 |
+
"loss": 0.1222,
|
| 578 |
+
"step": 1675
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 9.34206896551724,
|
| 582 |
+
"grad_norm": 9.006410598754883,
|
| 583 |
+
"learning_rate": 1.1849816849816852e-05,
|
| 584 |
+
"loss": 0.0851,
|
| 585 |
+
"step": 1700
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 9.48,
|
| 589 |
+
"grad_norm": 1.6805857419967651,
|
| 590 |
+
"learning_rate": 1.1697191697191698e-05,
|
| 591 |
+
"loss": 0.0819,
|
| 592 |
+
"step": 1725
|
| 593 |
+
},
|
| 594 |
+
{
|
| 595 |
+
"epoch": 9.617931034482758,
|
| 596 |
+
"grad_norm": 0.058555152267217636,
|
| 597 |
+
"learning_rate": 1.1544566544566545e-05,
|
| 598 |
+
"loss": 0.2012,
|
| 599 |
+
"step": 1750
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 9.755862068965516,
|
| 603 |
+
"grad_norm": 0.14438162744045258,
|
| 604 |
+
"learning_rate": 1.1391941391941392e-05,
|
| 605 |
+
"loss": 0.2202,
|
| 606 |
+
"step": 1775
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"epoch": 9.893793103448276,
|
| 610 |
+
"grad_norm": 0.17045970261096954,
|
| 611 |
+
"learning_rate": 1.123931623931624e-05,
|
| 612 |
+
"loss": 0.2002,
|
| 613 |
+
"step": 1800
|
| 614 |
+
},
|
| 615 |
+
{
|
| 616 |
+
"epoch": 10.0,
|
| 617 |
+
"eval_accuracy": 0.7910685805422647,
|
| 618 |
+
"eval_loss": 1.2801365852355957,
|
| 619 |
+
"eval_macro_f1": 0.7905016517834779,
|
| 620 |
+
"eval_runtime": 1.0838,
|
| 621 |
+
"eval_samples_per_second": 578.51,
|
| 622 |
+
"eval_steps_per_second": 36.907,
|
| 623 |
+
"eval_weighted_f1": 0.7910418809868928,
|
| 624 |
+
"step": 1820
|
| 625 |
+
}
|
| 626 |
+
],
|
| 627 |
+
"logging_steps": 25,
|
| 628 |
+
"max_steps": 3640,
|
| 629 |
+
"num_input_tokens_seen": 0,
|
| 630 |
+
"num_train_epochs": 20,
|
| 631 |
+
"save_steps": 500,
|
| 632 |
+
"stateful_callbacks": {
|
| 633 |
+
"EarlyStoppingCallback": {
|
| 634 |
+
"args": {
|
| 635 |
+
"early_stopping_patience": 3,
|
| 636 |
+
"early_stopping_threshold": 0.0
|
| 637 |
+
},
|
| 638 |
+
"attributes": {
|
| 639 |
+
"early_stopping_patience_counter": 3
|
| 640 |
+
}
|
| 641 |
+
},
|
| 642 |
+
"TrainerControl": {
|
| 643 |
+
"args": {
|
| 644 |
+
"should_epoch_stop": false,
|
| 645 |
+
"should_evaluate": false,
|
| 646 |
+
"should_log": false,
|
| 647 |
+
"should_save": true,
|
| 648 |
+
"should_training_stop": true
|
| 649 |
+
},
|
| 650 |
+
"attributes": {}
|
| 651 |
+
}
|
| 652 |
+
},
|
| 653 |
+
"total_flos": 0.0,
|
| 654 |
+
"train_batch_size": 8,
|
| 655 |
+
"trial_name": null,
|
| 656 |
+
"trial_params": null
|
| 657 |
+
}
|
ood-vifactcheck-bamibert-source-source-01-s44-20260829T015815Z-496d6871/trainer/checkpoint-1026/special_tokens_map.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "<s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "</s>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"mask_token": {
|
| 24 |
+
"content": "<mask>",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "<pad>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "</s>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "<unk>",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": false,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
+
}
|
ood-vifactcheck-bamibert-source-source-01-s44-20260829T015815Z-496d6871/trainer/checkpoint-1026/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-bamibert-source-source-01-s44-20260829T015815Z-496d6871/trainer/checkpoint-1026/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"20480": {
|
| 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": 2048,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "XLMRobertaTokenizerFast",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-bamibert-source-source-01-s44-20260829T015815Z-496d6871/trainer/checkpoint-1026/trainer_state.json
ADDED
|
@@ -0,0 +1,396 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1026,
|
| 3 |
+
"best_metric": 0.8641237155817039,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-bamibert-source-source-01-s44-20260829T015815Z-496d6871/trainer/checkpoint-1026",
|
| 5 |
+
"epoch": 6.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1026,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.14684287812041116,
|
| 14 |
+
"grad_norm": 8.273736000061035,
|
| 15 |
+
"learning_rate": 1.4035087719298246e-06,
|
| 16 |
+
"loss": 4.422,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.2936857562408223,
|
| 21 |
+
"grad_norm": 12.088432312011719,
|
| 22 |
+
"learning_rate": 2.865497076023392e-06,
|
| 23 |
+
"loss": 4.3982,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.44052863436123346,
|
| 28 |
+
"grad_norm": 12.162792205810547,
|
| 29 |
+
"learning_rate": 4.3274853801169596e-06,
|
| 30 |
+
"loss": 4.3863,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5873715124816447,
|
| 35 |
+
"grad_norm": 22.633127212524414,
|
| 36 |
+
"learning_rate": 5.789473684210527e-06,
|
| 37 |
+
"loss": 4.4795,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.7342143906020558,
|
| 42 |
+
"grad_norm": 21.444299697875977,
|
| 43 |
+
"learning_rate": 7.251461988304094e-06,
|
| 44 |
+
"loss": 4.3992,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8810572687224669,
|
| 49 |
+
"grad_norm": 24.679353713989258,
|
| 50 |
+
"learning_rate": 8.713450292397661e-06,
|
| 51 |
+
"loss": 4.3647,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 1.0,
|
| 56 |
+
"eval_accuracy": 0.4288256227758007,
|
| 57 |
+
"eval_loss": 1.0411720275878906,
|
| 58 |
+
"eval_macro_f1": 0.3821193715068592,
|
| 59 |
+
"eval_runtime": 1.0186,
|
| 60 |
+
"eval_samples_per_second": 551.72,
|
| 61 |
+
"eval_steps_per_second": 35.342,
|
| 62 |
+
"eval_weighted_f1": 0.38588009971236087,
|
| 63 |
+
"step": 171
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 1.0234948604992657,
|
| 67 |
+
"grad_norm": 15.236064910888672,
|
| 68 |
+
"learning_rate": 1.017543859649123e-05,
|
| 69 |
+
"loss": 4.2312,
|
| 70 |
+
"step": 175
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.170337738619677,
|
| 74 |
+
"grad_norm": 23.977218627929688,
|
| 75 |
+
"learning_rate": 1.1637426900584797e-05,
|
| 76 |
+
"loss": 4.1405,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.3171806167400881,
|
| 81 |
+
"grad_norm": 40.55796432495117,
|
| 82 |
+
"learning_rate": 1.3099415204678362e-05,
|
| 83 |
+
"loss": 3.4681,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.4640234948604993,
|
| 88 |
+
"grad_norm": 37.805301666259766,
|
| 89 |
+
"learning_rate": 1.4561403508771931e-05,
|
| 90 |
+
"loss": 3.1242,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.6108663729809103,
|
| 95 |
+
"grad_norm": 34.96382141113281,
|
| 96 |
+
"learning_rate": 1.60233918128655e-05,
|
| 97 |
+
"loss": 2.9267,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.7577092511013217,
|
| 102 |
+
"grad_norm": 40.63682174682617,
|
| 103 |
+
"learning_rate": 1.7485380116959064e-05,
|
| 104 |
+
"loss": 2.5733,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.9045521292217327,
|
| 109 |
+
"grad_norm": 37.963436126708984,
|
| 110 |
+
"learning_rate": 1.894736842105263e-05,
|
| 111 |
+
"loss": 2.6616,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 2.0,
|
| 116 |
+
"eval_accuracy": 0.8220640569395018,
|
| 117 |
+
"eval_loss": 0.46709010004997253,
|
| 118 |
+
"eval_macro_f1": 0.8248864312855869,
|
| 119 |
+
"eval_runtime": 1.0579,
|
| 120 |
+
"eval_samples_per_second": 531.265,
|
| 121 |
+
"eval_steps_per_second": 34.031,
|
| 122 |
+
"eval_weighted_f1": 0.8246116663111106,
|
| 123 |
+
"step": 342
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 2.0469897209985315,
|
| 127 |
+
"grad_norm": 68.306884765625,
|
| 128 |
+
"learning_rate": 1.99545159194282e-05,
|
| 129 |
+
"loss": 2.316,
|
| 130 |
+
"step": 350
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.193832599118943,
|
| 134 |
+
"grad_norm": 44.6381950378418,
|
| 135 |
+
"learning_rate": 1.9792072774528916e-05,
|
| 136 |
+
"loss": 2.1868,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.340675477239354,
|
| 141 |
+
"grad_norm": 39.388938903808594,
|
| 142 |
+
"learning_rate": 1.962962962962963e-05,
|
| 143 |
+
"loss": 1.9574,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.4875183553597653,
|
| 148 |
+
"grad_norm": 33.40509796142578,
|
| 149 |
+
"learning_rate": 1.9467186484730346e-05,
|
| 150 |
+
"loss": 1.8035,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.6343612334801763,
|
| 155 |
+
"grad_norm": 72.0527572631836,
|
| 156 |
+
"learning_rate": 1.930474333983106e-05,
|
| 157 |
+
"loss": 1.6116,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.7812041116005872,
|
| 162 |
+
"grad_norm": 34.48876190185547,
|
| 163 |
+
"learning_rate": 1.9142300194931776e-05,
|
| 164 |
+
"loss": 1.8546,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.9280469897209986,
|
| 169 |
+
"grad_norm": 63.485958099365234,
|
| 170 |
+
"learning_rate": 1.897985705003249e-05,
|
| 171 |
+
"loss": 1.9325,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 3.0,
|
| 176 |
+
"eval_accuracy": 0.8327402135231317,
|
| 177 |
+
"eval_loss": 0.46560126543045044,
|
| 178 |
+
"eval_macro_f1": 0.8284610703739349,
|
| 179 |
+
"eval_runtime": 1.1029,
|
| 180 |
+
"eval_samples_per_second": 509.56,
|
| 181 |
+
"eval_steps_per_second": 32.641,
|
| 182 |
+
"eval_weighted_f1": 0.8284524213651301,
|
| 183 |
+
"step": 513
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 3.0704845814977975,
|
| 187 |
+
"grad_norm": 67.29328155517578,
|
| 188 |
+
"learning_rate": 1.8817413905133203e-05,
|
| 189 |
+
"loss": 1.6186,
|
| 190 |
+
"step": 525
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.2173274596182084,
|
| 194 |
+
"grad_norm": 46.54673767089844,
|
| 195 |
+
"learning_rate": 1.8654970760233918e-05,
|
| 196 |
+
"loss": 1.1584,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.36417033773862,
|
| 201 |
+
"grad_norm": 59.3428955078125,
|
| 202 |
+
"learning_rate": 1.8492527615334633e-05,
|
| 203 |
+
"loss": 1.4225,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.511013215859031,
|
| 208 |
+
"grad_norm": 45.025882720947266,
|
| 209 |
+
"learning_rate": 1.833008447043535e-05,
|
| 210 |
+
"loss": 1.1203,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.657856093979442,
|
| 215 |
+
"grad_norm": 55.70996856689453,
|
| 216 |
+
"learning_rate": 1.8167641325536064e-05,
|
| 217 |
+
"loss": 1.2534,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.804698972099853,
|
| 222 |
+
"grad_norm": 50.27566909790039,
|
| 223 |
+
"learning_rate": 1.800519818063678e-05,
|
| 224 |
+
"loss": 1.2485,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.951541850220264,
|
| 229 |
+
"grad_norm": 45.697898864746094,
|
| 230 |
+
"learning_rate": 1.7842755035737494e-05,
|
| 231 |
+
"loss": 1.2695,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 4.0,
|
| 236 |
+
"eval_accuracy": 0.8451957295373665,
|
| 237 |
+
"eval_loss": 0.4495495557785034,
|
| 238 |
+
"eval_macro_f1": 0.8437875605148412,
|
| 239 |
+
"eval_runtime": 1.0865,
|
| 240 |
+
"eval_samples_per_second": 517.245,
|
| 241 |
+
"eval_steps_per_second": 33.133,
|
| 242 |
+
"eval_weighted_f1": 0.8436326979295947,
|
| 243 |
+
"step": 684
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 4.093979441997063,
|
| 247 |
+
"grad_norm": 154.74810791015625,
|
| 248 |
+
"learning_rate": 1.768031189083821e-05,
|
| 249 |
+
"loss": 0.9622,
|
| 250 |
+
"step": 700
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 4.240822320117474,
|
| 254 |
+
"grad_norm": 27.19001007080078,
|
| 255 |
+
"learning_rate": 1.751786874593892e-05,
|
| 256 |
+
"loss": 0.9477,
|
| 257 |
+
"step": 725
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.387665198237886,
|
| 261 |
+
"grad_norm": 43.845211029052734,
|
| 262 |
+
"learning_rate": 1.7355425601039636e-05,
|
| 263 |
+
"loss": 0.8213,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.534508076358296,
|
| 268 |
+
"grad_norm": 43.46588134765625,
|
| 269 |
+
"learning_rate": 1.719298245614035e-05,
|
| 270 |
+
"loss": 0.72,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.681350954478708,
|
| 275 |
+
"grad_norm": 103.8373031616211,
|
| 276 |
+
"learning_rate": 1.7030539311241066e-05,
|
| 277 |
+
"loss": 0.7151,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.828193832599119,
|
| 282 |
+
"grad_norm": 28.66697120666504,
|
| 283 |
+
"learning_rate": 1.686809616634178e-05,
|
| 284 |
+
"loss": 0.9425,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.975036710719531,
|
| 289 |
+
"grad_norm": 17.575389862060547,
|
| 290 |
+
"learning_rate": 1.6705653021442497e-05,
|
| 291 |
+
"loss": 0.7765,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 5.0,
|
| 296 |
+
"eval_accuracy": 0.8540925266903915,
|
| 297 |
+
"eval_loss": 0.4969077408313751,
|
| 298 |
+
"eval_macro_f1": 0.853654548995971,
|
| 299 |
+
"eval_runtime": 1.2382,
|
| 300 |
+
"eval_samples_per_second": 453.884,
|
| 301 |
+
"eval_steps_per_second": 29.074,
|
| 302 |
+
"eval_weighted_f1": 0.8533161289536633,
|
| 303 |
+
"step": 855
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 5.117474302496329,
|
| 307 |
+
"grad_norm": 38.30165481567383,
|
| 308 |
+
"learning_rate": 1.654320987654321e-05,
|
| 309 |
+
"loss": 0.6183,
|
| 310 |
+
"step": 875
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 5.26431718061674,
|
| 314 |
+
"grad_norm": 44.2156982421875,
|
| 315 |
+
"learning_rate": 1.6380766731643927e-05,
|
| 316 |
+
"loss": 0.4517,
|
| 317 |
+
"step": 900
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.411160058737151,
|
| 321 |
+
"grad_norm": 58.340965270996094,
|
| 322 |
+
"learning_rate": 1.621832358674464e-05,
|
| 323 |
+
"loss": 0.6013,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.558002936857562,
|
| 328 |
+
"grad_norm": 17.9237060546875,
|
| 329 |
+
"learning_rate": 1.6055880441845354e-05,
|
| 330 |
+
"loss": 0.6837,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.704845814977974,
|
| 335 |
+
"grad_norm": 70.25029754638672,
|
| 336 |
+
"learning_rate": 1.589343729694607e-05,
|
| 337 |
+
"loss": 0.5012,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.851688693098385,
|
| 342 |
+
"grad_norm": 102.47537231445312,
|
| 343 |
+
"learning_rate": 1.5730994152046787e-05,
|
| 344 |
+
"loss": 0.5637,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.998531571218796,
|
| 349 |
+
"grad_norm": 22.837568283081055,
|
| 350 |
+
"learning_rate": 1.55685510071475e-05,
|
| 351 |
+
"loss": 0.802,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 6.0,
|
| 356 |
+
"eval_accuracy": 0.8647686832740213,
|
| 357 |
+
"eval_loss": 0.5260095000267029,
|
| 358 |
+
"eval_macro_f1": 0.8641237155817039,
|
| 359 |
+
"eval_runtime": 1.0492,
|
| 360 |
+
"eval_samples_per_second": 535.644,
|
| 361 |
+
"eval_steps_per_second": 34.312,
|
| 362 |
+
"eval_weighted_f1": 0.8638597707690122,
|
| 363 |
+
"step": 1026
|
| 364 |
+
}
|
| 365 |
+
],
|
| 366 |
+
"logging_steps": 25,
|
| 367 |
+
"max_steps": 3420,
|
| 368 |
+
"num_input_tokens_seen": 0,
|
| 369 |
+
"num_train_epochs": 20,
|
| 370 |
+
"save_steps": 500,
|
| 371 |
+
"stateful_callbacks": {
|
| 372 |
+
"EarlyStoppingCallback": {
|
| 373 |
+
"args": {
|
| 374 |
+
"early_stopping_patience": 3,
|
| 375 |
+
"early_stopping_threshold": 0.0
|
| 376 |
+
},
|
| 377 |
+
"attributes": {
|
| 378 |
+
"early_stopping_patience_counter": 0
|
| 379 |
+
}
|
| 380 |
+
},
|
| 381 |
+
"TrainerControl": {
|
| 382 |
+
"args": {
|
| 383 |
+
"should_epoch_stop": false,
|
| 384 |
+
"should_evaluate": false,
|
| 385 |
+
"should_log": false,
|
| 386 |
+
"should_save": true,
|
| 387 |
+
"should_training_stop": false
|
| 388 |
+
},
|
| 389 |
+
"attributes": {}
|
| 390 |
+
}
|
| 391 |
+
},
|
| 392 |
+
"total_flos": 0.0,
|
| 393 |
+
"train_batch_size": 8,
|
| 394 |
+
"trial_name": null,
|
| 395 |
+
"trial_params": null
|
| 396 |
+
}
|
ood-vifactcheck-bamibert-source-source-01-s44-20260829T015815Z-496d6871/trainer/checkpoint-1539/special_tokens_map.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "<s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "</s>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"mask_token": {
|
| 24 |
+
"content": "<mask>",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "<pad>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "</s>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "<unk>",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": false,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
+
}
|
ood-vifactcheck-bamibert-source-source-01-s44-20260829T015815Z-496d6871/trainer/checkpoint-1539/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-bamibert-source-source-01-s44-20260829T015815Z-496d6871/trainer/checkpoint-1539/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"20480": {
|
| 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": 2048,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "XLMRobertaTokenizerFast",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-bamibert-source-source-01-s44-20260829T015815Z-496d6871/trainer/checkpoint-1539/trainer_state.json
ADDED
|
@@ -0,0 +1,569 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1026,
|
| 3 |
+
"best_metric": 0.8641237155817039,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-bamibert-source-source-01-s44-20260829T015815Z-496d6871/trainer/checkpoint-1026",
|
| 5 |
+
"epoch": 9.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1539,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.14684287812041116,
|
| 14 |
+
"grad_norm": 8.273736000061035,
|
| 15 |
+
"learning_rate": 1.4035087719298246e-06,
|
| 16 |
+
"loss": 4.422,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.2936857562408223,
|
| 21 |
+
"grad_norm": 12.088432312011719,
|
| 22 |
+
"learning_rate": 2.865497076023392e-06,
|
| 23 |
+
"loss": 4.3982,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.44052863436123346,
|
| 28 |
+
"grad_norm": 12.162792205810547,
|
| 29 |
+
"learning_rate": 4.3274853801169596e-06,
|
| 30 |
+
"loss": 4.3863,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5873715124816447,
|
| 35 |
+
"grad_norm": 22.633127212524414,
|
| 36 |
+
"learning_rate": 5.789473684210527e-06,
|
| 37 |
+
"loss": 4.4795,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.7342143906020558,
|
| 42 |
+
"grad_norm": 21.444299697875977,
|
| 43 |
+
"learning_rate": 7.251461988304094e-06,
|
| 44 |
+
"loss": 4.3992,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8810572687224669,
|
| 49 |
+
"grad_norm": 24.679353713989258,
|
| 50 |
+
"learning_rate": 8.713450292397661e-06,
|
| 51 |
+
"loss": 4.3647,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 1.0,
|
| 56 |
+
"eval_accuracy": 0.4288256227758007,
|
| 57 |
+
"eval_loss": 1.0411720275878906,
|
| 58 |
+
"eval_macro_f1": 0.3821193715068592,
|
| 59 |
+
"eval_runtime": 1.0186,
|
| 60 |
+
"eval_samples_per_second": 551.72,
|
| 61 |
+
"eval_steps_per_second": 35.342,
|
| 62 |
+
"eval_weighted_f1": 0.38588009971236087,
|
| 63 |
+
"step": 171
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 1.0234948604992657,
|
| 67 |
+
"grad_norm": 15.236064910888672,
|
| 68 |
+
"learning_rate": 1.017543859649123e-05,
|
| 69 |
+
"loss": 4.2312,
|
| 70 |
+
"step": 175
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.170337738619677,
|
| 74 |
+
"grad_norm": 23.977218627929688,
|
| 75 |
+
"learning_rate": 1.1637426900584797e-05,
|
| 76 |
+
"loss": 4.1405,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.3171806167400881,
|
| 81 |
+
"grad_norm": 40.55796432495117,
|
| 82 |
+
"learning_rate": 1.3099415204678362e-05,
|
| 83 |
+
"loss": 3.4681,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.4640234948604993,
|
| 88 |
+
"grad_norm": 37.805301666259766,
|
| 89 |
+
"learning_rate": 1.4561403508771931e-05,
|
| 90 |
+
"loss": 3.1242,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.6108663729809103,
|
| 95 |
+
"grad_norm": 34.96382141113281,
|
| 96 |
+
"learning_rate": 1.60233918128655e-05,
|
| 97 |
+
"loss": 2.9267,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.7577092511013217,
|
| 102 |
+
"grad_norm": 40.63682174682617,
|
| 103 |
+
"learning_rate": 1.7485380116959064e-05,
|
| 104 |
+
"loss": 2.5733,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.9045521292217327,
|
| 109 |
+
"grad_norm": 37.963436126708984,
|
| 110 |
+
"learning_rate": 1.894736842105263e-05,
|
| 111 |
+
"loss": 2.6616,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 2.0,
|
| 116 |
+
"eval_accuracy": 0.8220640569395018,
|
| 117 |
+
"eval_loss": 0.46709010004997253,
|
| 118 |
+
"eval_macro_f1": 0.8248864312855869,
|
| 119 |
+
"eval_runtime": 1.0579,
|
| 120 |
+
"eval_samples_per_second": 531.265,
|
| 121 |
+
"eval_steps_per_second": 34.031,
|
| 122 |
+
"eval_weighted_f1": 0.8246116663111106,
|
| 123 |
+
"step": 342
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 2.0469897209985315,
|
| 127 |
+
"grad_norm": 68.306884765625,
|
| 128 |
+
"learning_rate": 1.99545159194282e-05,
|
| 129 |
+
"loss": 2.316,
|
| 130 |
+
"step": 350
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.193832599118943,
|
| 134 |
+
"grad_norm": 44.6381950378418,
|
| 135 |
+
"learning_rate": 1.9792072774528916e-05,
|
| 136 |
+
"loss": 2.1868,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.340675477239354,
|
| 141 |
+
"grad_norm": 39.388938903808594,
|
| 142 |
+
"learning_rate": 1.962962962962963e-05,
|
| 143 |
+
"loss": 1.9574,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.4875183553597653,
|
| 148 |
+
"grad_norm": 33.40509796142578,
|
| 149 |
+
"learning_rate": 1.9467186484730346e-05,
|
| 150 |
+
"loss": 1.8035,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.6343612334801763,
|
| 155 |
+
"grad_norm": 72.0527572631836,
|
| 156 |
+
"learning_rate": 1.930474333983106e-05,
|
| 157 |
+
"loss": 1.6116,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.7812041116005872,
|
| 162 |
+
"grad_norm": 34.48876190185547,
|
| 163 |
+
"learning_rate": 1.9142300194931776e-05,
|
| 164 |
+
"loss": 1.8546,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.9280469897209986,
|
| 169 |
+
"grad_norm": 63.485958099365234,
|
| 170 |
+
"learning_rate": 1.897985705003249e-05,
|
| 171 |
+
"loss": 1.9325,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 3.0,
|
| 176 |
+
"eval_accuracy": 0.8327402135231317,
|
| 177 |
+
"eval_loss": 0.46560126543045044,
|
| 178 |
+
"eval_macro_f1": 0.8284610703739349,
|
| 179 |
+
"eval_runtime": 1.1029,
|
| 180 |
+
"eval_samples_per_second": 509.56,
|
| 181 |
+
"eval_steps_per_second": 32.641,
|
| 182 |
+
"eval_weighted_f1": 0.8284524213651301,
|
| 183 |
+
"step": 513
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 3.0704845814977975,
|
| 187 |
+
"grad_norm": 67.29328155517578,
|
| 188 |
+
"learning_rate": 1.8817413905133203e-05,
|
| 189 |
+
"loss": 1.6186,
|
| 190 |
+
"step": 525
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.2173274596182084,
|
| 194 |
+
"grad_norm": 46.54673767089844,
|
| 195 |
+
"learning_rate": 1.8654970760233918e-05,
|
| 196 |
+
"loss": 1.1584,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.36417033773862,
|
| 201 |
+
"grad_norm": 59.3428955078125,
|
| 202 |
+
"learning_rate": 1.8492527615334633e-05,
|
| 203 |
+
"loss": 1.4225,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.511013215859031,
|
| 208 |
+
"grad_norm": 45.025882720947266,
|
| 209 |
+
"learning_rate": 1.833008447043535e-05,
|
| 210 |
+
"loss": 1.1203,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.657856093979442,
|
| 215 |
+
"grad_norm": 55.70996856689453,
|
| 216 |
+
"learning_rate": 1.8167641325536064e-05,
|
| 217 |
+
"loss": 1.2534,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.804698972099853,
|
| 222 |
+
"grad_norm": 50.27566909790039,
|
| 223 |
+
"learning_rate": 1.800519818063678e-05,
|
| 224 |
+
"loss": 1.2485,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.951541850220264,
|
| 229 |
+
"grad_norm": 45.697898864746094,
|
| 230 |
+
"learning_rate": 1.7842755035737494e-05,
|
| 231 |
+
"loss": 1.2695,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 4.0,
|
| 236 |
+
"eval_accuracy": 0.8451957295373665,
|
| 237 |
+
"eval_loss": 0.4495495557785034,
|
| 238 |
+
"eval_macro_f1": 0.8437875605148412,
|
| 239 |
+
"eval_runtime": 1.0865,
|
| 240 |
+
"eval_samples_per_second": 517.245,
|
| 241 |
+
"eval_steps_per_second": 33.133,
|
| 242 |
+
"eval_weighted_f1": 0.8436326979295947,
|
| 243 |
+
"step": 684
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 4.093979441997063,
|
| 247 |
+
"grad_norm": 154.74810791015625,
|
| 248 |
+
"learning_rate": 1.768031189083821e-05,
|
| 249 |
+
"loss": 0.9622,
|
| 250 |
+
"step": 700
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 4.240822320117474,
|
| 254 |
+
"grad_norm": 27.19001007080078,
|
| 255 |
+
"learning_rate": 1.751786874593892e-05,
|
| 256 |
+
"loss": 0.9477,
|
| 257 |
+
"step": 725
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.387665198237886,
|
| 261 |
+
"grad_norm": 43.845211029052734,
|
| 262 |
+
"learning_rate": 1.7355425601039636e-05,
|
| 263 |
+
"loss": 0.8213,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.534508076358296,
|
| 268 |
+
"grad_norm": 43.46588134765625,
|
| 269 |
+
"learning_rate": 1.719298245614035e-05,
|
| 270 |
+
"loss": 0.72,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.681350954478708,
|
| 275 |
+
"grad_norm": 103.8373031616211,
|
| 276 |
+
"learning_rate": 1.7030539311241066e-05,
|
| 277 |
+
"loss": 0.7151,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.828193832599119,
|
| 282 |
+
"grad_norm": 28.66697120666504,
|
| 283 |
+
"learning_rate": 1.686809616634178e-05,
|
| 284 |
+
"loss": 0.9425,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.975036710719531,
|
| 289 |
+
"grad_norm": 17.575389862060547,
|
| 290 |
+
"learning_rate": 1.6705653021442497e-05,
|
| 291 |
+
"loss": 0.7765,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 5.0,
|
| 296 |
+
"eval_accuracy": 0.8540925266903915,
|
| 297 |
+
"eval_loss": 0.4969077408313751,
|
| 298 |
+
"eval_macro_f1": 0.853654548995971,
|
| 299 |
+
"eval_runtime": 1.2382,
|
| 300 |
+
"eval_samples_per_second": 453.884,
|
| 301 |
+
"eval_steps_per_second": 29.074,
|
| 302 |
+
"eval_weighted_f1": 0.8533161289536633,
|
| 303 |
+
"step": 855
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 5.117474302496329,
|
| 307 |
+
"grad_norm": 38.30165481567383,
|
| 308 |
+
"learning_rate": 1.654320987654321e-05,
|
| 309 |
+
"loss": 0.6183,
|
| 310 |
+
"step": 875
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 5.26431718061674,
|
| 314 |
+
"grad_norm": 44.2156982421875,
|
| 315 |
+
"learning_rate": 1.6380766731643927e-05,
|
| 316 |
+
"loss": 0.4517,
|
| 317 |
+
"step": 900
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.411160058737151,
|
| 321 |
+
"grad_norm": 58.340965270996094,
|
| 322 |
+
"learning_rate": 1.621832358674464e-05,
|
| 323 |
+
"loss": 0.6013,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.558002936857562,
|
| 328 |
+
"grad_norm": 17.9237060546875,
|
| 329 |
+
"learning_rate": 1.6055880441845354e-05,
|
| 330 |
+
"loss": 0.6837,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.704845814977974,
|
| 335 |
+
"grad_norm": 70.25029754638672,
|
| 336 |
+
"learning_rate": 1.589343729694607e-05,
|
| 337 |
+
"loss": 0.5012,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.851688693098385,
|
| 342 |
+
"grad_norm": 102.47537231445312,
|
| 343 |
+
"learning_rate": 1.5730994152046787e-05,
|
| 344 |
+
"loss": 0.5637,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.998531571218796,
|
| 349 |
+
"grad_norm": 22.837568283081055,
|
| 350 |
+
"learning_rate": 1.55685510071475e-05,
|
| 351 |
+
"loss": 0.802,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 6.0,
|
| 356 |
+
"eval_accuracy": 0.8647686832740213,
|
| 357 |
+
"eval_loss": 0.5260095000267029,
|
| 358 |
+
"eval_macro_f1": 0.8641237155817039,
|
| 359 |
+
"eval_runtime": 1.0492,
|
| 360 |
+
"eval_samples_per_second": 535.644,
|
| 361 |
+
"eval_steps_per_second": 34.312,
|
| 362 |
+
"eval_weighted_f1": 0.8638597707690122,
|
| 363 |
+
"step": 1026
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 6.140969162995595,
|
| 367 |
+
"grad_norm": 8.656144142150879,
|
| 368 |
+
"learning_rate": 1.5406107862248214e-05,
|
| 369 |
+
"loss": 0.4042,
|
| 370 |
+
"step": 1050
|
| 371 |
+
},
|
| 372 |
+
{
|
| 373 |
+
"epoch": 6.2878120411160054,
|
| 374 |
+
"grad_norm": 89.54744720458984,
|
| 375 |
+
"learning_rate": 1.524366471734893e-05,
|
| 376 |
+
"loss": 0.5664,
|
| 377 |
+
"step": 1075
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 6.434654919236417,
|
| 381 |
+
"grad_norm": 34.81669616699219,
|
| 382 |
+
"learning_rate": 1.5081221572449645e-05,
|
| 383 |
+
"loss": 0.6061,
|
| 384 |
+
"step": 1100
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 6.581497797356828,
|
| 388 |
+
"grad_norm": 0.9173683524131775,
|
| 389 |
+
"learning_rate": 1.4918778427550358e-05,
|
| 390 |
+
"loss": 0.2905,
|
| 391 |
+
"step": 1125
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 6.72834067547724,
|
| 395 |
+
"grad_norm": 126.47044372558594,
|
| 396 |
+
"learning_rate": 1.4756335282651073e-05,
|
| 397 |
+
"loss": 0.4918,
|
| 398 |
+
"step": 1150
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.87518355359765,
|
| 402 |
+
"grad_norm": 101.17224884033203,
|
| 403 |
+
"learning_rate": 1.4593892137751788e-05,
|
| 404 |
+
"loss": 0.551,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 7.0,
|
| 409 |
+
"eval_accuracy": 0.8558718861209964,
|
| 410 |
+
"eval_loss": 0.6425853371620178,
|
| 411 |
+
"eval_macro_f1": 0.8541409539613731,
|
| 412 |
+
"eval_runtime": 1.0408,
|
| 413 |
+
"eval_samples_per_second": 539.987,
|
| 414 |
+
"eval_steps_per_second": 34.59,
|
| 415 |
+
"eval_weighted_f1": 0.8541780749133843,
|
| 416 |
+
"step": 1197
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 7.0176211453744495,
|
| 420 |
+
"grad_norm": 0.5547441840171814,
|
| 421 |
+
"learning_rate": 1.4431448992852503e-05,
|
| 422 |
+
"loss": 0.4194,
|
| 423 |
+
"step": 1200
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 7.164464023494861,
|
| 427 |
+
"grad_norm": 2.8303141593933105,
|
| 428 |
+
"learning_rate": 1.4269005847953217e-05,
|
| 429 |
+
"loss": 0.4013,
|
| 430 |
+
"step": 1225
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 7.311306901615271,
|
| 434 |
+
"grad_norm": 56.89619445800781,
|
| 435 |
+
"learning_rate": 1.4106562703053932e-05,
|
| 436 |
+
"loss": 0.3428,
|
| 437 |
+
"step": 1250
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 7.458149779735683,
|
| 441 |
+
"grad_norm": 85.9038314819336,
|
| 442 |
+
"learning_rate": 1.3944119558154647e-05,
|
| 443 |
+
"loss": 0.3005,
|
| 444 |
+
"step": 1275
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 7.604992657856094,
|
| 448 |
+
"grad_norm": 86.68646240234375,
|
| 449 |
+
"learning_rate": 1.3781676413255362e-05,
|
| 450 |
+
"loss": 0.2139,
|
| 451 |
+
"step": 1300
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 7.751835535976505,
|
| 455 |
+
"grad_norm": 5.080887794494629,
|
| 456 |
+
"learning_rate": 1.3619233268356076e-05,
|
| 457 |
+
"loss": 0.3366,
|
| 458 |
+
"step": 1325
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 7.898678414096916,
|
| 462 |
+
"grad_norm": 143.64382934570312,
|
| 463 |
+
"learning_rate": 1.345679012345679e-05,
|
| 464 |
+
"loss": 0.519,
|
| 465 |
+
"step": 1350
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 8.0,
|
| 469 |
+
"eval_accuracy": 0.8594306049822064,
|
| 470 |
+
"eval_loss": 0.765890896320343,
|
| 471 |
+
"eval_macro_f1": 0.8581732242538562,
|
| 472 |
+
"eval_runtime": 1.1062,
|
| 473 |
+
"eval_samples_per_second": 508.047,
|
| 474 |
+
"eval_steps_per_second": 32.544,
|
| 475 |
+
"eval_weighted_f1": 0.8579046105349788,
|
| 476 |
+
"step": 1368
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 8.041116005873715,
|
| 480 |
+
"grad_norm": 95.24165344238281,
|
| 481 |
+
"learning_rate": 1.3294346978557506e-05,
|
| 482 |
+
"loss": 0.225,
|
| 483 |
+
"step": 1375
|
| 484 |
+
},
|
| 485 |
+
{
|
| 486 |
+
"epoch": 8.187958883994126,
|
| 487 |
+
"grad_norm": 0.49346044659614563,
|
| 488 |
+
"learning_rate": 1.3131903833658221e-05,
|
| 489 |
+
"loss": 0.1617,
|
| 490 |
+
"step": 1400
|
| 491 |
+
},
|
| 492 |
+
{
|
| 493 |
+
"epoch": 8.334801762114537,
|
| 494 |
+
"grad_norm": 70.15455627441406,
|
| 495 |
+
"learning_rate": 1.2969460688758935e-05,
|
| 496 |
+
"loss": 0.2043,
|
| 497 |
+
"step": 1425
|
| 498 |
+
},
|
| 499 |
+
{
|
| 500 |
+
"epoch": 8.481644640234949,
|
| 501 |
+
"grad_norm": 98.88150024414062,
|
| 502 |
+
"learning_rate": 1.280701754385965e-05,
|
| 503 |
+
"loss": 0.2602,
|
| 504 |
+
"step": 1450
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 8.62848751835536,
|
| 508 |
+
"grad_norm": 102.87335205078125,
|
| 509 |
+
"learning_rate": 1.2644574398960365e-05,
|
| 510 |
+
"loss": 0.3087,
|
| 511 |
+
"step": 1475
|
| 512 |
+
},
|
| 513 |
+
{
|
| 514 |
+
"epoch": 8.775330396475772,
|
| 515 |
+
"grad_norm": 18.848819732666016,
|
| 516 |
+
"learning_rate": 1.248213125406108e-05,
|
| 517 |
+
"loss": 0.2501,
|
| 518 |
+
"step": 1500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 8.922173274596181,
|
| 522 |
+
"grad_norm": 46.649505615234375,
|
| 523 |
+
"learning_rate": 1.2319688109161793e-05,
|
| 524 |
+
"loss": 0.3888,
|
| 525 |
+
"step": 1525
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 9.0,
|
| 529 |
+
"eval_accuracy": 0.8487544483985765,
|
| 530 |
+
"eval_loss": 0.8159970045089722,
|
| 531 |
+
"eval_macro_f1": 0.8485051429087919,
|
| 532 |
+
"eval_runtime": 1.2154,
|
| 533 |
+
"eval_samples_per_second": 462.4,
|
| 534 |
+
"eval_steps_per_second": 29.62,
|
| 535 |
+
"eval_weighted_f1": 0.848345048961446,
|
| 536 |
+
"step": 1539
|
| 537 |
+
}
|
| 538 |
+
],
|
| 539 |
+
"logging_steps": 25,
|
| 540 |
+
"max_steps": 3420,
|
| 541 |
+
"num_input_tokens_seen": 0,
|
| 542 |
+
"num_train_epochs": 20,
|
| 543 |
+
"save_steps": 500,
|
| 544 |
+
"stateful_callbacks": {
|
| 545 |
+
"EarlyStoppingCallback": {
|
| 546 |
+
"args": {
|
| 547 |
+
"early_stopping_patience": 3,
|
| 548 |
+
"early_stopping_threshold": 0.0
|
| 549 |
+
},
|
| 550 |
+
"attributes": {
|
| 551 |
+
"early_stopping_patience_counter": 3
|
| 552 |
+
}
|
| 553 |
+
},
|
| 554 |
+
"TrainerControl": {
|
| 555 |
+
"args": {
|
| 556 |
+
"should_epoch_stop": false,
|
| 557 |
+
"should_evaluate": false,
|
| 558 |
+
"should_log": false,
|
| 559 |
+
"should_save": true,
|
| 560 |
+
"should_training_stop": true
|
| 561 |
+
},
|
| 562 |
+
"attributes": {}
|
| 563 |
+
}
|
| 564 |
+
},
|
| 565 |
+
"total_flos": 0.0,
|
| 566 |
+
"train_batch_size": 8,
|
| 567 |
+
"trial_name": null,
|
| 568 |
+
"trial_params": null
|
| 569 |
+
}
|
ood-vifactcheck-bamibert-source-source-02-s44-20260829T020641Z-496d6871/trainer/checkpoint-1086/special_tokens_map.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "<s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "</s>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"mask_token": {
|
| 24 |
+
"content": "<mask>",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "<pad>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "</s>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "<unk>",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": false,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
+
}
|
ood-vifactcheck-bamibert-source-source-02-s44-20260829T020641Z-496d6871/trainer/checkpoint-1086/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-bamibert-source-source-02-s44-20260829T020641Z-496d6871/trainer/checkpoint-1086/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"20480": {
|
| 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": 2048,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "XLMRobertaTokenizerFast",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-bamibert-source-source-02-s44-20260829T020641Z-496d6871/trainer/checkpoint-1086/trainer_state.json
ADDED
|
@@ -0,0 +1,410 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1086,
|
| 3 |
+
"best_metric": 0.8228459689277442,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-bamibert-source-source-02-s44-20260829T020641Z-496d6871/trainer/checkpoint-1086",
|
| 5 |
+
"epoch": 6.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1086,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.13812154696132597,
|
| 14 |
+
"grad_norm": 8.378602981567383,
|
| 15 |
+
"learning_rate": 1.3259668508287293e-06,
|
| 16 |
+
"loss": 4.4221,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.27624309392265195,
|
| 21 |
+
"grad_norm": 9.934577941894531,
|
| 22 |
+
"learning_rate": 2.707182320441989e-06,
|
| 23 |
+
"loss": 4.4009,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.4143646408839779,
|
| 28 |
+
"grad_norm": 25.300434112548828,
|
| 29 |
+
"learning_rate": 4.088397790055249e-06,
|
| 30 |
+
"loss": 4.385,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5524861878453039,
|
| 35 |
+
"grad_norm": 15.242926597595215,
|
| 36 |
+
"learning_rate": 5.469613259668509e-06,
|
| 37 |
+
"loss": 4.3731,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6906077348066298,
|
| 42 |
+
"grad_norm": 17.174333572387695,
|
| 43 |
+
"learning_rate": 6.850828729281769e-06,
|
| 44 |
+
"loss": 4.4033,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8287292817679558,
|
| 49 |
+
"grad_norm": 64.0993423461914,
|
| 50 |
+
"learning_rate": 8.232044198895029e-06,
|
| 51 |
+
"loss": 4.3248,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9668508287292817,
|
| 56 |
+
"grad_norm": 26.487071990966797,
|
| 57 |
+
"learning_rate": 9.613259668508288e-06,
|
| 58 |
+
"loss": 4.3959,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.43470790378006874,
|
| 64 |
+
"eval_loss": 1.0628154277801514,
|
| 65 |
+
"eval_macro_f1": 0.37803412350196014,
|
| 66 |
+
"eval_runtime": 1.0213,
|
| 67 |
+
"eval_samples_per_second": 569.885,
|
| 68 |
+
"eval_steps_per_second": 36.23,
|
| 69 |
+
"eval_weighted_f1": 0.3846268801016823,
|
| 70 |
+
"step": 181
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.1049723756906078,
|
| 74 |
+
"grad_norm": 40.06175231933594,
|
| 75 |
+
"learning_rate": 1.0994475138121547e-05,
|
| 76 |
+
"loss": 4.3543,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.2430939226519337,
|
| 81 |
+
"grad_norm": 13.009339332580566,
|
| 82 |
+
"learning_rate": 1.2375690607734806e-05,
|
| 83 |
+
"loss": 4.3613,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.3812154696132597,
|
| 88 |
+
"grad_norm": 23.42742347717285,
|
| 89 |
+
"learning_rate": 1.3756906077348067e-05,
|
| 90 |
+
"loss": 4.3199,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.5193370165745856,
|
| 95 |
+
"grad_norm": 23.87099838256836,
|
| 96 |
+
"learning_rate": 1.5138121546961326e-05,
|
| 97 |
+
"loss": 4.0754,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.6574585635359116,
|
| 102 |
+
"grad_norm": 50.17503356933594,
|
| 103 |
+
"learning_rate": 1.6519337016574586e-05,
|
| 104 |
+
"loss": 3.4262,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7955801104972375,
|
| 109 |
+
"grad_norm": 74.37056732177734,
|
| 110 |
+
"learning_rate": 1.7900552486187846e-05,
|
| 111 |
+
"loss": 3.0862,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.9337016574585635,
|
| 116 |
+
"grad_norm": 52.36674118041992,
|
| 117 |
+
"learning_rate": 1.9281767955801107e-05,
|
| 118 |
+
"loss": 2.9627,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.7250859106529209,
|
| 124 |
+
"eval_loss": 0.6534662842750549,
|
| 125 |
+
"eval_macro_f1": 0.7189938408771342,
|
| 126 |
+
"eval_runtime": 1.043,
|
| 127 |
+
"eval_samples_per_second": 558.031,
|
| 128 |
+
"eval_steps_per_second": 35.476,
|
| 129 |
+
"eval_weighted_f1": 0.7210239812769866,
|
| 130 |
+
"step": 362
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.0718232044198897,
|
| 134 |
+
"grad_norm": 58.514095306396484,
|
| 135 |
+
"learning_rate": 1.9926335174953962e-05,
|
| 136 |
+
"loss": 2.6949,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.2099447513812156,
|
| 141 |
+
"grad_norm": 58.527103424072266,
|
| 142 |
+
"learning_rate": 1.9772866789441377e-05,
|
| 143 |
+
"loss": 2.2771,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.3480662983425415,
|
| 148 |
+
"grad_norm": 57.325462341308594,
|
| 149 |
+
"learning_rate": 1.9619398403928794e-05,
|
| 150 |
+
"loss": 2.3478,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.4861878453038675,
|
| 155 |
+
"grad_norm": 42.870723724365234,
|
| 156 |
+
"learning_rate": 1.946593001841621e-05,
|
| 157 |
+
"loss": 2.2489,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.6243093922651934,
|
| 162 |
+
"grad_norm": 56.36940383911133,
|
| 163 |
+
"learning_rate": 1.9312461632903623e-05,
|
| 164 |
+
"loss": 2.2231,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.7624309392265194,
|
| 169 |
+
"grad_norm": 68.46783447265625,
|
| 170 |
+
"learning_rate": 1.915899324739104e-05,
|
| 171 |
+
"loss": 2.0115,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.9005524861878453,
|
| 176 |
+
"grad_norm": 31.32921600341797,
|
| 177 |
+
"learning_rate": 1.9005524861878455e-05,
|
| 178 |
+
"loss": 2.1916,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 3.0,
|
| 183 |
+
"eval_accuracy": 0.7714776632302406,
|
| 184 |
+
"eval_loss": 0.6084756255149841,
|
| 185 |
+
"eval_macro_f1": 0.770967523163708,
|
| 186 |
+
"eval_runtime": 1.0515,
|
| 187 |
+
"eval_samples_per_second": 553.512,
|
| 188 |
+
"eval_steps_per_second": 35.189,
|
| 189 |
+
"eval_weighted_f1": 0.7708288136356287,
|
| 190 |
+
"step": 543
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.0386740331491713,
|
| 194 |
+
"grad_norm": 38.462303161621094,
|
| 195 |
+
"learning_rate": 1.885205647636587e-05,
|
| 196 |
+
"loss": 1.8529,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.1767955801104972,
|
| 201 |
+
"grad_norm": 54.62997055053711,
|
| 202 |
+
"learning_rate": 1.8698588090853287e-05,
|
| 203 |
+
"loss": 1.4638,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.314917127071823,
|
| 208 |
+
"grad_norm": 52.252376556396484,
|
| 209 |
+
"learning_rate": 1.85451197053407e-05,
|
| 210 |
+
"loss": 1.3945,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.453038674033149,
|
| 215 |
+
"grad_norm": 25.21811866760254,
|
| 216 |
+
"learning_rate": 1.8391651319828116e-05,
|
| 217 |
+
"loss": 1.49,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.591160220994475,
|
| 222 |
+
"grad_norm": 65.29776000976562,
|
| 223 |
+
"learning_rate": 1.823818293431553e-05,
|
| 224 |
+
"loss": 1.3792,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.729281767955801,
|
| 229 |
+
"grad_norm": 54.638092041015625,
|
| 230 |
+
"learning_rate": 1.8084714548802948e-05,
|
| 231 |
+
"loss": 1.3482,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.867403314917127,
|
| 236 |
+
"grad_norm": 71.57609558105469,
|
| 237 |
+
"learning_rate": 1.7931246163290362e-05,
|
| 238 |
+
"loss": 1.7038,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 4.0,
|
| 243 |
+
"eval_accuracy": 0.7989690721649485,
|
| 244 |
+
"eval_loss": 0.5463073253631592,
|
| 245 |
+
"eval_macro_f1": 0.7996686705393685,
|
| 246 |
+
"eval_runtime": 1.3303,
|
| 247 |
+
"eval_samples_per_second": 437.51,
|
| 248 |
+
"eval_steps_per_second": 27.814,
|
| 249 |
+
"eval_weighted_f1": 0.8003280157649901,
|
| 250 |
+
"step": 724
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 4.005524861878453,
|
| 254 |
+
"grad_norm": 38.35980224609375,
|
| 255 |
+
"learning_rate": 1.7777777777777777e-05,
|
| 256 |
+
"loss": 1.3231,
|
| 257 |
+
"step": 725
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.143646408839779,
|
| 261 |
+
"grad_norm": 58.87895584106445,
|
| 262 |
+
"learning_rate": 1.7624309392265194e-05,
|
| 263 |
+
"loss": 1.0074,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.281767955801105,
|
| 268 |
+
"grad_norm": 48.42424011230469,
|
| 269 |
+
"learning_rate": 1.747084100675261e-05,
|
| 270 |
+
"loss": 0.8949,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.419889502762431,
|
| 275 |
+
"grad_norm": 49.594940185546875,
|
| 276 |
+
"learning_rate": 1.7317372621240026e-05,
|
| 277 |
+
"loss": 1.1707,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.558011049723757,
|
| 282 |
+
"grad_norm": 41.483360290527344,
|
| 283 |
+
"learning_rate": 1.716390423572744e-05,
|
| 284 |
+
"loss": 1.1846,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.696132596685083,
|
| 289 |
+
"grad_norm": 38.58932876586914,
|
| 290 |
+
"learning_rate": 1.701043585021486e-05,
|
| 291 |
+
"loss": 0.9835,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.834254143646409,
|
| 296 |
+
"grad_norm": 22.224456787109375,
|
| 297 |
+
"learning_rate": 1.6856967464702273e-05,
|
| 298 |
+
"loss": 1.003,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.972375690607735,
|
| 303 |
+
"grad_norm": 45.71381378173828,
|
| 304 |
+
"learning_rate": 1.670349907918969e-05,
|
| 305 |
+
"loss": 0.9936,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 5.0,
|
| 310 |
+
"eval_accuracy": 0.8075601374570447,
|
| 311 |
+
"eval_loss": 0.6210135817527771,
|
| 312 |
+
"eval_macro_f1": 0.8058015769195027,
|
| 313 |
+
"eval_runtime": 1.0733,
|
| 314 |
+
"eval_samples_per_second": 542.257,
|
| 315 |
+
"eval_steps_per_second": 34.473,
|
| 316 |
+
"eval_weighted_f1": 0.8064412410634019,
|
| 317 |
+
"step": 905
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.110497237569061,
|
| 321 |
+
"grad_norm": 28.920438766479492,
|
| 322 |
+
"learning_rate": 1.6550030693677105e-05,
|
| 323 |
+
"loss": 0.7549,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.248618784530387,
|
| 328 |
+
"grad_norm": 36.169921875,
|
| 329 |
+
"learning_rate": 1.639656230816452e-05,
|
| 330 |
+
"loss": 0.5909,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.386740331491713,
|
| 335 |
+
"grad_norm": 71.56083679199219,
|
| 336 |
+
"learning_rate": 1.6243093922651937e-05,
|
| 337 |
+
"loss": 0.7266,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.524861878453039,
|
| 342 |
+
"grad_norm": 62.731964111328125,
|
| 343 |
+
"learning_rate": 1.608962553713935e-05,
|
| 344 |
+
"loss": 0.6539,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.662983425414365,
|
| 349 |
+
"grad_norm": 38.15754699707031,
|
| 350 |
+
"learning_rate": 1.5936157151626766e-05,
|
| 351 |
+
"loss": 0.6357,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.801104972375691,
|
| 356 |
+
"grad_norm": 14.287630081176758,
|
| 357 |
+
"learning_rate": 1.5782688766114183e-05,
|
| 358 |
+
"loss": 0.6742,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.939226519337017,
|
| 363 |
+
"grad_norm": 85.8550033569336,
|
| 364 |
+
"learning_rate": 1.5629220380601598e-05,
|
| 365 |
+
"loss": 0.8734,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 6.0,
|
| 370 |
+
"eval_accuracy": 0.8230240549828178,
|
| 371 |
+
"eval_loss": 0.6619381308555603,
|
| 372 |
+
"eval_macro_f1": 0.8228459689277442,
|
| 373 |
+
"eval_runtime": 1.1018,
|
| 374 |
+
"eval_samples_per_second": 528.206,
|
| 375 |
+
"eval_steps_per_second": 33.58,
|
| 376 |
+
"eval_weighted_f1": 0.8229351102028948,
|
| 377 |
+
"step": 1086
|
| 378 |
+
}
|
| 379 |
+
],
|
| 380 |
+
"logging_steps": 25,
|
| 381 |
+
"max_steps": 3620,
|
| 382 |
+
"num_input_tokens_seen": 0,
|
| 383 |
+
"num_train_epochs": 20,
|
| 384 |
+
"save_steps": 500,
|
| 385 |
+
"stateful_callbacks": {
|
| 386 |
+
"EarlyStoppingCallback": {
|
| 387 |
+
"args": {
|
| 388 |
+
"early_stopping_patience": 3,
|
| 389 |
+
"early_stopping_threshold": 0.0
|
| 390 |
+
},
|
| 391 |
+
"attributes": {
|
| 392 |
+
"early_stopping_patience_counter": 0
|
| 393 |
+
}
|
| 394 |
+
},
|
| 395 |
+
"TrainerControl": {
|
| 396 |
+
"args": {
|
| 397 |
+
"should_epoch_stop": false,
|
| 398 |
+
"should_evaluate": false,
|
| 399 |
+
"should_log": false,
|
| 400 |
+
"should_save": true,
|
| 401 |
+
"should_training_stop": false
|
| 402 |
+
},
|
| 403 |
+
"attributes": {}
|
| 404 |
+
}
|
| 405 |
+
},
|
| 406 |
+
"total_flos": 0.0,
|
| 407 |
+
"train_batch_size": 8,
|
| 408 |
+
"trial_name": null,
|
| 409 |
+
"trial_params": null
|
| 410 |
+
}
|
ood-vifactcheck-bamibert-source-source-02-s44-20260829T020641Z-496d6871/trainer/checkpoint-1629/special_tokens_map.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "<s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "</s>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"mask_token": {
|
| 24 |
+
"content": "<mask>",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "<pad>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "</s>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "<unk>",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": false,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
+
}
|
ood-vifactcheck-bamibert-source-source-02-s44-20260829T020641Z-496d6871/trainer/checkpoint-1629/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-bamibert-source-source-02-s44-20260829T020641Z-496d6871/trainer/checkpoint-1629/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"20480": {
|
| 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": 2048,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "XLMRobertaTokenizerFast",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-bamibert-source-source-02-s44-20260829T020641Z-496d6871/trainer/checkpoint-1629/trainer_state.json
ADDED
|
@@ -0,0 +1,597 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1086,
|
| 3 |
+
"best_metric": 0.8228459689277442,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-bamibert-source-source-02-s44-20260829T020641Z-496d6871/trainer/checkpoint-1086",
|
| 5 |
+
"epoch": 9.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1629,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.13812154696132597,
|
| 14 |
+
"grad_norm": 8.378602981567383,
|
| 15 |
+
"learning_rate": 1.3259668508287293e-06,
|
| 16 |
+
"loss": 4.4221,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.27624309392265195,
|
| 21 |
+
"grad_norm": 9.934577941894531,
|
| 22 |
+
"learning_rate": 2.707182320441989e-06,
|
| 23 |
+
"loss": 4.4009,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.4143646408839779,
|
| 28 |
+
"grad_norm": 25.300434112548828,
|
| 29 |
+
"learning_rate": 4.088397790055249e-06,
|
| 30 |
+
"loss": 4.385,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5524861878453039,
|
| 35 |
+
"grad_norm": 15.242926597595215,
|
| 36 |
+
"learning_rate": 5.469613259668509e-06,
|
| 37 |
+
"loss": 4.3731,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6906077348066298,
|
| 42 |
+
"grad_norm": 17.174333572387695,
|
| 43 |
+
"learning_rate": 6.850828729281769e-06,
|
| 44 |
+
"loss": 4.4033,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8287292817679558,
|
| 49 |
+
"grad_norm": 64.0993423461914,
|
| 50 |
+
"learning_rate": 8.232044198895029e-06,
|
| 51 |
+
"loss": 4.3248,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9668508287292817,
|
| 56 |
+
"grad_norm": 26.487071990966797,
|
| 57 |
+
"learning_rate": 9.613259668508288e-06,
|
| 58 |
+
"loss": 4.3959,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.43470790378006874,
|
| 64 |
+
"eval_loss": 1.0628154277801514,
|
| 65 |
+
"eval_macro_f1": 0.37803412350196014,
|
| 66 |
+
"eval_runtime": 1.0213,
|
| 67 |
+
"eval_samples_per_second": 569.885,
|
| 68 |
+
"eval_steps_per_second": 36.23,
|
| 69 |
+
"eval_weighted_f1": 0.3846268801016823,
|
| 70 |
+
"step": 181
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.1049723756906078,
|
| 74 |
+
"grad_norm": 40.06175231933594,
|
| 75 |
+
"learning_rate": 1.0994475138121547e-05,
|
| 76 |
+
"loss": 4.3543,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.2430939226519337,
|
| 81 |
+
"grad_norm": 13.009339332580566,
|
| 82 |
+
"learning_rate": 1.2375690607734806e-05,
|
| 83 |
+
"loss": 4.3613,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.3812154696132597,
|
| 88 |
+
"grad_norm": 23.42742347717285,
|
| 89 |
+
"learning_rate": 1.3756906077348067e-05,
|
| 90 |
+
"loss": 4.3199,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.5193370165745856,
|
| 95 |
+
"grad_norm": 23.87099838256836,
|
| 96 |
+
"learning_rate": 1.5138121546961326e-05,
|
| 97 |
+
"loss": 4.0754,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.6574585635359116,
|
| 102 |
+
"grad_norm": 50.17503356933594,
|
| 103 |
+
"learning_rate": 1.6519337016574586e-05,
|
| 104 |
+
"loss": 3.4262,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7955801104972375,
|
| 109 |
+
"grad_norm": 74.37056732177734,
|
| 110 |
+
"learning_rate": 1.7900552486187846e-05,
|
| 111 |
+
"loss": 3.0862,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.9337016574585635,
|
| 116 |
+
"grad_norm": 52.36674118041992,
|
| 117 |
+
"learning_rate": 1.9281767955801107e-05,
|
| 118 |
+
"loss": 2.9627,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.7250859106529209,
|
| 124 |
+
"eval_loss": 0.6534662842750549,
|
| 125 |
+
"eval_macro_f1": 0.7189938408771342,
|
| 126 |
+
"eval_runtime": 1.043,
|
| 127 |
+
"eval_samples_per_second": 558.031,
|
| 128 |
+
"eval_steps_per_second": 35.476,
|
| 129 |
+
"eval_weighted_f1": 0.7210239812769866,
|
| 130 |
+
"step": 362
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.0718232044198897,
|
| 134 |
+
"grad_norm": 58.514095306396484,
|
| 135 |
+
"learning_rate": 1.9926335174953962e-05,
|
| 136 |
+
"loss": 2.6949,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.2099447513812156,
|
| 141 |
+
"grad_norm": 58.527103424072266,
|
| 142 |
+
"learning_rate": 1.9772866789441377e-05,
|
| 143 |
+
"loss": 2.2771,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.3480662983425415,
|
| 148 |
+
"grad_norm": 57.325462341308594,
|
| 149 |
+
"learning_rate": 1.9619398403928794e-05,
|
| 150 |
+
"loss": 2.3478,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.4861878453038675,
|
| 155 |
+
"grad_norm": 42.870723724365234,
|
| 156 |
+
"learning_rate": 1.946593001841621e-05,
|
| 157 |
+
"loss": 2.2489,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.6243093922651934,
|
| 162 |
+
"grad_norm": 56.36940383911133,
|
| 163 |
+
"learning_rate": 1.9312461632903623e-05,
|
| 164 |
+
"loss": 2.2231,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.7624309392265194,
|
| 169 |
+
"grad_norm": 68.46783447265625,
|
| 170 |
+
"learning_rate": 1.915899324739104e-05,
|
| 171 |
+
"loss": 2.0115,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.9005524861878453,
|
| 176 |
+
"grad_norm": 31.32921600341797,
|
| 177 |
+
"learning_rate": 1.9005524861878455e-05,
|
| 178 |
+
"loss": 2.1916,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 3.0,
|
| 183 |
+
"eval_accuracy": 0.7714776632302406,
|
| 184 |
+
"eval_loss": 0.6084756255149841,
|
| 185 |
+
"eval_macro_f1": 0.770967523163708,
|
| 186 |
+
"eval_runtime": 1.0515,
|
| 187 |
+
"eval_samples_per_second": 553.512,
|
| 188 |
+
"eval_steps_per_second": 35.189,
|
| 189 |
+
"eval_weighted_f1": 0.7708288136356287,
|
| 190 |
+
"step": 543
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.0386740331491713,
|
| 194 |
+
"grad_norm": 38.462303161621094,
|
| 195 |
+
"learning_rate": 1.885205647636587e-05,
|
| 196 |
+
"loss": 1.8529,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.1767955801104972,
|
| 201 |
+
"grad_norm": 54.62997055053711,
|
| 202 |
+
"learning_rate": 1.8698588090853287e-05,
|
| 203 |
+
"loss": 1.4638,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.314917127071823,
|
| 208 |
+
"grad_norm": 52.252376556396484,
|
| 209 |
+
"learning_rate": 1.85451197053407e-05,
|
| 210 |
+
"loss": 1.3945,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.453038674033149,
|
| 215 |
+
"grad_norm": 25.21811866760254,
|
| 216 |
+
"learning_rate": 1.8391651319828116e-05,
|
| 217 |
+
"loss": 1.49,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.591160220994475,
|
| 222 |
+
"grad_norm": 65.29776000976562,
|
| 223 |
+
"learning_rate": 1.823818293431553e-05,
|
| 224 |
+
"loss": 1.3792,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.729281767955801,
|
| 229 |
+
"grad_norm": 54.638092041015625,
|
| 230 |
+
"learning_rate": 1.8084714548802948e-05,
|
| 231 |
+
"loss": 1.3482,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.867403314917127,
|
| 236 |
+
"grad_norm": 71.57609558105469,
|
| 237 |
+
"learning_rate": 1.7931246163290362e-05,
|
| 238 |
+
"loss": 1.7038,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 4.0,
|
| 243 |
+
"eval_accuracy": 0.7989690721649485,
|
| 244 |
+
"eval_loss": 0.5463073253631592,
|
| 245 |
+
"eval_macro_f1": 0.7996686705393685,
|
| 246 |
+
"eval_runtime": 1.3303,
|
| 247 |
+
"eval_samples_per_second": 437.51,
|
| 248 |
+
"eval_steps_per_second": 27.814,
|
| 249 |
+
"eval_weighted_f1": 0.8003280157649901,
|
| 250 |
+
"step": 724
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 4.005524861878453,
|
| 254 |
+
"grad_norm": 38.35980224609375,
|
| 255 |
+
"learning_rate": 1.7777777777777777e-05,
|
| 256 |
+
"loss": 1.3231,
|
| 257 |
+
"step": 725
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.143646408839779,
|
| 261 |
+
"grad_norm": 58.87895584106445,
|
| 262 |
+
"learning_rate": 1.7624309392265194e-05,
|
| 263 |
+
"loss": 1.0074,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.281767955801105,
|
| 268 |
+
"grad_norm": 48.42424011230469,
|
| 269 |
+
"learning_rate": 1.747084100675261e-05,
|
| 270 |
+
"loss": 0.8949,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.419889502762431,
|
| 275 |
+
"grad_norm": 49.594940185546875,
|
| 276 |
+
"learning_rate": 1.7317372621240026e-05,
|
| 277 |
+
"loss": 1.1707,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.558011049723757,
|
| 282 |
+
"grad_norm": 41.483360290527344,
|
| 283 |
+
"learning_rate": 1.716390423572744e-05,
|
| 284 |
+
"loss": 1.1846,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.696132596685083,
|
| 289 |
+
"grad_norm": 38.58932876586914,
|
| 290 |
+
"learning_rate": 1.701043585021486e-05,
|
| 291 |
+
"loss": 0.9835,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.834254143646409,
|
| 296 |
+
"grad_norm": 22.224456787109375,
|
| 297 |
+
"learning_rate": 1.6856967464702273e-05,
|
| 298 |
+
"loss": 1.003,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.972375690607735,
|
| 303 |
+
"grad_norm": 45.71381378173828,
|
| 304 |
+
"learning_rate": 1.670349907918969e-05,
|
| 305 |
+
"loss": 0.9936,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 5.0,
|
| 310 |
+
"eval_accuracy": 0.8075601374570447,
|
| 311 |
+
"eval_loss": 0.6210135817527771,
|
| 312 |
+
"eval_macro_f1": 0.8058015769195027,
|
| 313 |
+
"eval_runtime": 1.0733,
|
| 314 |
+
"eval_samples_per_second": 542.257,
|
| 315 |
+
"eval_steps_per_second": 34.473,
|
| 316 |
+
"eval_weighted_f1": 0.8064412410634019,
|
| 317 |
+
"step": 905
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.110497237569061,
|
| 321 |
+
"grad_norm": 28.920438766479492,
|
| 322 |
+
"learning_rate": 1.6550030693677105e-05,
|
| 323 |
+
"loss": 0.7549,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.248618784530387,
|
| 328 |
+
"grad_norm": 36.169921875,
|
| 329 |
+
"learning_rate": 1.639656230816452e-05,
|
| 330 |
+
"loss": 0.5909,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.386740331491713,
|
| 335 |
+
"grad_norm": 71.56083679199219,
|
| 336 |
+
"learning_rate": 1.6243093922651937e-05,
|
| 337 |
+
"loss": 0.7266,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.524861878453039,
|
| 342 |
+
"grad_norm": 62.731964111328125,
|
| 343 |
+
"learning_rate": 1.608962553713935e-05,
|
| 344 |
+
"loss": 0.6539,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.662983425414365,
|
| 349 |
+
"grad_norm": 38.15754699707031,
|
| 350 |
+
"learning_rate": 1.5936157151626766e-05,
|
| 351 |
+
"loss": 0.6357,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.801104972375691,
|
| 356 |
+
"grad_norm": 14.287630081176758,
|
| 357 |
+
"learning_rate": 1.5782688766114183e-05,
|
| 358 |
+
"loss": 0.6742,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.939226519337017,
|
| 363 |
+
"grad_norm": 85.8550033569336,
|
| 364 |
+
"learning_rate": 1.5629220380601598e-05,
|
| 365 |
+
"loss": 0.8734,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 6.0,
|
| 370 |
+
"eval_accuracy": 0.8230240549828178,
|
| 371 |
+
"eval_loss": 0.6619381308555603,
|
| 372 |
+
"eval_macro_f1": 0.8228459689277442,
|
| 373 |
+
"eval_runtime": 1.1018,
|
| 374 |
+
"eval_samples_per_second": 528.206,
|
| 375 |
+
"eval_steps_per_second": 33.58,
|
| 376 |
+
"eval_weighted_f1": 0.8229351102028948,
|
| 377 |
+
"step": 1086
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 6.077348066298343,
|
| 381 |
+
"grad_norm": 226.38731384277344,
|
| 382 |
+
"learning_rate": 1.5475751995089012e-05,
|
| 383 |
+
"loss": 0.4655,
|
| 384 |
+
"step": 1100
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 6.2154696132596685,
|
| 388 |
+
"grad_norm": 3.9470162391662598,
|
| 389 |
+
"learning_rate": 1.532228360957643e-05,
|
| 390 |
+
"loss": 0.4045,
|
| 391 |
+
"step": 1125
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 6.3535911602209945,
|
| 395 |
+
"grad_norm": 67.44049072265625,
|
| 396 |
+
"learning_rate": 1.5168815224063844e-05,
|
| 397 |
+
"loss": 0.4966,
|
| 398 |
+
"step": 1150
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.49171270718232,
|
| 402 |
+
"grad_norm": 80.22216796875,
|
| 403 |
+
"learning_rate": 1.5015346838551258e-05,
|
| 404 |
+
"loss": 0.5009,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.629834254143646,
|
| 409 |
+
"grad_norm": 3.586699962615967,
|
| 410 |
+
"learning_rate": 1.4861878453038676e-05,
|
| 411 |
+
"loss": 0.4624,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.767955801104972,
|
| 416 |
+
"grad_norm": 6.727713108062744,
|
| 417 |
+
"learning_rate": 1.470841006752609e-05,
|
| 418 |
+
"loss": 0.4791,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.906077348066298,
|
| 423 |
+
"grad_norm": 88.3924789428711,
|
| 424 |
+
"learning_rate": 1.4554941682013506e-05,
|
| 425 |
+
"loss": 0.626,
|
| 426 |
+
"step": 1250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 7.0,
|
| 430 |
+
"eval_accuracy": 0.8006872852233677,
|
| 431 |
+
"eval_loss": 0.8543733954429626,
|
| 432 |
+
"eval_macro_f1": 0.8004268395757309,
|
| 433 |
+
"eval_runtime": 1.2574,
|
| 434 |
+
"eval_samples_per_second": 462.877,
|
| 435 |
+
"eval_steps_per_second": 29.427,
|
| 436 |
+
"eval_weighted_f1": 0.8005602318558923,
|
| 437 |
+
"step": 1267
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 7.044198895027624,
|
| 441 |
+
"grad_norm": 110.76263427734375,
|
| 442 |
+
"learning_rate": 1.4401473296500922e-05,
|
| 443 |
+
"loss": 0.4659,
|
| 444 |
+
"step": 1275
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 7.18232044198895,
|
| 448 |
+
"grad_norm": 4.755856037139893,
|
| 449 |
+
"learning_rate": 1.4248004910988338e-05,
|
| 450 |
+
"loss": 0.2255,
|
| 451 |
+
"step": 1300
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 7.320441988950276,
|
| 455 |
+
"grad_norm": 83.01264190673828,
|
| 456 |
+
"learning_rate": 1.4094536525475753e-05,
|
| 457 |
+
"loss": 0.644,
|
| 458 |
+
"step": 1325
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 7.458563535911602,
|
| 462 |
+
"grad_norm": 21.05301284790039,
|
| 463 |
+
"learning_rate": 1.394106813996317e-05,
|
| 464 |
+
"loss": 0.4958,
|
| 465 |
+
"step": 1350
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 7.596685082872928,
|
| 469 |
+
"grad_norm": 69.73882293701172,
|
| 470 |
+
"learning_rate": 1.3787599754450585e-05,
|
| 471 |
+
"loss": 0.5525,
|
| 472 |
+
"step": 1375
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 7.734806629834254,
|
| 476 |
+
"grad_norm": 74.1106185913086,
|
| 477 |
+
"learning_rate": 1.3634131368937999e-05,
|
| 478 |
+
"loss": 0.6454,
|
| 479 |
+
"step": 1400
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 7.87292817679558,
|
| 483 |
+
"grad_norm": 58.45998764038086,
|
| 484 |
+
"learning_rate": 1.3480662983425417e-05,
|
| 485 |
+
"loss": 0.2582,
|
| 486 |
+
"step": 1425
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 8.0,
|
| 490 |
+
"eval_accuracy": 0.8109965635738832,
|
| 491 |
+
"eval_loss": 1.0342952013015747,
|
| 492 |
+
"eval_macro_f1": 0.8122542387307309,
|
| 493 |
+
"eval_runtime": 1.0686,
|
| 494 |
+
"eval_samples_per_second": 544.637,
|
| 495 |
+
"eval_steps_per_second": 34.625,
|
| 496 |
+
"eval_weighted_f1": 0.8129682064346981,
|
| 497 |
+
"step": 1448
|
| 498 |
+
},
|
| 499 |
+
{
|
| 500 |
+
"epoch": 8.011049723756907,
|
| 501 |
+
"grad_norm": 5.1458282470703125,
|
| 502 |
+
"learning_rate": 1.3327194597912831e-05,
|
| 503 |
+
"loss": 0.3905,
|
| 504 |
+
"step": 1450
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 8.149171270718233,
|
| 508 |
+
"grad_norm": 152.94375610351562,
|
| 509 |
+
"learning_rate": 1.3173726212400245e-05,
|
| 510 |
+
"loss": 0.3086,
|
| 511 |
+
"step": 1475
|
| 512 |
+
},
|
| 513 |
+
{
|
| 514 |
+
"epoch": 8.287292817679559,
|
| 515 |
+
"grad_norm": 1.4745954275131226,
|
| 516 |
+
"learning_rate": 1.3020257826887662e-05,
|
| 517 |
+
"loss": 0.2907,
|
| 518 |
+
"step": 1500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 8.425414364640885,
|
| 522 |
+
"grad_norm": 0.21353238821029663,
|
| 523 |
+
"learning_rate": 1.2866789441375078e-05,
|
| 524 |
+
"loss": 0.229,
|
| 525 |
+
"step": 1525
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 8.56353591160221,
|
| 529 |
+
"grad_norm": 0.42275407910346985,
|
| 530 |
+
"learning_rate": 1.2713321055862494e-05,
|
| 531 |
+
"loss": 0.2829,
|
| 532 |
+
"step": 1550
|
| 533 |
+
},
|
| 534 |
+
{
|
| 535 |
+
"epoch": 8.701657458563536,
|
| 536 |
+
"grad_norm": 150.2773895263672,
|
| 537 |
+
"learning_rate": 1.2559852670349908e-05,
|
| 538 |
+
"loss": 0.375,
|
| 539 |
+
"step": 1575
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 8.839779005524862,
|
| 543 |
+
"grad_norm": 0.09542777389287949,
|
| 544 |
+
"learning_rate": 1.2406384284837326e-05,
|
| 545 |
+
"loss": 0.282,
|
| 546 |
+
"step": 1600
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 8.977900552486188,
|
| 550 |
+
"grad_norm": 146.8008270263672,
|
| 551 |
+
"learning_rate": 1.225291589932474e-05,
|
| 552 |
+
"loss": 0.3413,
|
| 553 |
+
"step": 1625
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 9.0,
|
| 557 |
+
"eval_accuracy": 0.8161512027491409,
|
| 558 |
+
"eval_loss": 1.0832351446151733,
|
| 559 |
+
"eval_macro_f1": 0.8143389349954658,
|
| 560 |
+
"eval_runtime": 1.0404,
|
| 561 |
+
"eval_samples_per_second": 559.376,
|
| 562 |
+
"eval_steps_per_second": 35.562,
|
| 563 |
+
"eval_weighted_f1": 0.8151272246270538,
|
| 564 |
+
"step": 1629
|
| 565 |
+
}
|
| 566 |
+
],
|
| 567 |
+
"logging_steps": 25,
|
| 568 |
+
"max_steps": 3620,
|
| 569 |
+
"num_input_tokens_seen": 0,
|
| 570 |
+
"num_train_epochs": 20,
|
| 571 |
+
"save_steps": 500,
|
| 572 |
+
"stateful_callbacks": {
|
| 573 |
+
"EarlyStoppingCallback": {
|
| 574 |
+
"args": {
|
| 575 |
+
"early_stopping_patience": 3,
|
| 576 |
+
"early_stopping_threshold": 0.0
|
| 577 |
+
},
|
| 578 |
+
"attributes": {
|
| 579 |
+
"early_stopping_patience_counter": 3
|
| 580 |
+
}
|
| 581 |
+
},
|
| 582 |
+
"TrainerControl": {
|
| 583 |
+
"args": {
|
| 584 |
+
"should_epoch_stop": false,
|
| 585 |
+
"should_evaluate": false,
|
| 586 |
+
"should_log": false,
|
| 587 |
+
"should_save": true,
|
| 588 |
+
"should_training_stop": true
|
| 589 |
+
},
|
| 590 |
+
"attributes": {}
|
| 591 |
+
}
|
| 592 |
+
},
|
| 593 |
+
"total_flos": 0.0,
|
| 594 |
+
"train_batch_size": 8,
|
| 595 |
+
"trial_name": null,
|
| 596 |
+
"trial_params": null
|
| 597 |
+
}
|
ood-vifactcheck-bamibert-source-source-03-s44-20260829T021529Z-496d6871/trainer/checkpoint-1281/special_tokens_map.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "<s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "</s>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"mask_token": {
|
| 24 |
+
"content": "<mask>",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "<pad>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "</s>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "<unk>",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": false,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
+
}
|
ood-vifactcheck-bamibert-source-source-03-s44-20260829T021529Z-496d6871/trainer/checkpoint-1281/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-bamibert-source-source-03-s44-20260829T021529Z-496d6871/trainer/checkpoint-1281/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 |
+
"20480": {
|
| 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": 2048,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "XLMRobertaTokenizerFast",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-bamibert-source-source-03-s44-20260829T021529Z-496d6871/trainer/checkpoint-1281/trainer_state.json
ADDED
|
@@ -0,0 +1,477 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 732,
|
| 3 |
+
"best_metric": 0.8414306483731392,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-bamibert-source-source-03-s44-20260829T021529Z-496d6871/trainer/checkpoint-732",
|
| 5 |
+
"epoch": 7.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1281,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.13717421124828533,
|
| 14 |
+
"grad_norm": 11.189960479736328,
|
| 15 |
+
"learning_rate": 1.3114754098360657e-06,
|
| 16 |
+
"loss": 4.4541,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.27434842249657065,
|
| 21 |
+
"grad_norm": 9.076894760131836,
|
| 22 |
+
"learning_rate": 2.677595628415301e-06,
|
| 23 |
+
"loss": 4.4018,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.411522633744856,
|
| 28 |
+
"grad_norm": 11.297986030578613,
|
| 29 |
+
"learning_rate": 4.043715846994536e-06,
|
| 30 |
+
"loss": 4.3934,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5486968449931413,
|
| 35 |
+
"grad_norm": 49.99574661254883,
|
| 36 |
+
"learning_rate": 5.409836065573772e-06,
|
| 37 |
+
"loss": 4.3786,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6858710562414266,
|
| 42 |
+
"grad_norm": 9.995810508728027,
|
| 43 |
+
"learning_rate": 6.775956284153006e-06,
|
| 44 |
+
"loss": 4.411,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.823045267489712,
|
| 49 |
+
"grad_norm": 20.65247917175293,
|
| 50 |
+
"learning_rate": 8.14207650273224e-06,
|
| 51 |
+
"loss": 4.3927,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9602194787379973,
|
| 56 |
+
"grad_norm": 9.190462112426758,
|
| 57 |
+
"learning_rate": 9.508196721311476e-06,
|
| 58 |
+
"loss": 4.2631,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.43594009983361065,
|
| 64 |
+
"eval_loss": 1.0505797863006592,
|
| 65 |
+
"eval_macro_f1": 0.39786815293151473,
|
| 66 |
+
"eval_runtime": 1.1489,
|
| 67 |
+
"eval_samples_per_second": 523.108,
|
| 68 |
+
"eval_steps_per_second": 33.075,
|
| 69 |
+
"eval_weighted_f1": 0.40225063836462943,
|
| 70 |
+
"step": 183
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.093278463648834,
|
| 74 |
+
"grad_norm": 11.409358024597168,
|
| 75 |
+
"learning_rate": 1.0874316939890712e-05,
|
| 76 |
+
"loss": 4.1089,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.2304526748971194,
|
| 81 |
+
"grad_norm": 35.67272186279297,
|
| 82 |
+
"learning_rate": 1.2240437158469946e-05,
|
| 83 |
+
"loss": 4.0287,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.3676268861454046,
|
| 88 |
+
"grad_norm": 33.38080978393555,
|
| 89 |
+
"learning_rate": 1.3606557377049181e-05,
|
| 90 |
+
"loss": 3.7965,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.50480109739369,
|
| 95 |
+
"grad_norm": 40.98576736450195,
|
| 96 |
+
"learning_rate": 1.4972677595628417e-05,
|
| 97 |
+
"loss": 3.1235,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.6419753086419753,
|
| 102 |
+
"grad_norm": 33.52834701538086,
|
| 103 |
+
"learning_rate": 1.633879781420765e-05,
|
| 104 |
+
"loss": 2.9775,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7791495198902605,
|
| 109 |
+
"grad_norm": 60.922996520996094,
|
| 110 |
+
"learning_rate": 1.7704918032786887e-05,
|
| 111 |
+
"loss": 2.6854,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.916323731138546,
|
| 116 |
+
"grad_norm": 39.36262512207031,
|
| 117 |
+
"learning_rate": 1.9071038251366122e-05,
|
| 118 |
+
"loss": 2.7661,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.7587354409317804,
|
| 124 |
+
"eval_loss": 0.596989631652832,
|
| 125 |
+
"eval_macro_f1": 0.7600921442023137,
|
| 126 |
+
"eval_runtime": 1.2378,
|
| 127 |
+
"eval_samples_per_second": 485.548,
|
| 128 |
+
"eval_steps_per_second": 30.7,
|
| 129 |
+
"eval_weighted_f1": 0.7599400347013789,
|
| 130 |
+
"step": 366
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.049382716049383,
|
| 134 |
+
"grad_norm": 39.22695541381836,
|
| 135 |
+
"learning_rate": 1.995142683667274e-05,
|
| 136 |
+
"loss": 2.3734,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.186556927297668,
|
| 141 |
+
"grad_norm": 73.93245697021484,
|
| 142 |
+
"learning_rate": 1.9799635701275045e-05,
|
| 143 |
+
"loss": 2.0911,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.3237311385459534,
|
| 148 |
+
"grad_norm": 38.73860549926758,
|
| 149 |
+
"learning_rate": 1.9647844565877354e-05,
|
| 150 |
+
"loss": 2.0367,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.460905349794239,
|
| 155 |
+
"grad_norm": 50.98875045776367,
|
| 156 |
+
"learning_rate": 1.9496053430479663e-05,
|
| 157 |
+
"loss": 1.9306,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.598079561042524,
|
| 162 |
+
"grad_norm": 45.027034759521484,
|
| 163 |
+
"learning_rate": 1.934426229508197e-05,
|
| 164 |
+
"loss": 2.2272,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.7352537722908092,
|
| 169 |
+
"grad_norm": 61.15818786621094,
|
| 170 |
+
"learning_rate": 1.9192471159684274e-05,
|
| 171 |
+
"loss": 2.1014,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.8724279835390947,
|
| 176 |
+
"grad_norm": 50.471458435058594,
|
| 177 |
+
"learning_rate": 1.9040680024286583e-05,
|
| 178 |
+
"loss": 1.8755,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 3.0,
|
| 183 |
+
"eval_accuracy": 0.8119800332778702,
|
| 184 |
+
"eval_loss": 0.4782460927963257,
|
| 185 |
+
"eval_macro_f1": 0.8112290008841733,
|
| 186 |
+
"eval_runtime": 1.2262,
|
| 187 |
+
"eval_samples_per_second": 490.145,
|
| 188 |
+
"eval_steps_per_second": 30.991,
|
| 189 |
+
"eval_weighted_f1": 0.8118388009374298,
|
| 190 |
+
"step": 549
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.0054869684499312,
|
| 194 |
+
"grad_norm": 38.042625427246094,
|
| 195 |
+
"learning_rate": 1.888888888888889e-05,
|
| 196 |
+
"loss": 1.7792,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.1426611796982167,
|
| 201 |
+
"grad_norm": 71.85594177246094,
|
| 202 |
+
"learning_rate": 1.8737097753491198e-05,
|
| 203 |
+
"loss": 1.591,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.279835390946502,
|
| 208 |
+
"grad_norm": 49.27700424194336,
|
| 209 |
+
"learning_rate": 1.8585306618093504e-05,
|
| 210 |
+
"loss": 1.1804,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.4170096021947876,
|
| 215 |
+
"grad_norm": 50.078330993652344,
|
| 216 |
+
"learning_rate": 1.8433515482695813e-05,
|
| 217 |
+
"loss": 1.6091,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.5541838134430725,
|
| 222 |
+
"grad_norm": 43.49482345581055,
|
| 223 |
+
"learning_rate": 1.828172434729812e-05,
|
| 224 |
+
"loss": 1.2848,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.691358024691358,
|
| 229 |
+
"grad_norm": 27.69275665283203,
|
| 230 |
+
"learning_rate": 1.8129933211900424e-05,
|
| 231 |
+
"loss": 1.498,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.8285322359396434,
|
| 236 |
+
"grad_norm": 26.647626876831055,
|
| 237 |
+
"learning_rate": 1.7978142076502733e-05,
|
| 238 |
+
"loss": 1.4792,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.9657064471879284,
|
| 243 |
+
"grad_norm": 41.61360549926758,
|
| 244 |
+
"learning_rate": 1.7826350941105042e-05,
|
| 245 |
+
"loss": 1.4132,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 4.0,
|
| 250 |
+
"eval_accuracy": 0.8419301164725458,
|
| 251 |
+
"eval_loss": 0.46716317534446716,
|
| 252 |
+
"eval_macro_f1": 0.8414306483731392,
|
| 253 |
+
"eval_runtime": 1.0555,
|
| 254 |
+
"eval_samples_per_second": 569.371,
|
| 255 |
+
"eval_steps_per_second": 36.0,
|
| 256 |
+
"eval_weighted_f1": 0.8415059206166405,
|
| 257 |
+
"step": 732
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.098765432098766,
|
| 261 |
+
"grad_norm": 46.99260711669922,
|
| 262 |
+
"learning_rate": 1.7674559805707348e-05,
|
| 263 |
+
"loss": 1.1452,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.23593964334705,
|
| 268 |
+
"grad_norm": 57.263004302978516,
|
| 269 |
+
"learning_rate": 1.7522768670309657e-05,
|
| 270 |
+
"loss": 0.8888,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.373113854595336,
|
| 275 |
+
"grad_norm": 104.96063232421875,
|
| 276 |
+
"learning_rate": 1.7370977534911962e-05,
|
| 277 |
+
"loss": 1.1757,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.510288065843621,
|
| 282 |
+
"grad_norm": 45.959556579589844,
|
| 283 |
+
"learning_rate": 1.7219186399514268e-05,
|
| 284 |
+
"loss": 1.2513,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.647462277091907,
|
| 289 |
+
"grad_norm": 54.99401092529297,
|
| 290 |
+
"learning_rate": 1.7067395264116577e-05,
|
| 291 |
+
"loss": 0.7526,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.784636488340192,
|
| 296 |
+
"grad_norm": 32.72242736816406,
|
| 297 |
+
"learning_rate": 1.6915604128718883e-05,
|
| 298 |
+
"loss": 0.8225,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.921810699588478,
|
| 303 |
+
"grad_norm": 53.15626907348633,
|
| 304 |
+
"learning_rate": 1.6763812993321192e-05,
|
| 305 |
+
"loss": 0.8733,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 5.0,
|
| 310 |
+
"eval_accuracy": 0.831946755407654,
|
| 311 |
+
"eval_loss": 0.5456264615058899,
|
| 312 |
+
"eval_macro_f1": 0.8303880365633477,
|
| 313 |
+
"eval_runtime": 1.0665,
|
| 314 |
+
"eval_samples_per_second": 563.526,
|
| 315 |
+
"eval_steps_per_second": 35.631,
|
| 316 |
+
"eval_weighted_f1": 0.830614198617727,
|
| 317 |
+
"step": 915
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.054869684499314,
|
| 321 |
+
"grad_norm": 16.261497497558594,
|
| 322 |
+
"learning_rate": 1.66120218579235e-05,
|
| 323 |
+
"loss": 0.6468,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.1920438957476,
|
| 328 |
+
"grad_norm": 28.834444046020508,
|
| 329 |
+
"learning_rate": 1.6460230722525807e-05,
|
| 330 |
+
"loss": 0.5945,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.329218106995885,
|
| 335 |
+
"grad_norm": 51.509403228759766,
|
| 336 |
+
"learning_rate": 1.6308439587128112e-05,
|
| 337 |
+
"loss": 0.7388,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.4663923182441705,
|
| 342 |
+
"grad_norm": 91.80178833007812,
|
| 343 |
+
"learning_rate": 1.615664845173042e-05,
|
| 344 |
+
"loss": 0.6379,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.603566529492455,
|
| 349 |
+
"grad_norm": 79.57178497314453,
|
| 350 |
+
"learning_rate": 1.6004857316332727e-05,
|
| 351 |
+
"loss": 0.6748,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.7407407407407405,
|
| 356 |
+
"grad_norm": 68.71125030517578,
|
| 357 |
+
"learning_rate": 1.5853066180935036e-05,
|
| 358 |
+
"loss": 0.5015,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.877914951989026,
|
| 363 |
+
"grad_norm": 35.61964416503906,
|
| 364 |
+
"learning_rate": 1.570127504553734e-05,
|
| 365 |
+
"loss": 0.6242,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 6.0,
|
| 370 |
+
"eval_accuracy": 0.8169717138103162,
|
| 371 |
+
"eval_loss": 0.7627720832824707,
|
| 372 |
+
"eval_macro_f1": 0.8177629715719975,
|
| 373 |
+
"eval_runtime": 1.2397,
|
| 374 |
+
"eval_samples_per_second": 484.789,
|
| 375 |
+
"eval_steps_per_second": 30.652,
|
| 376 |
+
"eval_weighted_f1": 0.8176887581655143,
|
| 377 |
+
"step": 1098
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 6.0109739368998625,
|
| 381 |
+
"grad_norm": 68.40625,
|
| 382 |
+
"learning_rate": 1.5549483910139647e-05,
|
| 383 |
+
"loss": 0.7008,
|
| 384 |
+
"step": 1100
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 6.148148148148148,
|
| 388 |
+
"grad_norm": 49.65940856933594,
|
| 389 |
+
"learning_rate": 1.5397692774741956e-05,
|
| 390 |
+
"loss": 0.3347,
|
| 391 |
+
"step": 1125
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 6.285322359396433,
|
| 395 |
+
"grad_norm": 14.913434028625488,
|
| 396 |
+
"learning_rate": 1.5245901639344264e-05,
|
| 397 |
+
"loss": 0.4066,
|
| 398 |
+
"step": 1150
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.422496570644719,
|
| 402 |
+
"grad_norm": 18.121706008911133,
|
| 403 |
+
"learning_rate": 1.5094110503946571e-05,
|
| 404 |
+
"loss": 0.4203,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.559670781893004,
|
| 409 |
+
"grad_norm": 83.77801513671875,
|
| 410 |
+
"learning_rate": 1.4942319368548878e-05,
|
| 411 |
+
"loss": 0.4346,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.69684499314129,
|
| 416 |
+
"grad_norm": 100.65118408203125,
|
| 417 |
+
"learning_rate": 1.4790528233151184e-05,
|
| 418 |
+
"loss": 0.5967,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.834019204389575,
|
| 423 |
+
"grad_norm": 82.48768615722656,
|
| 424 |
+
"learning_rate": 1.4638737097753493e-05,
|
| 425 |
+
"loss": 0.6467,
|
| 426 |
+
"step": 1250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 6.97119341563786,
|
| 430 |
+
"grad_norm": 129.8446044921875,
|
| 431 |
+
"learning_rate": 1.44869459623558e-05,
|
| 432 |
+
"loss": 0.7176,
|
| 433 |
+
"step": 1275
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 7.0,
|
| 437 |
+
"eval_accuracy": 0.8386023294509152,
|
| 438 |
+
"eval_loss": 0.7942222356796265,
|
| 439 |
+
"eval_macro_f1": 0.8373103683701532,
|
| 440 |
+
"eval_runtime": 1.0948,
|
| 441 |
+
"eval_samples_per_second": 548.947,
|
| 442 |
+
"eval_steps_per_second": 34.709,
|
| 443 |
+
"eval_weighted_f1": 0.8375298826757408,
|
| 444 |
+
"step": 1281
|
| 445 |
+
}
|
| 446 |
+
],
|
| 447 |
+
"logging_steps": 25,
|
| 448 |
+
"max_steps": 3660,
|
| 449 |
+
"num_input_tokens_seen": 0,
|
| 450 |
+
"num_train_epochs": 20,
|
| 451 |
+
"save_steps": 500,
|
| 452 |
+
"stateful_callbacks": {
|
| 453 |
+
"EarlyStoppingCallback": {
|
| 454 |
+
"args": {
|
| 455 |
+
"early_stopping_patience": 3,
|
| 456 |
+
"early_stopping_threshold": 0.0
|
| 457 |
+
},
|
| 458 |
+
"attributes": {
|
| 459 |
+
"early_stopping_patience_counter": 3
|
| 460 |
+
}
|
| 461 |
+
},
|
| 462 |
+
"TrainerControl": {
|
| 463 |
+
"args": {
|
| 464 |
+
"should_epoch_stop": false,
|
| 465 |
+
"should_evaluate": false,
|
| 466 |
+
"should_log": false,
|
| 467 |
+
"should_save": true,
|
| 468 |
+
"should_training_stop": true
|
| 469 |
+
},
|
| 470 |
+
"attributes": {}
|
| 471 |
+
}
|
| 472 |
+
},
|
| 473 |
+
"total_flos": 0.0,
|
| 474 |
+
"train_batch_size": 8,
|
| 475 |
+
"trial_name": null,
|
| 476 |
+
"trial_params": null
|
| 477 |
+
}
|
ood-vifactcheck-bamibert-source-source-03-s44-20260829T021529Z-496d6871/trainer/checkpoint-732/special_tokens_map.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "<s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "</s>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"mask_token": {
|
| 24 |
+
"content": "<mask>",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "<pad>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "</s>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "<unk>",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": false,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
+
}
|
ood-vifactcheck-bamibert-source-source-03-s44-20260829T021529Z-496d6871/trainer/checkpoint-732/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-bamibert-source-source-03-s44-20260829T021529Z-496d6871/trainer/checkpoint-732/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 |
+
"20480": {
|
| 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": 2048,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "XLMRobertaTokenizerFast",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-bamibert-source-source-03-s44-20260829T021529Z-496d6871/trainer/checkpoint-732/trainer_state.json
ADDED
|
@@ -0,0 +1,290 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 732,
|
| 3 |
+
"best_metric": 0.8414306483731392,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-bamibert-source-source-03-s44-20260829T021529Z-496d6871/trainer/checkpoint-732",
|
| 5 |
+
"epoch": 4.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 732,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.13717421124828533,
|
| 14 |
+
"grad_norm": 11.189960479736328,
|
| 15 |
+
"learning_rate": 1.3114754098360657e-06,
|
| 16 |
+
"loss": 4.4541,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.27434842249657065,
|
| 21 |
+
"grad_norm": 9.076894760131836,
|
| 22 |
+
"learning_rate": 2.677595628415301e-06,
|
| 23 |
+
"loss": 4.4018,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.411522633744856,
|
| 28 |
+
"grad_norm": 11.297986030578613,
|
| 29 |
+
"learning_rate": 4.043715846994536e-06,
|
| 30 |
+
"loss": 4.3934,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5486968449931413,
|
| 35 |
+
"grad_norm": 49.99574661254883,
|
| 36 |
+
"learning_rate": 5.409836065573772e-06,
|
| 37 |
+
"loss": 4.3786,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6858710562414266,
|
| 42 |
+
"grad_norm": 9.995810508728027,
|
| 43 |
+
"learning_rate": 6.775956284153006e-06,
|
| 44 |
+
"loss": 4.411,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.823045267489712,
|
| 49 |
+
"grad_norm": 20.65247917175293,
|
| 50 |
+
"learning_rate": 8.14207650273224e-06,
|
| 51 |
+
"loss": 4.3927,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9602194787379973,
|
| 56 |
+
"grad_norm": 9.190462112426758,
|
| 57 |
+
"learning_rate": 9.508196721311476e-06,
|
| 58 |
+
"loss": 4.2631,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.43594009983361065,
|
| 64 |
+
"eval_loss": 1.0505797863006592,
|
| 65 |
+
"eval_macro_f1": 0.39786815293151473,
|
| 66 |
+
"eval_runtime": 1.1489,
|
| 67 |
+
"eval_samples_per_second": 523.108,
|
| 68 |
+
"eval_steps_per_second": 33.075,
|
| 69 |
+
"eval_weighted_f1": 0.40225063836462943,
|
| 70 |
+
"step": 183
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.093278463648834,
|
| 74 |
+
"grad_norm": 11.409358024597168,
|
| 75 |
+
"learning_rate": 1.0874316939890712e-05,
|
| 76 |
+
"loss": 4.1089,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.2304526748971194,
|
| 81 |
+
"grad_norm": 35.67272186279297,
|
| 82 |
+
"learning_rate": 1.2240437158469946e-05,
|
| 83 |
+
"loss": 4.0287,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.3676268861454046,
|
| 88 |
+
"grad_norm": 33.38080978393555,
|
| 89 |
+
"learning_rate": 1.3606557377049181e-05,
|
| 90 |
+
"loss": 3.7965,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.50480109739369,
|
| 95 |
+
"grad_norm": 40.98576736450195,
|
| 96 |
+
"learning_rate": 1.4972677595628417e-05,
|
| 97 |
+
"loss": 3.1235,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.6419753086419753,
|
| 102 |
+
"grad_norm": 33.52834701538086,
|
| 103 |
+
"learning_rate": 1.633879781420765e-05,
|
| 104 |
+
"loss": 2.9775,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7791495198902605,
|
| 109 |
+
"grad_norm": 60.922996520996094,
|
| 110 |
+
"learning_rate": 1.7704918032786887e-05,
|
| 111 |
+
"loss": 2.6854,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.916323731138546,
|
| 116 |
+
"grad_norm": 39.36262512207031,
|
| 117 |
+
"learning_rate": 1.9071038251366122e-05,
|
| 118 |
+
"loss": 2.7661,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.7587354409317804,
|
| 124 |
+
"eval_loss": 0.596989631652832,
|
| 125 |
+
"eval_macro_f1": 0.7600921442023137,
|
| 126 |
+
"eval_runtime": 1.2378,
|
| 127 |
+
"eval_samples_per_second": 485.548,
|
| 128 |
+
"eval_steps_per_second": 30.7,
|
| 129 |
+
"eval_weighted_f1": 0.7599400347013789,
|
| 130 |
+
"step": 366
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.049382716049383,
|
| 134 |
+
"grad_norm": 39.22695541381836,
|
| 135 |
+
"learning_rate": 1.995142683667274e-05,
|
| 136 |
+
"loss": 2.3734,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.186556927297668,
|
| 141 |
+
"grad_norm": 73.93245697021484,
|
| 142 |
+
"learning_rate": 1.9799635701275045e-05,
|
| 143 |
+
"loss": 2.0911,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.3237311385459534,
|
| 148 |
+
"grad_norm": 38.73860549926758,
|
| 149 |
+
"learning_rate": 1.9647844565877354e-05,
|
| 150 |
+
"loss": 2.0367,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.460905349794239,
|
| 155 |
+
"grad_norm": 50.98875045776367,
|
| 156 |
+
"learning_rate": 1.9496053430479663e-05,
|
| 157 |
+
"loss": 1.9306,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.598079561042524,
|
| 162 |
+
"grad_norm": 45.027034759521484,
|
| 163 |
+
"learning_rate": 1.934426229508197e-05,
|
| 164 |
+
"loss": 2.2272,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.7352537722908092,
|
| 169 |
+
"grad_norm": 61.15818786621094,
|
| 170 |
+
"learning_rate": 1.9192471159684274e-05,
|
| 171 |
+
"loss": 2.1014,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.8724279835390947,
|
| 176 |
+
"grad_norm": 50.471458435058594,
|
| 177 |
+
"learning_rate": 1.9040680024286583e-05,
|
| 178 |
+
"loss": 1.8755,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 3.0,
|
| 183 |
+
"eval_accuracy": 0.8119800332778702,
|
| 184 |
+
"eval_loss": 0.4782460927963257,
|
| 185 |
+
"eval_macro_f1": 0.8112290008841733,
|
| 186 |
+
"eval_runtime": 1.2262,
|
| 187 |
+
"eval_samples_per_second": 490.145,
|
| 188 |
+
"eval_steps_per_second": 30.991,
|
| 189 |
+
"eval_weighted_f1": 0.8118388009374298,
|
| 190 |
+
"step": 549
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.0054869684499312,
|
| 194 |
+
"grad_norm": 38.042625427246094,
|
| 195 |
+
"learning_rate": 1.888888888888889e-05,
|
| 196 |
+
"loss": 1.7792,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.1426611796982167,
|
| 201 |
+
"grad_norm": 71.85594177246094,
|
| 202 |
+
"learning_rate": 1.8737097753491198e-05,
|
| 203 |
+
"loss": 1.591,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.279835390946502,
|
| 208 |
+
"grad_norm": 49.27700424194336,
|
| 209 |
+
"learning_rate": 1.8585306618093504e-05,
|
| 210 |
+
"loss": 1.1804,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.4170096021947876,
|
| 215 |
+
"grad_norm": 50.078330993652344,
|
| 216 |
+
"learning_rate": 1.8433515482695813e-05,
|
| 217 |
+
"loss": 1.6091,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.5541838134430725,
|
| 222 |
+
"grad_norm": 43.49482345581055,
|
| 223 |
+
"learning_rate": 1.828172434729812e-05,
|
| 224 |
+
"loss": 1.2848,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.691358024691358,
|
| 229 |
+
"grad_norm": 27.69275665283203,
|
| 230 |
+
"learning_rate": 1.8129933211900424e-05,
|
| 231 |
+
"loss": 1.498,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.8285322359396434,
|
| 236 |
+
"grad_norm": 26.647626876831055,
|
| 237 |
+
"learning_rate": 1.7978142076502733e-05,
|
| 238 |
+
"loss": 1.4792,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.9657064471879284,
|
| 243 |
+
"grad_norm": 41.61360549926758,
|
| 244 |
+
"learning_rate": 1.7826350941105042e-05,
|
| 245 |
+
"loss": 1.4132,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 4.0,
|
| 250 |
+
"eval_accuracy": 0.8419301164725458,
|
| 251 |
+
"eval_loss": 0.46716317534446716,
|
| 252 |
+
"eval_macro_f1": 0.8414306483731392,
|
| 253 |
+
"eval_runtime": 1.0555,
|
| 254 |
+
"eval_samples_per_second": 569.371,
|
| 255 |
+
"eval_steps_per_second": 36.0,
|
| 256 |
+
"eval_weighted_f1": 0.8415059206166405,
|
| 257 |
+
"step": 732
|
| 258 |
+
}
|
| 259 |
+
],
|
| 260 |
+
"logging_steps": 25,
|
| 261 |
+
"max_steps": 3660,
|
| 262 |
+
"num_input_tokens_seen": 0,
|
| 263 |
+
"num_train_epochs": 20,
|
| 264 |
+
"save_steps": 500,
|
| 265 |
+
"stateful_callbacks": {
|
| 266 |
+
"EarlyStoppingCallback": {
|
| 267 |
+
"args": {
|
| 268 |
+
"early_stopping_patience": 3,
|
| 269 |
+
"early_stopping_threshold": 0.0
|
| 270 |
+
},
|
| 271 |
+
"attributes": {
|
| 272 |
+
"early_stopping_patience_counter": 0
|
| 273 |
+
}
|
| 274 |
+
},
|
| 275 |
+
"TrainerControl": {
|
| 276 |
+
"args": {
|
| 277 |
+
"should_epoch_stop": false,
|
| 278 |
+
"should_evaluate": false,
|
| 279 |
+
"should_log": false,
|
| 280 |
+
"should_save": true,
|
| 281 |
+
"should_training_stop": false
|
| 282 |
+
},
|
| 283 |
+
"attributes": {}
|
| 284 |
+
}
|
| 285 |
+
},
|
| 286 |
+
"total_flos": 0.0,
|
| 287 |
+
"train_batch_size": 8,
|
| 288 |
+
"trial_name": null,
|
| 289 |
+
"trial_params": null
|
| 290 |
+
}
|
ood-vifactcheck-bamibert-source-source-04-s44-20260829T022221Z-496d6871/trainer/checkpoint-1232/special_tokens_map.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "<s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "</s>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"mask_token": {
|
| 24 |
+
"content": "<mask>",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "<pad>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "</s>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "<unk>",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": false,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
+
}
|
ood-vifactcheck-bamibert-source-source-04-s44-20260829T022221Z-496d6871/trainer/checkpoint-1232/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-bamibert-source-source-04-s44-20260829T022221Z-496d6871/trainer/checkpoint-1232/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 |
+
"20480": {
|
| 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": 2048,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "XLMRobertaTokenizerFast",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-bamibert-source-source-04-s44-20260829T022221Z-496d6871/trainer/checkpoint-1232/trainer_state.json
ADDED
|
@@ -0,0 +1,463 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 704,
|
| 3 |
+
"best_metric": 0.8483499462179144,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-bamibert-source-source-04-s44-20260829T022221Z-496d6871/trainer/checkpoint-704",
|
| 5 |
+
"epoch": 7.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1232,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.14245014245014245,
|
| 14 |
+
"grad_norm": 10.823077201843262,
|
| 15 |
+
"learning_rate": 1.3636363636363636e-06,
|
| 16 |
+
"loss": 4.4188,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.2849002849002849,
|
| 21 |
+
"grad_norm": 13.990646362304688,
|
| 22 |
+
"learning_rate": 2.784090909090909e-06,
|
| 23 |
+
"loss": 4.4175,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.42735042735042733,
|
| 28 |
+
"grad_norm": 13.875608444213867,
|
| 29 |
+
"learning_rate": 4.204545454545455e-06,
|
| 30 |
+
"loss": 4.4015,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5698005698005698,
|
| 35 |
+
"grad_norm": 18.352867126464844,
|
| 36 |
+
"learning_rate": 5.625e-06,
|
| 37 |
+
"loss": 4.3708,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.7122507122507122,
|
| 42 |
+
"grad_norm": 32.78199768066406,
|
| 43 |
+
"learning_rate": 7.045454545454546e-06,
|
| 44 |
+
"loss": 4.3972,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8547008547008547,
|
| 49 |
+
"grad_norm": 22.737611770629883,
|
| 50 |
+
"learning_rate": 8.465909090909091e-06,
|
| 51 |
+
"loss": 4.3434,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9971509971509972,
|
| 56 |
+
"grad_norm": 17.734363555908203,
|
| 57 |
+
"learning_rate": 9.886363636363637e-06,
|
| 58 |
+
"loss": 4.3278,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.4397677793904209,
|
| 64 |
+
"eval_loss": 1.0458204746246338,
|
| 65 |
+
"eval_macro_f1": 0.39764758904320363,
|
| 66 |
+
"eval_runtime": 1.1926,
|
| 67 |
+
"eval_samples_per_second": 577.742,
|
| 68 |
+
"eval_steps_per_second": 36.895,
|
| 69 |
+
"eval_weighted_f1": 0.4020705339010283,
|
| 70 |
+
"step": 176
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.1367521367521367,
|
| 74 |
+
"grad_norm": 24.788105010986328,
|
| 75 |
+
"learning_rate": 1.1306818181818183e-05,
|
| 76 |
+
"loss": 4.1453,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.2792022792022792,
|
| 81 |
+
"grad_norm": 19.455854415893555,
|
| 82 |
+
"learning_rate": 1.2727272727272728e-05,
|
| 83 |
+
"loss": 4.0938,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.4216524216524218,
|
| 88 |
+
"grad_norm": 56.08881759643555,
|
| 89 |
+
"learning_rate": 1.4147727272727274e-05,
|
| 90 |
+
"loss": 3.6979,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.564102564102564,
|
| 95 |
+
"grad_norm": 40.092071533203125,
|
| 96 |
+
"learning_rate": 1.5568181818181822e-05,
|
| 97 |
+
"loss": 3.2124,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.7065527065527064,
|
| 102 |
+
"grad_norm": 29.906169891357422,
|
| 103 |
+
"learning_rate": 1.6988636363636368e-05,
|
| 104 |
+
"loss": 2.9068,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.849002849002849,
|
| 109 |
+
"grad_norm": 45.58308792114258,
|
| 110 |
+
"learning_rate": 1.840909090909091e-05,
|
| 111 |
+
"loss": 2.6334,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.9914529914529915,
|
| 116 |
+
"grad_norm": 37.55397415161133,
|
| 117 |
+
"learning_rate": 1.9829545454545456e-05,
|
| 118 |
+
"loss": 2.8086,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.7880986937590712,
|
| 124 |
+
"eval_loss": 0.5228174328804016,
|
| 125 |
+
"eval_macro_f1": 0.7827058179168409,
|
| 126 |
+
"eval_runtime": 1.2241,
|
| 127 |
+
"eval_samples_per_second": 562.885,
|
| 128 |
+
"eval_steps_per_second": 35.946,
|
| 129 |
+
"eval_weighted_f1": 0.7838739986441651,
|
| 130 |
+
"step": 352
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.131054131054131,
|
| 134 |
+
"grad_norm": 59.070594787597656,
|
| 135 |
+
"learning_rate": 1.9861111111111114e-05,
|
| 136 |
+
"loss": 2.2068,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.2735042735042734,
|
| 141 |
+
"grad_norm": 56.52789306640625,
|
| 142 |
+
"learning_rate": 1.9703282828282832e-05,
|
| 143 |
+
"loss": 2.1596,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.415954415954416,
|
| 148 |
+
"grad_norm": 51.37003707885742,
|
| 149 |
+
"learning_rate": 1.9545454545454546e-05,
|
| 150 |
+
"loss": 2.167,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.5584045584045585,
|
| 155 |
+
"grad_norm": 34.09475326538086,
|
| 156 |
+
"learning_rate": 1.9387626262626263e-05,
|
| 157 |
+
"loss": 2.281,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.700854700854701,
|
| 162 |
+
"grad_norm": 43.252925872802734,
|
| 163 |
+
"learning_rate": 1.922979797979798e-05,
|
| 164 |
+
"loss": 2.0887,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.8433048433048436,
|
| 169 |
+
"grad_norm": 55.824424743652344,
|
| 170 |
+
"learning_rate": 1.90719696969697e-05,
|
| 171 |
+
"loss": 1.8943,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.9857549857549857,
|
| 176 |
+
"grad_norm": 49.144142150878906,
|
| 177 |
+
"learning_rate": 1.8914141414141416e-05,
|
| 178 |
+
"loss": 2.0231,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 3.0,
|
| 183 |
+
"eval_accuracy": 0.8258345428156749,
|
| 184 |
+
"eval_loss": 0.445341020822525,
|
| 185 |
+
"eval_macro_f1": 0.8255258037770004,
|
| 186 |
+
"eval_runtime": 1.3874,
|
| 187 |
+
"eval_samples_per_second": 496.607,
|
| 188 |
+
"eval_steps_per_second": 31.714,
|
| 189 |
+
"eval_weighted_f1": 0.825927481026209,
|
| 190 |
+
"step": 528
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.1253561253561255,
|
| 194 |
+
"grad_norm": 32.462013244628906,
|
| 195 |
+
"learning_rate": 1.8756313131313133e-05,
|
| 196 |
+
"loss": 1.3643,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.267806267806268,
|
| 201 |
+
"grad_norm": 31.293603897094727,
|
| 202 |
+
"learning_rate": 1.859848484848485e-05,
|
| 203 |
+
"loss": 1.6046,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.41025641025641,
|
| 208 |
+
"grad_norm": 86.7500991821289,
|
| 209 |
+
"learning_rate": 1.844065656565657e-05,
|
| 210 |
+
"loss": 1.3924,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.5527065527065527,
|
| 215 |
+
"grad_norm": 59.56466293334961,
|
| 216 |
+
"learning_rate": 1.8282828282828286e-05,
|
| 217 |
+
"loss": 1.3943,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.695156695156695,
|
| 222 |
+
"grad_norm": 38.38946533203125,
|
| 223 |
+
"learning_rate": 1.8125e-05,
|
| 224 |
+
"loss": 1.4906,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.8376068376068377,
|
| 229 |
+
"grad_norm": 27.51655387878418,
|
| 230 |
+
"learning_rate": 1.7967171717171718e-05,
|
| 231 |
+
"loss": 1.5182,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.98005698005698,
|
| 236 |
+
"grad_norm": 64.71507263183594,
|
| 237 |
+
"learning_rate": 1.7809343434343435e-05,
|
| 238 |
+
"loss": 1.4228,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 4.0,
|
| 243 |
+
"eval_accuracy": 0.8490566037735849,
|
| 244 |
+
"eval_loss": 0.4359929859638214,
|
| 245 |
+
"eval_macro_f1": 0.8483499462179144,
|
| 246 |
+
"eval_runtime": 1.4555,
|
| 247 |
+
"eval_samples_per_second": 473.37,
|
| 248 |
+
"eval_steps_per_second": 30.23,
|
| 249 |
+
"eval_weighted_f1": 0.8488789991438674,
|
| 250 |
+
"step": 704
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 4.119658119658119,
|
| 254 |
+
"grad_norm": 29.26097869873047,
|
| 255 |
+
"learning_rate": 1.7651515151515153e-05,
|
| 256 |
+
"loss": 1.2091,
|
| 257 |
+
"step": 725
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.262108262108262,
|
| 261 |
+
"grad_norm": 29.644058227539062,
|
| 262 |
+
"learning_rate": 1.749368686868687e-05,
|
| 263 |
+
"loss": 0.8902,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.404558404558404,
|
| 268 |
+
"grad_norm": 48.6378173828125,
|
| 269 |
+
"learning_rate": 1.7335858585858588e-05,
|
| 270 |
+
"loss": 0.9743,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.547008547008547,
|
| 275 |
+
"grad_norm": 92.021484375,
|
| 276 |
+
"learning_rate": 1.7178030303030305e-05,
|
| 277 |
+
"loss": 1.0164,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.689458689458689,
|
| 282 |
+
"grad_norm": 9.568416595458984,
|
| 283 |
+
"learning_rate": 1.7020202020202023e-05,
|
| 284 |
+
"loss": 1.0546,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.831908831908832,
|
| 289 |
+
"grad_norm": 36.156185150146484,
|
| 290 |
+
"learning_rate": 1.6862373737373737e-05,
|
| 291 |
+
"loss": 1.1209,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.9743589743589745,
|
| 296 |
+
"grad_norm": 29.737831115722656,
|
| 297 |
+
"learning_rate": 1.6704545454545454e-05,
|
| 298 |
+
"loss": 0.908,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 5.0,
|
| 303 |
+
"eval_accuracy": 0.8243831640058055,
|
| 304 |
+
"eval_loss": 0.5420055389404297,
|
| 305 |
+
"eval_macro_f1": 0.8251747484574237,
|
| 306 |
+
"eval_runtime": 1.2315,
|
| 307 |
+
"eval_samples_per_second": 559.459,
|
| 308 |
+
"eval_steps_per_second": 35.727,
|
| 309 |
+
"eval_weighted_f1": 0.8251288029282543,
|
| 310 |
+
"step": 880
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 5.113960113960114,
|
| 314 |
+
"grad_norm": 56.39426040649414,
|
| 315 |
+
"learning_rate": 1.6546717171717172e-05,
|
| 316 |
+
"loss": 0.5789,
|
| 317 |
+
"step": 900
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.256410256410256,
|
| 321 |
+
"grad_norm": 30.42128562927246,
|
| 322 |
+
"learning_rate": 1.638888888888889e-05,
|
| 323 |
+
"loss": 0.6145,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.398860398860399,
|
| 328 |
+
"grad_norm": 22.22464370727539,
|
| 329 |
+
"learning_rate": 1.6231060606060607e-05,
|
| 330 |
+
"loss": 0.6476,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.5413105413105415,
|
| 335 |
+
"grad_norm": 55.62599563598633,
|
| 336 |
+
"learning_rate": 1.6073232323232324e-05,
|
| 337 |
+
"loss": 0.7281,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.683760683760684,
|
| 342 |
+
"grad_norm": 13.225717544555664,
|
| 343 |
+
"learning_rate": 1.5915404040404042e-05,
|
| 344 |
+
"loss": 0.9048,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.8262108262108265,
|
| 349 |
+
"grad_norm": 13.551254272460938,
|
| 350 |
+
"learning_rate": 1.575757575757576e-05,
|
| 351 |
+
"loss": 0.7543,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.968660968660968,
|
| 356 |
+
"grad_norm": 16.735212326049805,
|
| 357 |
+
"learning_rate": 1.5599747474747477e-05,
|
| 358 |
+
"loss": 0.6195,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 6.0,
|
| 363 |
+
"eval_accuracy": 0.8214804063860668,
|
| 364 |
+
"eval_loss": 0.6844180226325989,
|
| 365 |
+
"eval_macro_f1": 0.8217933282762834,
|
| 366 |
+
"eval_runtime": 1.2601,
|
| 367 |
+
"eval_samples_per_second": 546.803,
|
| 368 |
+
"eval_steps_per_second": 34.919,
|
| 369 |
+
"eval_weighted_f1": 0.8221663119500343,
|
| 370 |
+
"step": 1056
|
| 371 |
+
},
|
| 372 |
+
{
|
| 373 |
+
"epoch": 6.1082621082621085,
|
| 374 |
+
"grad_norm": 57.83783721923828,
|
| 375 |
+
"learning_rate": 1.544191919191919e-05,
|
| 376 |
+
"loss": 0.5132,
|
| 377 |
+
"step": 1075
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 6.250712250712251,
|
| 381 |
+
"grad_norm": 98.25761413574219,
|
| 382 |
+
"learning_rate": 1.528409090909091e-05,
|
| 383 |
+
"loss": 0.5871,
|
| 384 |
+
"step": 1100
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 6.3931623931623935,
|
| 388 |
+
"grad_norm": 93.90412902832031,
|
| 389 |
+
"learning_rate": 1.5126262626262626e-05,
|
| 390 |
+
"loss": 0.5177,
|
| 391 |
+
"step": 1125
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 6.535612535612536,
|
| 395 |
+
"grad_norm": 46.37567901611328,
|
| 396 |
+
"learning_rate": 1.4968434343434344e-05,
|
| 397 |
+
"loss": 0.5626,
|
| 398 |
+
"step": 1150
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.678062678062678,
|
| 402 |
+
"grad_norm": 76.79432678222656,
|
| 403 |
+
"learning_rate": 1.4810606060606061e-05,
|
| 404 |
+
"loss": 0.5575,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.82051282051282,
|
| 409 |
+
"grad_norm": 80.60455322265625,
|
| 410 |
+
"learning_rate": 1.4652777777777779e-05,
|
| 411 |
+
"loss": 0.5444,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.962962962962963,
|
| 416 |
+
"grad_norm": 93.20869445800781,
|
| 417 |
+
"learning_rate": 1.4494949494949494e-05,
|
| 418 |
+
"loss": 0.4129,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 7.0,
|
| 423 |
+
"eval_accuracy": 0.818577648766328,
|
| 424 |
+
"eval_loss": 0.8065328001976013,
|
| 425 |
+
"eval_macro_f1": 0.818560482393932,
|
| 426 |
+
"eval_runtime": 1.1617,
|
| 427 |
+
"eval_samples_per_second": 593.12,
|
| 428 |
+
"eval_steps_per_second": 37.877,
|
| 429 |
+
"eval_weighted_f1": 0.8189957883678276,
|
| 430 |
+
"step": 1232
|
| 431 |
+
}
|
| 432 |
+
],
|
| 433 |
+
"logging_steps": 25,
|
| 434 |
+
"max_steps": 3520,
|
| 435 |
+
"num_input_tokens_seen": 0,
|
| 436 |
+
"num_train_epochs": 20,
|
| 437 |
+
"save_steps": 500,
|
| 438 |
+
"stateful_callbacks": {
|
| 439 |
+
"EarlyStoppingCallback": {
|
| 440 |
+
"args": {
|
| 441 |
+
"early_stopping_patience": 3,
|
| 442 |
+
"early_stopping_threshold": 0.0
|
| 443 |
+
},
|
| 444 |
+
"attributes": {
|
| 445 |
+
"early_stopping_patience_counter": 3
|
| 446 |
+
}
|
| 447 |
+
},
|
| 448 |
+
"TrainerControl": {
|
| 449 |
+
"args": {
|
| 450 |
+
"should_epoch_stop": false,
|
| 451 |
+
"should_evaluate": false,
|
| 452 |
+
"should_log": false,
|
| 453 |
+
"should_save": true,
|
| 454 |
+
"should_training_stop": true
|
| 455 |
+
},
|
| 456 |
+
"attributes": {}
|
| 457 |
+
}
|
| 458 |
+
},
|
| 459 |
+
"total_flos": 0.0,
|
| 460 |
+
"train_batch_size": 8,
|
| 461 |
+
"trial_name": null,
|
| 462 |
+
"trial_params": null
|
| 463 |
+
}
|
ood-vifactcheck-bamibert-source-source-04-s44-20260829T022221Z-496d6871/trainer/checkpoint-704/special_tokens_map.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "<s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "</s>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"mask_token": {
|
| 24 |
+
"content": "<mask>",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "<pad>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "</s>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "<unk>",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": false,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
+
}
|
ood-vifactcheck-bamibert-source-source-04-s44-20260829T022221Z-496d6871/trainer/checkpoint-704/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-bamibert-source-source-04-s44-20260829T022221Z-496d6871/trainer/checkpoint-704/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 |
+
"20480": {
|
| 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": 2048,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "XLMRobertaTokenizerFast",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-bamibert-source-source-04-s44-20260829T022221Z-496d6871/trainer/checkpoint-704/trainer_state.json
ADDED
|
@@ -0,0 +1,283 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 704,
|
| 3 |
+
"best_metric": 0.8483499462179144,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-bamibert-source-source-04-s44-20260829T022221Z-496d6871/trainer/checkpoint-704",
|
| 5 |
+
"epoch": 4.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 704,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.14245014245014245,
|
| 14 |
+
"grad_norm": 10.823077201843262,
|
| 15 |
+
"learning_rate": 1.3636363636363636e-06,
|
| 16 |
+
"loss": 4.4188,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.2849002849002849,
|
| 21 |
+
"grad_norm": 13.990646362304688,
|
| 22 |
+
"learning_rate": 2.784090909090909e-06,
|
| 23 |
+
"loss": 4.4175,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.42735042735042733,
|
| 28 |
+
"grad_norm": 13.875608444213867,
|
| 29 |
+
"learning_rate": 4.204545454545455e-06,
|
| 30 |
+
"loss": 4.4015,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5698005698005698,
|
| 35 |
+
"grad_norm": 18.352867126464844,
|
| 36 |
+
"learning_rate": 5.625e-06,
|
| 37 |
+
"loss": 4.3708,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.7122507122507122,
|
| 42 |
+
"grad_norm": 32.78199768066406,
|
| 43 |
+
"learning_rate": 7.045454545454546e-06,
|
| 44 |
+
"loss": 4.3972,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8547008547008547,
|
| 49 |
+
"grad_norm": 22.737611770629883,
|
| 50 |
+
"learning_rate": 8.465909090909091e-06,
|
| 51 |
+
"loss": 4.3434,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9971509971509972,
|
| 56 |
+
"grad_norm": 17.734363555908203,
|
| 57 |
+
"learning_rate": 9.886363636363637e-06,
|
| 58 |
+
"loss": 4.3278,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.4397677793904209,
|
| 64 |
+
"eval_loss": 1.0458204746246338,
|
| 65 |
+
"eval_macro_f1": 0.39764758904320363,
|
| 66 |
+
"eval_runtime": 1.1926,
|
| 67 |
+
"eval_samples_per_second": 577.742,
|
| 68 |
+
"eval_steps_per_second": 36.895,
|
| 69 |
+
"eval_weighted_f1": 0.4020705339010283,
|
| 70 |
+
"step": 176
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.1367521367521367,
|
| 74 |
+
"grad_norm": 24.788105010986328,
|
| 75 |
+
"learning_rate": 1.1306818181818183e-05,
|
| 76 |
+
"loss": 4.1453,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.2792022792022792,
|
| 81 |
+
"grad_norm": 19.455854415893555,
|
| 82 |
+
"learning_rate": 1.2727272727272728e-05,
|
| 83 |
+
"loss": 4.0938,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.4216524216524218,
|
| 88 |
+
"grad_norm": 56.08881759643555,
|
| 89 |
+
"learning_rate": 1.4147727272727274e-05,
|
| 90 |
+
"loss": 3.6979,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.564102564102564,
|
| 95 |
+
"grad_norm": 40.092071533203125,
|
| 96 |
+
"learning_rate": 1.5568181818181822e-05,
|
| 97 |
+
"loss": 3.2124,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.7065527065527064,
|
| 102 |
+
"grad_norm": 29.906169891357422,
|
| 103 |
+
"learning_rate": 1.6988636363636368e-05,
|
| 104 |
+
"loss": 2.9068,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.849002849002849,
|
| 109 |
+
"grad_norm": 45.58308792114258,
|
| 110 |
+
"learning_rate": 1.840909090909091e-05,
|
| 111 |
+
"loss": 2.6334,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.9914529914529915,
|
| 116 |
+
"grad_norm": 37.55397415161133,
|
| 117 |
+
"learning_rate": 1.9829545454545456e-05,
|
| 118 |
+
"loss": 2.8086,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.7880986937590712,
|
| 124 |
+
"eval_loss": 0.5228174328804016,
|
| 125 |
+
"eval_macro_f1": 0.7827058179168409,
|
| 126 |
+
"eval_runtime": 1.2241,
|
| 127 |
+
"eval_samples_per_second": 562.885,
|
| 128 |
+
"eval_steps_per_second": 35.946,
|
| 129 |
+
"eval_weighted_f1": 0.7838739986441651,
|
| 130 |
+
"step": 352
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.131054131054131,
|
| 134 |
+
"grad_norm": 59.070594787597656,
|
| 135 |
+
"learning_rate": 1.9861111111111114e-05,
|
| 136 |
+
"loss": 2.2068,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.2735042735042734,
|
| 141 |
+
"grad_norm": 56.52789306640625,
|
| 142 |
+
"learning_rate": 1.9703282828282832e-05,
|
| 143 |
+
"loss": 2.1596,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.415954415954416,
|
| 148 |
+
"grad_norm": 51.37003707885742,
|
| 149 |
+
"learning_rate": 1.9545454545454546e-05,
|
| 150 |
+
"loss": 2.167,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.5584045584045585,
|
| 155 |
+
"grad_norm": 34.09475326538086,
|
| 156 |
+
"learning_rate": 1.9387626262626263e-05,
|
| 157 |
+
"loss": 2.281,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.700854700854701,
|
| 162 |
+
"grad_norm": 43.252925872802734,
|
| 163 |
+
"learning_rate": 1.922979797979798e-05,
|
| 164 |
+
"loss": 2.0887,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.8433048433048436,
|
| 169 |
+
"grad_norm": 55.824424743652344,
|
| 170 |
+
"learning_rate": 1.90719696969697e-05,
|
| 171 |
+
"loss": 1.8943,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.9857549857549857,
|
| 176 |
+
"grad_norm": 49.144142150878906,
|
| 177 |
+
"learning_rate": 1.8914141414141416e-05,
|
| 178 |
+
"loss": 2.0231,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 3.0,
|
| 183 |
+
"eval_accuracy": 0.8258345428156749,
|
| 184 |
+
"eval_loss": 0.445341020822525,
|
| 185 |
+
"eval_macro_f1": 0.8255258037770004,
|
| 186 |
+
"eval_runtime": 1.3874,
|
| 187 |
+
"eval_samples_per_second": 496.607,
|
| 188 |
+
"eval_steps_per_second": 31.714,
|
| 189 |
+
"eval_weighted_f1": 0.825927481026209,
|
| 190 |
+
"step": 528
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.1253561253561255,
|
| 194 |
+
"grad_norm": 32.462013244628906,
|
| 195 |
+
"learning_rate": 1.8756313131313133e-05,
|
| 196 |
+
"loss": 1.3643,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.267806267806268,
|
| 201 |
+
"grad_norm": 31.293603897094727,
|
| 202 |
+
"learning_rate": 1.859848484848485e-05,
|
| 203 |
+
"loss": 1.6046,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.41025641025641,
|
| 208 |
+
"grad_norm": 86.7500991821289,
|
| 209 |
+
"learning_rate": 1.844065656565657e-05,
|
| 210 |
+
"loss": 1.3924,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.5527065527065527,
|
| 215 |
+
"grad_norm": 59.56466293334961,
|
| 216 |
+
"learning_rate": 1.8282828282828286e-05,
|
| 217 |
+
"loss": 1.3943,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.695156695156695,
|
| 222 |
+
"grad_norm": 38.38946533203125,
|
| 223 |
+
"learning_rate": 1.8125e-05,
|
| 224 |
+
"loss": 1.4906,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.8376068376068377,
|
| 229 |
+
"grad_norm": 27.51655387878418,
|
| 230 |
+
"learning_rate": 1.7967171717171718e-05,
|
| 231 |
+
"loss": 1.5182,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.98005698005698,
|
| 236 |
+
"grad_norm": 64.71507263183594,
|
| 237 |
+
"learning_rate": 1.7809343434343435e-05,
|
| 238 |
+
"loss": 1.4228,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 4.0,
|
| 243 |
+
"eval_accuracy": 0.8490566037735849,
|
| 244 |
+
"eval_loss": 0.4359929859638214,
|
| 245 |
+
"eval_macro_f1": 0.8483499462179144,
|
| 246 |
+
"eval_runtime": 1.4555,
|
| 247 |
+
"eval_samples_per_second": 473.37,
|
| 248 |
+
"eval_steps_per_second": 30.23,
|
| 249 |
+
"eval_weighted_f1": 0.8488789991438674,
|
| 250 |
+
"step": 704
|
| 251 |
+
}
|
| 252 |
+
],
|
| 253 |
+
"logging_steps": 25,
|
| 254 |
+
"max_steps": 3520,
|
| 255 |
+
"num_input_tokens_seen": 0,
|
| 256 |
+
"num_train_epochs": 20,
|
| 257 |
+
"save_steps": 500,
|
| 258 |
+
"stateful_callbacks": {
|
| 259 |
+
"EarlyStoppingCallback": {
|
| 260 |
+
"args": {
|
| 261 |
+
"early_stopping_patience": 3,
|
| 262 |
+
"early_stopping_threshold": 0.0
|
| 263 |
+
},
|
| 264 |
+
"attributes": {
|
| 265 |
+
"early_stopping_patience_counter": 0
|
| 266 |
+
}
|
| 267 |
+
},
|
| 268 |
+
"TrainerControl": {
|
| 269 |
+
"args": {
|
| 270 |
+
"should_epoch_stop": false,
|
| 271 |
+
"should_evaluate": false,
|
| 272 |
+
"should_log": false,
|
| 273 |
+
"should_save": true,
|
| 274 |
+
"should_training_stop": false
|
| 275 |
+
},
|
| 276 |
+
"attributes": {}
|
| 277 |
+
}
|
| 278 |
+
},
|
| 279 |
+
"total_flos": 0.0,
|
| 280 |
+
"train_batch_size": 8,
|
| 281 |
+
"trial_name": null,
|
| 282 |
+
"trial_params": null
|
| 283 |
+
}
|
ood-vifactcheck-bamibert-source-source-05-s44-20260829T022906Z-496d6871/trainer/checkpoint-1620/special_tokens_map.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "<s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "</s>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"mask_token": {
|
| 24 |
+
"content": "<mask>",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "<pad>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "</s>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "<unk>",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": false,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
+
}
|
ood-vifactcheck-bamibert-source-source-05-s44-20260829T022906Z-496d6871/trainer/checkpoint-1620/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-bamibert-source-source-05-s44-20260829T022906Z-496d6871/trainer/checkpoint-1620/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 |
+
"20480": {
|
| 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": 2048,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "XLMRobertaTokenizerFast",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-bamibert-source-source-05-s44-20260829T022906Z-496d6871/trainer/checkpoint-1620/trainer_state.json
ADDED
|
@@ -0,0 +1,590 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1620,
|
| 3 |
+
"best_metric": 0.8148732842573422,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-bamibert-source-source-05-s44-20260829T022906Z-496d6871/trainer/checkpoint-1620",
|
| 5 |
+
"epoch": 9.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1620,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.1388888888888889,
|
| 14 |
+
"grad_norm": 11.068696975708008,
|
| 15 |
+
"learning_rate": 1.3333333333333334e-06,
|
| 16 |
+
"loss": 4.4184,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.2777777777777778,
|
| 21 |
+
"grad_norm": 9.884967803955078,
|
| 22 |
+
"learning_rate": 2.7222222222222224e-06,
|
| 23 |
+
"loss": 4.4316,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.4166666666666667,
|
| 28 |
+
"grad_norm": 9.222219467163086,
|
| 29 |
+
"learning_rate": 4.111111111111111e-06,
|
| 30 |
+
"loss": 4.3897,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5555555555555556,
|
| 35 |
+
"grad_norm": 9.293334007263184,
|
| 36 |
+
"learning_rate": 5.500000000000001e-06,
|
| 37 |
+
"loss": 4.3956,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6944444444444444,
|
| 42 |
+
"grad_norm": 7.660713195800781,
|
| 43 |
+
"learning_rate": 6.88888888888889e-06,
|
| 44 |
+
"loss": 4.4021,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8333333333333334,
|
| 49 |
+
"grad_norm": 16.73239517211914,
|
| 50 |
+
"learning_rate": 8.277777777777778e-06,
|
| 51 |
+
"loss": 4.3709,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9722222222222222,
|
| 56 |
+
"grad_norm": 16.58985137939453,
|
| 57 |
+
"learning_rate": 9.666666666666667e-06,
|
| 58 |
+
"loss": 4.3129,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.4253731343283582,
|
| 64 |
+
"eval_loss": 1.047044038772583,
|
| 65 |
+
"eval_macro_f1": 0.35760751361974935,
|
| 66 |
+
"eval_runtime": 1.2775,
|
| 67 |
+
"eval_samples_per_second": 524.461,
|
| 68 |
+
"eval_steps_per_second": 32.877,
|
| 69 |
+
"eval_weighted_f1": 0.3615937920167487,
|
| 70 |
+
"step": 180
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.1111111111111112,
|
| 74 |
+
"grad_norm": 24.939180374145508,
|
| 75 |
+
"learning_rate": 1.1055555555555557e-05,
|
| 76 |
+
"loss": 4.2038,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.25,
|
| 81 |
+
"grad_norm": 29.46761703491211,
|
| 82 |
+
"learning_rate": 1.2444444444444446e-05,
|
| 83 |
+
"loss": 4.0637,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.3888888888888888,
|
| 88 |
+
"grad_norm": 29.442646026611328,
|
| 89 |
+
"learning_rate": 1.3833333333333334e-05,
|
| 90 |
+
"loss": 3.658,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.5277777777777777,
|
| 95 |
+
"grad_norm": 34.75871658325195,
|
| 96 |
+
"learning_rate": 1.5222222222222223e-05,
|
| 97 |
+
"loss": 3.1041,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.6666666666666665,
|
| 102 |
+
"grad_norm": 81.986572265625,
|
| 103 |
+
"learning_rate": 1.6611111111111113e-05,
|
| 104 |
+
"loss": 2.9111,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.8055555555555556,
|
| 109 |
+
"grad_norm": 73.61305236816406,
|
| 110 |
+
"learning_rate": 1.8e-05,
|
| 111 |
+
"loss": 2.8138,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.9444444444444444,
|
| 116 |
+
"grad_norm": 58.9093017578125,
|
| 117 |
+
"learning_rate": 1.938888888888889e-05,
|
| 118 |
+
"loss": 2.4912,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.7328358208955223,
|
| 124 |
+
"eval_loss": 0.6236317157745361,
|
| 125 |
+
"eval_macro_f1": 0.7297010226134563,
|
| 126 |
+
"eval_runtime": 1.1683,
|
| 127 |
+
"eval_samples_per_second": 573.494,
|
| 128 |
+
"eval_steps_per_second": 35.95,
|
| 129 |
+
"eval_weighted_f1": 0.7297831143109896,
|
| 130 |
+
"step": 360
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.0833333333333335,
|
| 134 |
+
"grad_norm": 30.31173324584961,
|
| 135 |
+
"learning_rate": 1.991358024691358e-05,
|
| 136 |
+
"loss": 2.2326,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.2222222222222223,
|
| 141 |
+
"grad_norm": 58.152706146240234,
|
| 142 |
+
"learning_rate": 1.975925925925926e-05,
|
| 143 |
+
"loss": 2.196,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.361111111111111,
|
| 148 |
+
"grad_norm": 63.72364807128906,
|
| 149 |
+
"learning_rate": 1.960493827160494e-05,
|
| 150 |
+
"loss": 2.0034,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.5,
|
| 155 |
+
"grad_norm": 62.1304817199707,
|
| 156 |
+
"learning_rate": 1.945061728395062e-05,
|
| 157 |
+
"loss": 1.9878,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.638888888888889,
|
| 162 |
+
"grad_norm": 44.015201568603516,
|
| 163 |
+
"learning_rate": 1.92962962962963e-05,
|
| 164 |
+
"loss": 1.9173,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.7777777777777777,
|
| 169 |
+
"grad_norm": 32.35331726074219,
|
| 170 |
+
"learning_rate": 1.914197530864198e-05,
|
| 171 |
+
"loss": 2.0956,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.9166666666666665,
|
| 176 |
+
"grad_norm": 41.758148193359375,
|
| 177 |
+
"learning_rate": 1.8987654320987657e-05,
|
| 178 |
+
"loss": 2.094,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 3.0,
|
| 183 |
+
"eval_accuracy": 0.8014925373134328,
|
| 184 |
+
"eval_loss": 0.536288321018219,
|
| 185 |
+
"eval_macro_f1": 0.8026287106281544,
|
| 186 |
+
"eval_runtime": 1.1183,
|
| 187 |
+
"eval_samples_per_second": 599.112,
|
| 188 |
+
"eval_steps_per_second": 37.556,
|
| 189 |
+
"eval_weighted_f1": 0.8026043904369775,
|
| 190 |
+
"step": 540
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.0555555555555554,
|
| 194 |
+
"grad_norm": 23.91804313659668,
|
| 195 |
+
"learning_rate": 1.8833333333333335e-05,
|
| 196 |
+
"loss": 1.7222,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.1944444444444446,
|
| 201 |
+
"grad_norm": 36.21202087402344,
|
| 202 |
+
"learning_rate": 1.8679012345679013e-05,
|
| 203 |
+
"loss": 1.4497,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.3333333333333335,
|
| 208 |
+
"grad_norm": 33.39002227783203,
|
| 209 |
+
"learning_rate": 1.8524691358024694e-05,
|
| 210 |
+
"loss": 1.2607,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.4722222222222223,
|
| 215 |
+
"grad_norm": 24.654508590698242,
|
| 216 |
+
"learning_rate": 1.837037037037037e-05,
|
| 217 |
+
"loss": 1.3317,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.611111111111111,
|
| 222 |
+
"grad_norm": 46.853118896484375,
|
| 223 |
+
"learning_rate": 1.821604938271605e-05,
|
| 224 |
+
"loss": 1.4804,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.75,
|
| 229 |
+
"grad_norm": 40.93479919433594,
|
| 230 |
+
"learning_rate": 1.806172839506173e-05,
|
| 231 |
+
"loss": 1.4229,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.888888888888889,
|
| 236 |
+
"grad_norm": 56.005584716796875,
|
| 237 |
+
"learning_rate": 1.7907407407407408e-05,
|
| 238 |
+
"loss": 1.6409,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 4.0,
|
| 243 |
+
"eval_accuracy": 0.8059701492537313,
|
| 244 |
+
"eval_loss": 0.5493109226226807,
|
| 245 |
+
"eval_macro_f1": 0.8038556682550148,
|
| 246 |
+
"eval_runtime": 1.4713,
|
| 247 |
+
"eval_samples_per_second": 455.368,
|
| 248 |
+
"eval_steps_per_second": 28.545,
|
| 249 |
+
"eval_weighted_f1": 0.8038275574614349,
|
| 250 |
+
"step": 720
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 4.027777777777778,
|
| 254 |
+
"grad_norm": 35.491031646728516,
|
| 255 |
+
"learning_rate": 1.775308641975309e-05,
|
| 256 |
+
"loss": 1.368,
|
| 257 |
+
"step": 725
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.166666666666667,
|
| 261 |
+
"grad_norm": 48.94493865966797,
|
| 262 |
+
"learning_rate": 1.7598765432098767e-05,
|
| 263 |
+
"loss": 0.8571,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.305555555555555,
|
| 268 |
+
"grad_norm": 66.48535919189453,
|
| 269 |
+
"learning_rate": 1.7444444444444448e-05,
|
| 270 |
+
"loss": 0.869,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.444444444444445,
|
| 275 |
+
"grad_norm": 41.62992858886719,
|
| 276 |
+
"learning_rate": 1.7290123456790125e-05,
|
| 277 |
+
"loss": 1.0444,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.583333333333333,
|
| 282 |
+
"grad_norm": 64.55258178710938,
|
| 283 |
+
"learning_rate": 1.7135802469135806e-05,
|
| 284 |
+
"loss": 1.0393,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.722222222222222,
|
| 289 |
+
"grad_norm": 63.23442840576172,
|
| 290 |
+
"learning_rate": 1.6981481481481484e-05,
|
| 291 |
+
"loss": 1.0757,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.861111111111111,
|
| 296 |
+
"grad_norm": 60.5599365234375,
|
| 297 |
+
"learning_rate": 1.6827160493827162e-05,
|
| 298 |
+
"loss": 1.1253,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 5.0,
|
| 303 |
+
"grad_norm": 70.414794921875,
|
| 304 |
+
"learning_rate": 1.667283950617284e-05,
|
| 305 |
+
"loss": 0.8121,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 5.0,
|
| 310 |
+
"eval_accuracy": 0.7925373134328358,
|
| 311 |
+
"eval_loss": 0.6475654244422913,
|
| 312 |
+
"eval_macro_f1": 0.7925047568409894,
|
| 313 |
+
"eval_runtime": 1.2463,
|
| 314 |
+
"eval_samples_per_second": 537.579,
|
| 315 |
+
"eval_steps_per_second": 33.699,
|
| 316 |
+
"eval_weighted_f1": 0.7922318015364535,
|
| 317 |
+
"step": 900
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.138888888888889,
|
| 321 |
+
"grad_norm": 27.50987434387207,
|
| 322 |
+
"learning_rate": 1.651851851851852e-05,
|
| 323 |
+
"loss": 0.6812,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.277777777777778,
|
| 328 |
+
"grad_norm": 82.33159637451172,
|
| 329 |
+
"learning_rate": 1.6364197530864198e-05,
|
| 330 |
+
"loss": 0.823,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.416666666666667,
|
| 335 |
+
"grad_norm": 29.44865608215332,
|
| 336 |
+
"learning_rate": 1.6209876543209876e-05,
|
| 337 |
+
"loss": 0.5544,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.555555555555555,
|
| 342 |
+
"grad_norm": 10.798467636108398,
|
| 343 |
+
"learning_rate": 1.6055555555555557e-05,
|
| 344 |
+
"loss": 0.6244,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.694444444444445,
|
| 349 |
+
"grad_norm": 43.99272155761719,
|
| 350 |
+
"learning_rate": 1.5901234567901235e-05,
|
| 351 |
+
"loss": 0.7715,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.833333333333333,
|
| 356 |
+
"grad_norm": 62.74761962890625,
|
| 357 |
+
"learning_rate": 1.5746913580246916e-05,
|
| 358 |
+
"loss": 0.5751,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.972222222222222,
|
| 363 |
+
"grad_norm": 32.02378845214844,
|
| 364 |
+
"learning_rate": 1.5592592592592593e-05,
|
| 365 |
+
"loss": 0.7494,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 6.0,
|
| 370 |
+
"eval_accuracy": 0.7955223880597015,
|
| 371 |
+
"eval_loss": 0.75309157371521,
|
| 372 |
+
"eval_macro_f1": 0.7948037561843376,
|
| 373 |
+
"eval_runtime": 1.3022,
|
| 374 |
+
"eval_samples_per_second": 514.503,
|
| 375 |
+
"eval_steps_per_second": 32.252,
|
| 376 |
+
"eval_weighted_f1": 0.7948788066612027,
|
| 377 |
+
"step": 1080
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 6.111111111111111,
|
| 381 |
+
"grad_norm": 11.061698913574219,
|
| 382 |
+
"learning_rate": 1.5438271604938274e-05,
|
| 383 |
+
"loss": 0.5315,
|
| 384 |
+
"step": 1100
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 6.25,
|
| 388 |
+
"grad_norm": 26.717884063720703,
|
| 389 |
+
"learning_rate": 1.5283950617283952e-05,
|
| 390 |
+
"loss": 0.3229,
|
| 391 |
+
"step": 1125
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 6.388888888888889,
|
| 395 |
+
"grad_norm": 60.8264045715332,
|
| 396 |
+
"learning_rate": 1.5129629629629632e-05,
|
| 397 |
+
"loss": 0.5669,
|
| 398 |
+
"step": 1150
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.527777777777778,
|
| 402 |
+
"grad_norm": 66.8740463256836,
|
| 403 |
+
"learning_rate": 1.4975308641975311e-05,
|
| 404 |
+
"loss": 0.5008,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.666666666666667,
|
| 409 |
+
"grad_norm": 36.30829620361328,
|
| 410 |
+
"learning_rate": 1.4820987654320989e-05,
|
| 411 |
+
"loss": 0.6362,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.805555555555555,
|
| 416 |
+
"grad_norm": 155.0914306640625,
|
| 417 |
+
"learning_rate": 1.4666666666666666e-05,
|
| 418 |
+
"loss": 0.4549,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.944444444444445,
|
| 423 |
+
"grad_norm": 139.9639434814453,
|
| 424 |
+
"learning_rate": 1.4512345679012347e-05,
|
| 425 |
+
"loss": 0.6395,
|
| 426 |
+
"step": 1250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 7.0,
|
| 430 |
+
"eval_accuracy": 0.808955223880597,
|
| 431 |
+
"eval_loss": 0.8966561555862427,
|
| 432 |
+
"eval_macro_f1": 0.809128225575594,
|
| 433 |
+
"eval_runtime": 1.1567,
|
| 434 |
+
"eval_samples_per_second": 579.236,
|
| 435 |
+
"eval_steps_per_second": 36.31,
|
| 436 |
+
"eval_weighted_f1": 0.8091520864576639,
|
| 437 |
+
"step": 1260
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 7.083333333333333,
|
| 441 |
+
"grad_norm": 7.8447747230529785,
|
| 442 |
+
"learning_rate": 1.4358024691358025e-05,
|
| 443 |
+
"loss": 0.3873,
|
| 444 |
+
"step": 1275
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 7.222222222222222,
|
| 448 |
+
"grad_norm": 52.89217758178711,
|
| 449 |
+
"learning_rate": 1.4203703703703704e-05,
|
| 450 |
+
"loss": 0.3552,
|
| 451 |
+
"step": 1300
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 7.361111111111111,
|
| 455 |
+
"grad_norm": 185.01222229003906,
|
| 456 |
+
"learning_rate": 1.4049382716049384e-05,
|
| 457 |
+
"loss": 0.3863,
|
| 458 |
+
"step": 1325
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 7.5,
|
| 462 |
+
"grad_norm": 73.46562957763672,
|
| 463 |
+
"learning_rate": 1.3895061728395063e-05,
|
| 464 |
+
"loss": 0.3545,
|
| 465 |
+
"step": 1350
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 7.638888888888889,
|
| 469 |
+
"grad_norm": 64.76278686523438,
|
| 470 |
+
"learning_rate": 1.3740740740740741e-05,
|
| 471 |
+
"loss": 0.4902,
|
| 472 |
+
"step": 1375
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 7.777777777777778,
|
| 476 |
+
"grad_norm": 102.80863952636719,
|
| 477 |
+
"learning_rate": 1.3586419753086422e-05,
|
| 478 |
+
"loss": 0.2782,
|
| 479 |
+
"step": 1400
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 7.916666666666667,
|
| 483 |
+
"grad_norm": 30.765993118286133,
|
| 484 |
+
"learning_rate": 1.34320987654321e-05,
|
| 485 |
+
"loss": 0.5709,
|
| 486 |
+
"step": 1425
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 8.0,
|
| 490 |
+
"eval_accuracy": 0.8074626865671641,
|
| 491 |
+
"eval_loss": 1.0201267004013062,
|
| 492 |
+
"eval_macro_f1": 0.8082715436259896,
|
| 493 |
+
"eval_runtime": 1.1436,
|
| 494 |
+
"eval_samples_per_second": 585.878,
|
| 495 |
+
"eval_steps_per_second": 36.727,
|
| 496 |
+
"eval_weighted_f1": 0.8082993735096795,
|
| 497 |
+
"step": 1440
|
| 498 |
+
},
|
| 499 |
+
{
|
| 500 |
+
"epoch": 8.055555555555555,
|
| 501 |
+
"grad_norm": 0.40022704005241394,
|
| 502 |
+
"learning_rate": 1.3277777777777779e-05,
|
| 503 |
+
"loss": 0.318,
|
| 504 |
+
"step": 1450
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 8.194444444444445,
|
| 508 |
+
"grad_norm": 135.5516357421875,
|
| 509 |
+
"learning_rate": 1.3123456790123458e-05,
|
| 510 |
+
"loss": 0.3243,
|
| 511 |
+
"step": 1475
|
| 512 |
+
},
|
| 513 |
+
{
|
| 514 |
+
"epoch": 8.333333333333334,
|
| 515 |
+
"grad_norm": 98.12139129638672,
|
| 516 |
+
"learning_rate": 1.2969135802469138e-05,
|
| 517 |
+
"loss": 0.2718,
|
| 518 |
+
"step": 1500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 8.472222222222221,
|
| 522 |
+
"grad_norm": 118.84815216064453,
|
| 523 |
+
"learning_rate": 1.2814814814814815e-05,
|
| 524 |
+
"loss": 0.3389,
|
| 525 |
+
"step": 1525
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 8.61111111111111,
|
| 529 |
+
"grad_norm": 30.021512985229492,
|
| 530 |
+
"learning_rate": 1.2660493827160493e-05,
|
| 531 |
+
"loss": 0.225,
|
| 532 |
+
"step": 1550
|
| 533 |
+
},
|
| 534 |
+
{
|
| 535 |
+
"epoch": 8.75,
|
| 536 |
+
"grad_norm": 158.55189514160156,
|
| 537 |
+
"learning_rate": 1.2506172839506174e-05,
|
| 538 |
+
"loss": 0.3143,
|
| 539 |
+
"step": 1575
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 8.88888888888889,
|
| 543 |
+
"grad_norm": 46.29622268676758,
|
| 544 |
+
"learning_rate": 1.2351851851851852e-05,
|
| 545 |
+
"loss": 0.4041,
|
| 546 |
+
"step": 1600
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 9.0,
|
| 550 |
+
"eval_accuracy": 0.8149253731343283,
|
| 551 |
+
"eval_loss": 1.034593105316162,
|
| 552 |
+
"eval_macro_f1": 0.8148732842573422,
|
| 553 |
+
"eval_runtime": 1.1648,
|
| 554 |
+
"eval_samples_per_second": 575.212,
|
| 555 |
+
"eval_steps_per_second": 36.058,
|
| 556 |
+
"eval_weighted_f1": 0.8151889663140474,
|
| 557 |
+
"step": 1620
|
| 558 |
+
}
|
| 559 |
+
],
|
| 560 |
+
"logging_steps": 25,
|
| 561 |
+
"max_steps": 3600,
|
| 562 |
+
"num_input_tokens_seen": 0,
|
| 563 |
+
"num_train_epochs": 20,
|
| 564 |
+
"save_steps": 500,
|
| 565 |
+
"stateful_callbacks": {
|
| 566 |
+
"EarlyStoppingCallback": {
|
| 567 |
+
"args": {
|
| 568 |
+
"early_stopping_patience": 3,
|
| 569 |
+
"early_stopping_threshold": 0.0
|
| 570 |
+
},
|
| 571 |
+
"attributes": {
|
| 572 |
+
"early_stopping_patience_counter": 0
|
| 573 |
+
}
|
| 574 |
+
},
|
| 575 |
+
"TrainerControl": {
|
| 576 |
+
"args": {
|
| 577 |
+
"should_epoch_stop": false,
|
| 578 |
+
"should_evaluate": false,
|
| 579 |
+
"should_log": false,
|
| 580 |
+
"should_save": true,
|
| 581 |
+
"should_training_stop": false
|
| 582 |
+
},
|
| 583 |
+
"attributes": {}
|
| 584 |
+
}
|
| 585 |
+
},
|
| 586 |
+
"total_flos": 0.0,
|
| 587 |
+
"train_batch_size": 8,
|
| 588 |
+
"trial_name": null,
|
| 589 |
+
"trial_params": null
|
| 590 |
+
}
|
ood-vifactcheck-bamibert-source-source-05-s44-20260829T022906Z-496d6871/trainer/checkpoint-2160/special_tokens_map.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "<s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "</s>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"mask_token": {
|
| 24 |
+
"content": "<mask>",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "<pad>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "</s>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "<unk>",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": false,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
+
}
|
ood-vifactcheck-bamibert-source-source-05-s44-20260829T022906Z-496d6871/trainer/checkpoint-2160/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-bamibert-source-source-05-s44-20260829T022906Z-496d6871/trainer/checkpoint-2160/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 |
+
"20480": {
|
| 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": 2048,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "XLMRobertaTokenizerFast",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-bamibert-source-source-05-s44-20260829T022906Z-496d6871/trainer/checkpoint-2160/trainer_state.json
ADDED
|
@@ -0,0 +1,777 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1620,
|
| 3 |
+
"best_metric": 0.8148732842573422,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-bamibert-source-source-05-s44-20260829T022906Z-496d6871/trainer/checkpoint-1620",
|
| 5 |
+
"epoch": 12.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 2160,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.1388888888888889,
|
| 14 |
+
"grad_norm": 11.068696975708008,
|
| 15 |
+
"learning_rate": 1.3333333333333334e-06,
|
| 16 |
+
"loss": 4.4184,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.2777777777777778,
|
| 21 |
+
"grad_norm": 9.884967803955078,
|
| 22 |
+
"learning_rate": 2.7222222222222224e-06,
|
| 23 |
+
"loss": 4.4316,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.4166666666666667,
|
| 28 |
+
"grad_norm": 9.222219467163086,
|
| 29 |
+
"learning_rate": 4.111111111111111e-06,
|
| 30 |
+
"loss": 4.3897,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5555555555555556,
|
| 35 |
+
"grad_norm": 9.293334007263184,
|
| 36 |
+
"learning_rate": 5.500000000000001e-06,
|
| 37 |
+
"loss": 4.3956,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6944444444444444,
|
| 42 |
+
"grad_norm": 7.660713195800781,
|
| 43 |
+
"learning_rate": 6.88888888888889e-06,
|
| 44 |
+
"loss": 4.4021,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8333333333333334,
|
| 49 |
+
"grad_norm": 16.73239517211914,
|
| 50 |
+
"learning_rate": 8.277777777777778e-06,
|
| 51 |
+
"loss": 4.3709,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9722222222222222,
|
| 56 |
+
"grad_norm": 16.58985137939453,
|
| 57 |
+
"learning_rate": 9.666666666666667e-06,
|
| 58 |
+
"loss": 4.3129,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.4253731343283582,
|
| 64 |
+
"eval_loss": 1.047044038772583,
|
| 65 |
+
"eval_macro_f1": 0.35760751361974935,
|
| 66 |
+
"eval_runtime": 1.2775,
|
| 67 |
+
"eval_samples_per_second": 524.461,
|
| 68 |
+
"eval_steps_per_second": 32.877,
|
| 69 |
+
"eval_weighted_f1": 0.3615937920167487,
|
| 70 |
+
"step": 180
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.1111111111111112,
|
| 74 |
+
"grad_norm": 24.939180374145508,
|
| 75 |
+
"learning_rate": 1.1055555555555557e-05,
|
| 76 |
+
"loss": 4.2038,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.25,
|
| 81 |
+
"grad_norm": 29.46761703491211,
|
| 82 |
+
"learning_rate": 1.2444444444444446e-05,
|
| 83 |
+
"loss": 4.0637,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.3888888888888888,
|
| 88 |
+
"grad_norm": 29.442646026611328,
|
| 89 |
+
"learning_rate": 1.3833333333333334e-05,
|
| 90 |
+
"loss": 3.658,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.5277777777777777,
|
| 95 |
+
"grad_norm": 34.75871658325195,
|
| 96 |
+
"learning_rate": 1.5222222222222223e-05,
|
| 97 |
+
"loss": 3.1041,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.6666666666666665,
|
| 102 |
+
"grad_norm": 81.986572265625,
|
| 103 |
+
"learning_rate": 1.6611111111111113e-05,
|
| 104 |
+
"loss": 2.9111,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.8055555555555556,
|
| 109 |
+
"grad_norm": 73.61305236816406,
|
| 110 |
+
"learning_rate": 1.8e-05,
|
| 111 |
+
"loss": 2.8138,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.9444444444444444,
|
| 116 |
+
"grad_norm": 58.9093017578125,
|
| 117 |
+
"learning_rate": 1.938888888888889e-05,
|
| 118 |
+
"loss": 2.4912,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.7328358208955223,
|
| 124 |
+
"eval_loss": 0.6236317157745361,
|
| 125 |
+
"eval_macro_f1": 0.7297010226134563,
|
| 126 |
+
"eval_runtime": 1.1683,
|
| 127 |
+
"eval_samples_per_second": 573.494,
|
| 128 |
+
"eval_steps_per_second": 35.95,
|
| 129 |
+
"eval_weighted_f1": 0.7297831143109896,
|
| 130 |
+
"step": 360
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.0833333333333335,
|
| 134 |
+
"grad_norm": 30.31173324584961,
|
| 135 |
+
"learning_rate": 1.991358024691358e-05,
|
| 136 |
+
"loss": 2.2326,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.2222222222222223,
|
| 141 |
+
"grad_norm": 58.152706146240234,
|
| 142 |
+
"learning_rate": 1.975925925925926e-05,
|
| 143 |
+
"loss": 2.196,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.361111111111111,
|
| 148 |
+
"grad_norm": 63.72364807128906,
|
| 149 |
+
"learning_rate": 1.960493827160494e-05,
|
| 150 |
+
"loss": 2.0034,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.5,
|
| 155 |
+
"grad_norm": 62.1304817199707,
|
| 156 |
+
"learning_rate": 1.945061728395062e-05,
|
| 157 |
+
"loss": 1.9878,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.638888888888889,
|
| 162 |
+
"grad_norm": 44.015201568603516,
|
| 163 |
+
"learning_rate": 1.92962962962963e-05,
|
| 164 |
+
"loss": 1.9173,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.7777777777777777,
|
| 169 |
+
"grad_norm": 32.35331726074219,
|
| 170 |
+
"learning_rate": 1.914197530864198e-05,
|
| 171 |
+
"loss": 2.0956,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.9166666666666665,
|
| 176 |
+
"grad_norm": 41.758148193359375,
|
| 177 |
+
"learning_rate": 1.8987654320987657e-05,
|
| 178 |
+
"loss": 2.094,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 3.0,
|
| 183 |
+
"eval_accuracy": 0.8014925373134328,
|
| 184 |
+
"eval_loss": 0.536288321018219,
|
| 185 |
+
"eval_macro_f1": 0.8026287106281544,
|
| 186 |
+
"eval_runtime": 1.1183,
|
| 187 |
+
"eval_samples_per_second": 599.112,
|
| 188 |
+
"eval_steps_per_second": 37.556,
|
| 189 |
+
"eval_weighted_f1": 0.8026043904369775,
|
| 190 |
+
"step": 540
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.0555555555555554,
|
| 194 |
+
"grad_norm": 23.91804313659668,
|
| 195 |
+
"learning_rate": 1.8833333333333335e-05,
|
| 196 |
+
"loss": 1.7222,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.1944444444444446,
|
| 201 |
+
"grad_norm": 36.21202087402344,
|
| 202 |
+
"learning_rate": 1.8679012345679013e-05,
|
| 203 |
+
"loss": 1.4497,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.3333333333333335,
|
| 208 |
+
"grad_norm": 33.39002227783203,
|
| 209 |
+
"learning_rate": 1.8524691358024694e-05,
|
| 210 |
+
"loss": 1.2607,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.4722222222222223,
|
| 215 |
+
"grad_norm": 24.654508590698242,
|
| 216 |
+
"learning_rate": 1.837037037037037e-05,
|
| 217 |
+
"loss": 1.3317,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.611111111111111,
|
| 222 |
+
"grad_norm": 46.853118896484375,
|
| 223 |
+
"learning_rate": 1.821604938271605e-05,
|
| 224 |
+
"loss": 1.4804,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.75,
|
| 229 |
+
"grad_norm": 40.93479919433594,
|
| 230 |
+
"learning_rate": 1.806172839506173e-05,
|
| 231 |
+
"loss": 1.4229,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.888888888888889,
|
| 236 |
+
"grad_norm": 56.005584716796875,
|
| 237 |
+
"learning_rate": 1.7907407407407408e-05,
|
| 238 |
+
"loss": 1.6409,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 4.0,
|
| 243 |
+
"eval_accuracy": 0.8059701492537313,
|
| 244 |
+
"eval_loss": 0.5493109226226807,
|
| 245 |
+
"eval_macro_f1": 0.8038556682550148,
|
| 246 |
+
"eval_runtime": 1.4713,
|
| 247 |
+
"eval_samples_per_second": 455.368,
|
| 248 |
+
"eval_steps_per_second": 28.545,
|
| 249 |
+
"eval_weighted_f1": 0.8038275574614349,
|
| 250 |
+
"step": 720
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 4.027777777777778,
|
| 254 |
+
"grad_norm": 35.491031646728516,
|
| 255 |
+
"learning_rate": 1.775308641975309e-05,
|
| 256 |
+
"loss": 1.368,
|
| 257 |
+
"step": 725
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.166666666666667,
|
| 261 |
+
"grad_norm": 48.94493865966797,
|
| 262 |
+
"learning_rate": 1.7598765432098767e-05,
|
| 263 |
+
"loss": 0.8571,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.305555555555555,
|
| 268 |
+
"grad_norm": 66.48535919189453,
|
| 269 |
+
"learning_rate": 1.7444444444444448e-05,
|
| 270 |
+
"loss": 0.869,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.444444444444445,
|
| 275 |
+
"grad_norm": 41.62992858886719,
|
| 276 |
+
"learning_rate": 1.7290123456790125e-05,
|
| 277 |
+
"loss": 1.0444,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.583333333333333,
|
| 282 |
+
"grad_norm": 64.55258178710938,
|
| 283 |
+
"learning_rate": 1.7135802469135806e-05,
|
| 284 |
+
"loss": 1.0393,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.722222222222222,
|
| 289 |
+
"grad_norm": 63.23442840576172,
|
| 290 |
+
"learning_rate": 1.6981481481481484e-05,
|
| 291 |
+
"loss": 1.0757,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.861111111111111,
|
| 296 |
+
"grad_norm": 60.5599365234375,
|
| 297 |
+
"learning_rate": 1.6827160493827162e-05,
|
| 298 |
+
"loss": 1.1253,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 5.0,
|
| 303 |
+
"grad_norm": 70.414794921875,
|
| 304 |
+
"learning_rate": 1.667283950617284e-05,
|
| 305 |
+
"loss": 0.8121,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 5.0,
|
| 310 |
+
"eval_accuracy": 0.7925373134328358,
|
| 311 |
+
"eval_loss": 0.6475654244422913,
|
| 312 |
+
"eval_macro_f1": 0.7925047568409894,
|
| 313 |
+
"eval_runtime": 1.2463,
|
| 314 |
+
"eval_samples_per_second": 537.579,
|
| 315 |
+
"eval_steps_per_second": 33.699,
|
| 316 |
+
"eval_weighted_f1": 0.7922318015364535,
|
| 317 |
+
"step": 900
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.138888888888889,
|
| 321 |
+
"grad_norm": 27.50987434387207,
|
| 322 |
+
"learning_rate": 1.651851851851852e-05,
|
| 323 |
+
"loss": 0.6812,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.277777777777778,
|
| 328 |
+
"grad_norm": 82.33159637451172,
|
| 329 |
+
"learning_rate": 1.6364197530864198e-05,
|
| 330 |
+
"loss": 0.823,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.416666666666667,
|
| 335 |
+
"grad_norm": 29.44865608215332,
|
| 336 |
+
"learning_rate": 1.6209876543209876e-05,
|
| 337 |
+
"loss": 0.5544,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.555555555555555,
|
| 342 |
+
"grad_norm": 10.798467636108398,
|
| 343 |
+
"learning_rate": 1.6055555555555557e-05,
|
| 344 |
+
"loss": 0.6244,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.694444444444445,
|
| 349 |
+
"grad_norm": 43.99272155761719,
|
| 350 |
+
"learning_rate": 1.5901234567901235e-05,
|
| 351 |
+
"loss": 0.7715,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.833333333333333,
|
| 356 |
+
"grad_norm": 62.74761962890625,
|
| 357 |
+
"learning_rate": 1.5746913580246916e-05,
|
| 358 |
+
"loss": 0.5751,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.972222222222222,
|
| 363 |
+
"grad_norm": 32.02378845214844,
|
| 364 |
+
"learning_rate": 1.5592592592592593e-05,
|
| 365 |
+
"loss": 0.7494,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 6.0,
|
| 370 |
+
"eval_accuracy": 0.7955223880597015,
|
| 371 |
+
"eval_loss": 0.75309157371521,
|
| 372 |
+
"eval_macro_f1": 0.7948037561843376,
|
| 373 |
+
"eval_runtime": 1.3022,
|
| 374 |
+
"eval_samples_per_second": 514.503,
|
| 375 |
+
"eval_steps_per_second": 32.252,
|
| 376 |
+
"eval_weighted_f1": 0.7948788066612027,
|
| 377 |
+
"step": 1080
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 6.111111111111111,
|
| 381 |
+
"grad_norm": 11.061698913574219,
|
| 382 |
+
"learning_rate": 1.5438271604938274e-05,
|
| 383 |
+
"loss": 0.5315,
|
| 384 |
+
"step": 1100
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 6.25,
|
| 388 |
+
"grad_norm": 26.717884063720703,
|
| 389 |
+
"learning_rate": 1.5283950617283952e-05,
|
| 390 |
+
"loss": 0.3229,
|
| 391 |
+
"step": 1125
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 6.388888888888889,
|
| 395 |
+
"grad_norm": 60.8264045715332,
|
| 396 |
+
"learning_rate": 1.5129629629629632e-05,
|
| 397 |
+
"loss": 0.5669,
|
| 398 |
+
"step": 1150
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.527777777777778,
|
| 402 |
+
"grad_norm": 66.8740463256836,
|
| 403 |
+
"learning_rate": 1.4975308641975311e-05,
|
| 404 |
+
"loss": 0.5008,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.666666666666667,
|
| 409 |
+
"grad_norm": 36.30829620361328,
|
| 410 |
+
"learning_rate": 1.4820987654320989e-05,
|
| 411 |
+
"loss": 0.6362,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.805555555555555,
|
| 416 |
+
"grad_norm": 155.0914306640625,
|
| 417 |
+
"learning_rate": 1.4666666666666666e-05,
|
| 418 |
+
"loss": 0.4549,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.944444444444445,
|
| 423 |
+
"grad_norm": 139.9639434814453,
|
| 424 |
+
"learning_rate": 1.4512345679012347e-05,
|
| 425 |
+
"loss": 0.6395,
|
| 426 |
+
"step": 1250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 7.0,
|
| 430 |
+
"eval_accuracy": 0.808955223880597,
|
| 431 |
+
"eval_loss": 0.8966561555862427,
|
| 432 |
+
"eval_macro_f1": 0.809128225575594,
|
| 433 |
+
"eval_runtime": 1.1567,
|
| 434 |
+
"eval_samples_per_second": 579.236,
|
| 435 |
+
"eval_steps_per_second": 36.31,
|
| 436 |
+
"eval_weighted_f1": 0.8091520864576639,
|
| 437 |
+
"step": 1260
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 7.083333333333333,
|
| 441 |
+
"grad_norm": 7.8447747230529785,
|
| 442 |
+
"learning_rate": 1.4358024691358025e-05,
|
| 443 |
+
"loss": 0.3873,
|
| 444 |
+
"step": 1275
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 7.222222222222222,
|
| 448 |
+
"grad_norm": 52.89217758178711,
|
| 449 |
+
"learning_rate": 1.4203703703703704e-05,
|
| 450 |
+
"loss": 0.3552,
|
| 451 |
+
"step": 1300
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 7.361111111111111,
|
| 455 |
+
"grad_norm": 185.01222229003906,
|
| 456 |
+
"learning_rate": 1.4049382716049384e-05,
|
| 457 |
+
"loss": 0.3863,
|
| 458 |
+
"step": 1325
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 7.5,
|
| 462 |
+
"grad_norm": 73.46562957763672,
|
| 463 |
+
"learning_rate": 1.3895061728395063e-05,
|
| 464 |
+
"loss": 0.3545,
|
| 465 |
+
"step": 1350
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 7.638888888888889,
|
| 469 |
+
"grad_norm": 64.76278686523438,
|
| 470 |
+
"learning_rate": 1.3740740740740741e-05,
|
| 471 |
+
"loss": 0.4902,
|
| 472 |
+
"step": 1375
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 7.777777777777778,
|
| 476 |
+
"grad_norm": 102.80863952636719,
|
| 477 |
+
"learning_rate": 1.3586419753086422e-05,
|
| 478 |
+
"loss": 0.2782,
|
| 479 |
+
"step": 1400
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 7.916666666666667,
|
| 483 |
+
"grad_norm": 30.765993118286133,
|
| 484 |
+
"learning_rate": 1.34320987654321e-05,
|
| 485 |
+
"loss": 0.5709,
|
| 486 |
+
"step": 1425
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 8.0,
|
| 490 |
+
"eval_accuracy": 0.8074626865671641,
|
| 491 |
+
"eval_loss": 1.0201267004013062,
|
| 492 |
+
"eval_macro_f1": 0.8082715436259896,
|
| 493 |
+
"eval_runtime": 1.1436,
|
| 494 |
+
"eval_samples_per_second": 585.878,
|
| 495 |
+
"eval_steps_per_second": 36.727,
|
| 496 |
+
"eval_weighted_f1": 0.8082993735096795,
|
| 497 |
+
"step": 1440
|
| 498 |
+
},
|
| 499 |
+
{
|
| 500 |
+
"epoch": 8.055555555555555,
|
| 501 |
+
"grad_norm": 0.40022704005241394,
|
| 502 |
+
"learning_rate": 1.3277777777777779e-05,
|
| 503 |
+
"loss": 0.318,
|
| 504 |
+
"step": 1450
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 8.194444444444445,
|
| 508 |
+
"grad_norm": 135.5516357421875,
|
| 509 |
+
"learning_rate": 1.3123456790123458e-05,
|
| 510 |
+
"loss": 0.3243,
|
| 511 |
+
"step": 1475
|
| 512 |
+
},
|
| 513 |
+
{
|
| 514 |
+
"epoch": 8.333333333333334,
|
| 515 |
+
"grad_norm": 98.12139129638672,
|
| 516 |
+
"learning_rate": 1.2969135802469138e-05,
|
| 517 |
+
"loss": 0.2718,
|
| 518 |
+
"step": 1500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 8.472222222222221,
|
| 522 |
+
"grad_norm": 118.84815216064453,
|
| 523 |
+
"learning_rate": 1.2814814814814815e-05,
|
| 524 |
+
"loss": 0.3389,
|
| 525 |
+
"step": 1525
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 8.61111111111111,
|
| 529 |
+
"grad_norm": 30.021512985229492,
|
| 530 |
+
"learning_rate": 1.2660493827160493e-05,
|
| 531 |
+
"loss": 0.225,
|
| 532 |
+
"step": 1550
|
| 533 |
+
},
|
| 534 |
+
{
|
| 535 |
+
"epoch": 8.75,
|
| 536 |
+
"grad_norm": 158.55189514160156,
|
| 537 |
+
"learning_rate": 1.2506172839506174e-05,
|
| 538 |
+
"loss": 0.3143,
|
| 539 |
+
"step": 1575
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 8.88888888888889,
|
| 543 |
+
"grad_norm": 46.29622268676758,
|
| 544 |
+
"learning_rate": 1.2351851851851852e-05,
|
| 545 |
+
"loss": 0.4041,
|
| 546 |
+
"step": 1600
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 9.0,
|
| 550 |
+
"eval_accuracy": 0.8149253731343283,
|
| 551 |
+
"eval_loss": 1.034593105316162,
|
| 552 |
+
"eval_macro_f1": 0.8148732842573422,
|
| 553 |
+
"eval_runtime": 1.1648,
|
| 554 |
+
"eval_samples_per_second": 575.212,
|
| 555 |
+
"eval_steps_per_second": 36.058,
|
| 556 |
+
"eval_weighted_f1": 0.8151889663140474,
|
| 557 |
+
"step": 1620
|
| 558 |
+
},
|
| 559 |
+
{
|
| 560 |
+
"epoch": 9.027777777777779,
|
| 561 |
+
"grad_norm": 1.1533715724945068,
|
| 562 |
+
"learning_rate": 1.2197530864197531e-05,
|
| 563 |
+
"loss": 0.3719,
|
| 564 |
+
"step": 1625
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"epoch": 9.166666666666666,
|
| 568 |
+
"grad_norm": 104.78726196289062,
|
| 569 |
+
"learning_rate": 1.204320987654321e-05,
|
| 570 |
+
"loss": 0.3694,
|
| 571 |
+
"step": 1650
|
| 572 |
+
},
|
| 573 |
+
{
|
| 574 |
+
"epoch": 9.305555555555555,
|
| 575 |
+
"grad_norm": 41.63246536254883,
|
| 576 |
+
"learning_rate": 1.188888888888889e-05,
|
| 577 |
+
"loss": 0.4414,
|
| 578 |
+
"step": 1675
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 9.444444444444445,
|
| 582 |
+
"grad_norm": 0.31377339363098145,
|
| 583 |
+
"learning_rate": 1.1734567901234568e-05,
|
| 584 |
+
"loss": 0.301,
|
| 585 |
+
"step": 1700
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 9.583333333333334,
|
| 589 |
+
"grad_norm": 3.4063668251037598,
|
| 590 |
+
"learning_rate": 1.1580246913580249e-05,
|
| 591 |
+
"loss": 0.2679,
|
| 592 |
+
"step": 1725
|
| 593 |
+
},
|
| 594 |
+
{
|
| 595 |
+
"epoch": 9.722222222222221,
|
| 596 |
+
"grad_norm": 244.53662109375,
|
| 597 |
+
"learning_rate": 1.1425925925925927e-05,
|
| 598 |
+
"loss": 0.2287,
|
| 599 |
+
"step": 1750
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 9.86111111111111,
|
| 603 |
+
"grad_norm": 31.85007667541504,
|
| 604 |
+
"learning_rate": 1.1271604938271606e-05,
|
| 605 |
+
"loss": 0.3839,
|
| 606 |
+
"step": 1775
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"epoch": 10.0,
|
| 610 |
+
"grad_norm": 106.38321685791016,
|
| 611 |
+
"learning_rate": 1.1117283950617285e-05,
|
| 612 |
+
"loss": 0.2979,
|
| 613 |
+
"step": 1800
|
| 614 |
+
},
|
| 615 |
+
{
|
| 616 |
+
"epoch": 10.0,
|
| 617 |
+
"eval_accuracy": 0.8059701492537313,
|
| 618 |
+
"eval_loss": 1.2069175243377686,
|
| 619 |
+
"eval_macro_f1": 0.8064578380697731,
|
| 620 |
+
"eval_runtime": 1.2547,
|
| 621 |
+
"eval_samples_per_second": 533.988,
|
| 622 |
+
"eval_steps_per_second": 33.474,
|
| 623 |
+
"eval_weighted_f1": 0.8065746093568985,
|
| 624 |
+
"step": 1800
|
| 625 |
+
},
|
| 626 |
+
{
|
| 627 |
+
"epoch": 10.13888888888889,
|
| 628 |
+
"grad_norm": 12.738669395446777,
|
| 629 |
+
"learning_rate": 1.0962962962962965e-05,
|
| 630 |
+
"loss": 0.1478,
|
| 631 |
+
"step": 1825
|
| 632 |
+
},
|
| 633 |
+
{
|
| 634 |
+
"epoch": 10.277777777777779,
|
| 635 |
+
"grad_norm": 0.2007477581501007,
|
| 636 |
+
"learning_rate": 1.0808641975308642e-05,
|
| 637 |
+
"loss": 0.0605,
|
| 638 |
+
"step": 1850
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 10.416666666666666,
|
| 642 |
+
"grad_norm": 0.051435429602861404,
|
| 643 |
+
"learning_rate": 1.065432098765432e-05,
|
| 644 |
+
"loss": 0.1573,
|
| 645 |
+
"step": 1875
|
| 646 |
+
},
|
| 647 |
+
{
|
| 648 |
+
"epoch": 10.555555555555555,
|
| 649 |
+
"grad_norm": 0.11048978567123413,
|
| 650 |
+
"learning_rate": 1.0500000000000001e-05,
|
| 651 |
+
"loss": 0.159,
|
| 652 |
+
"step": 1900
|
| 653 |
+
},
|
| 654 |
+
{
|
| 655 |
+
"epoch": 10.694444444444445,
|
| 656 |
+
"grad_norm": 154.48573303222656,
|
| 657 |
+
"learning_rate": 1.0345679012345679e-05,
|
| 658 |
+
"loss": 0.1993,
|
| 659 |
+
"step": 1925
|
| 660 |
+
},
|
| 661 |
+
{
|
| 662 |
+
"epoch": 10.833333333333334,
|
| 663 |
+
"grad_norm": 0.1568298041820526,
|
| 664 |
+
"learning_rate": 1.0191358024691358e-05,
|
| 665 |
+
"loss": 0.1837,
|
| 666 |
+
"step": 1950
|
| 667 |
+
},
|
| 668 |
+
{
|
| 669 |
+
"epoch": 10.972222222222221,
|
| 670 |
+
"grad_norm": 0.07513098418712616,
|
| 671 |
+
"learning_rate": 1.003703703703704e-05,
|
| 672 |
+
"loss": 0.141,
|
| 673 |
+
"step": 1975
|
| 674 |
+
},
|
| 675 |
+
{
|
| 676 |
+
"epoch": 11.0,
|
| 677 |
+
"eval_accuracy": 0.8,
|
| 678 |
+
"eval_loss": 1.2891130447387695,
|
| 679 |
+
"eval_macro_f1": 0.8015924326487273,
|
| 680 |
+
"eval_runtime": 1.1496,
|
| 681 |
+
"eval_samples_per_second": 582.825,
|
| 682 |
+
"eval_steps_per_second": 36.535,
|
| 683 |
+
"eval_weighted_f1": 0.8016432834378326,
|
| 684 |
+
"step": 1980
|
| 685 |
+
},
|
| 686 |
+
{
|
| 687 |
+
"epoch": 11.11111111111111,
|
| 688 |
+
"grad_norm": 0.07965394854545593,
|
| 689 |
+
"learning_rate": 9.882716049382717e-06,
|
| 690 |
+
"loss": 0.148,
|
| 691 |
+
"step": 2000
|
| 692 |
+
},
|
| 693 |
+
{
|
| 694 |
+
"epoch": 11.25,
|
| 695 |
+
"grad_norm": 0.07134576886892319,
|
| 696 |
+
"learning_rate": 9.728395061728396e-06,
|
| 697 |
+
"loss": 0.0512,
|
| 698 |
+
"step": 2025
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 11.38888888888889,
|
| 702 |
+
"grad_norm": 3.5403780937194824,
|
| 703 |
+
"learning_rate": 9.574074074074074e-06,
|
| 704 |
+
"loss": 0.1094,
|
| 705 |
+
"step": 2050
|
| 706 |
+
},
|
| 707 |
+
{
|
| 708 |
+
"epoch": 11.527777777777779,
|
| 709 |
+
"grad_norm": 0.16378413140773773,
|
| 710 |
+
"learning_rate": 9.419753086419753e-06,
|
| 711 |
+
"loss": 0.1363,
|
| 712 |
+
"step": 2075
|
| 713 |
+
},
|
| 714 |
+
{
|
| 715 |
+
"epoch": 11.666666666666666,
|
| 716 |
+
"grad_norm": 17.97455406188965,
|
| 717 |
+
"learning_rate": 9.265432098765433e-06,
|
| 718 |
+
"loss": 0.1649,
|
| 719 |
+
"step": 2100
|
| 720 |
+
},
|
| 721 |
+
{
|
| 722 |
+
"epoch": 11.805555555555555,
|
| 723 |
+
"grad_norm": 0.041285913437604904,
|
| 724 |
+
"learning_rate": 9.111111111111112e-06,
|
| 725 |
+
"loss": 0.0738,
|
| 726 |
+
"step": 2125
|
| 727 |
+
},
|
| 728 |
+
{
|
| 729 |
+
"epoch": 11.944444444444445,
|
| 730 |
+
"grad_norm": 0.5761922001838684,
|
| 731 |
+
"learning_rate": 8.956790123456792e-06,
|
| 732 |
+
"loss": 0.1358,
|
| 733 |
+
"step": 2150
|
| 734 |
+
},
|
| 735 |
+
{
|
| 736 |
+
"epoch": 12.0,
|
| 737 |
+
"eval_accuracy": 0.8134328358208955,
|
| 738 |
+
"eval_loss": 1.2781795263290405,
|
| 739 |
+
"eval_macro_f1": 0.813865265373377,
|
| 740 |
+
"eval_runtime": 1.1357,
|
| 741 |
+
"eval_samples_per_second": 589.947,
|
| 742 |
+
"eval_steps_per_second": 36.982,
|
| 743 |
+
"eval_weighted_f1": 0.814003934180603,
|
| 744 |
+
"step": 2160
|
| 745 |
+
}
|
| 746 |
+
],
|
| 747 |
+
"logging_steps": 25,
|
| 748 |
+
"max_steps": 3600,
|
| 749 |
+
"num_input_tokens_seen": 0,
|
| 750 |
+
"num_train_epochs": 20,
|
| 751 |
+
"save_steps": 500,
|
| 752 |
+
"stateful_callbacks": {
|
| 753 |
+
"EarlyStoppingCallback": {
|
| 754 |
+
"args": {
|
| 755 |
+
"early_stopping_patience": 3,
|
| 756 |
+
"early_stopping_threshold": 0.0
|
| 757 |
+
},
|
| 758 |
+
"attributes": {
|
| 759 |
+
"early_stopping_patience_counter": 3
|
| 760 |
+
}
|
| 761 |
+
},
|
| 762 |
+
"TrainerControl": {
|
| 763 |
+
"args": {
|
| 764 |
+
"should_epoch_stop": false,
|
| 765 |
+
"should_evaluate": false,
|
| 766 |
+
"should_log": false,
|
| 767 |
+
"should_save": true,
|
| 768 |
+
"should_training_stop": true
|
| 769 |
+
},
|
| 770 |
+
"attributes": {}
|
| 771 |
+
}
|
| 772 |
+
},
|
| 773 |
+
"total_flos": 0.0,
|
| 774 |
+
"train_batch_size": 8,
|
| 775 |
+
"trial_name": null,
|
| 776 |
+
"trial_params": null
|
| 777 |
+
}
|
ood-vifactcheck-bamibert-source-source-06-s44-20260829T024021Z-496d6871/trainer/checkpoint-1480/special_tokens_map.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "<s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "</s>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"mask_token": {
|
| 24 |
+
"content": "<mask>",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "<pad>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "</s>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "<unk>",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": false,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
+
}
|
ood-vifactcheck-bamibert-source-source-06-s44-20260829T024021Z-496d6871/trainer/checkpoint-1480/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|