Add files using upload-large-folder tool
Browse filesThis view is limited to 50 files because it contains too many changes. See raw diff
- uit_vsmec-bamibert-s42-20260827T141858Z-45c80ccd/trainer/checkpoint-1044/special_tokens_map.json +51 -0
- uit_vsmec-bamibert-s42-20260827T141858Z-45c80ccd/trainer/checkpoint-1044/tokenizer.json +0 -0
- uit_vsmec-bamibert-s42-20260827T141858Z-45c80ccd/trainer/checkpoint-1044/tokenizer_config.json +55 -0
- uit_vsmec-bamibert-s42-20260827T141858Z-45c80ccd/trainer/checkpoint-1044/trainer_state.json +396 -0
- uit_vsmec-bamibert-s42-20260827T141858Z-45c80ccd/trainer/checkpoint-1566/tokenizer_config.json +55 -0
- viclickbait2025-bamibert-s42-20260827T130938Z-ee19aa8e/trainer/checkpoint-150/tokenizer.json +0 -0
- viclickbait2025-bamibert-s42-20260827T130938Z-ee19aa8e/trainer/checkpoint-150/trainer_state.json +107 -0
- viclickbait2025-bamibert-s42-20260827T130938Z-ee19aa8e/trainer/checkpoint-375/special_tokens_map.json +51 -0
- viclickbait2025-bamibert-s42-20260827T130938Z-ee19aa8e/trainer/checkpoint-375/tokenizer.json +0 -0
- viclickbait2025-bamibert-s42-20260827T130938Z-ee19aa8e/trainer/checkpoint-375/tokenizer_config.json +55 -0
- viclickbait2025-bamibert-s42-20260827T130938Z-ee19aa8e/trainer/checkpoint-375/trainer_state.json +203 -0
- viclickbait2025-bamibert-s43-20260827T131028Z-32d744c6/trainer/checkpoint-1050/special_tokens_map.json +51 -0
- viclickbait2025-bamibert-s43-20260827T131028Z-32d744c6/trainer/checkpoint-1050/tokenizer.json +0 -0
- viclickbait2025-bamibert-s43-20260827T131028Z-32d744c6/trainer/checkpoint-1050/tokenizer_config.json +55 -0
- viclickbait2025-bamibert-s43-20260827T131028Z-32d744c6/trainer/checkpoint-1050/trainer_state.json +491 -0
- viclickbait2025-bamibert-s43-20260827T131028Z-32d744c6/trainer/checkpoint-1275/special_tokens_map.json +51 -0
- viclickbait2025-bamibert-s43-20260827T131028Z-32d744c6/trainer/checkpoint-1275/tokenizer.json +0 -0
- viclickbait2025-bamibert-s43-20260827T131028Z-32d744c6/trainer/checkpoint-1275/tokenizer_config.json +55 -0
- viclickbait2025-bamibert-s43-20260827T131028Z-32d744c6/trainer/checkpoint-1275/trainer_state.json +587 -0
- viclickbait2025-bamibert-s44-20260827T131301Z-fe731097/trainer/checkpoint-300/special_tokens_map.json +51 -0
- viclickbait2025-bamibert-s44-20260827T131301Z-fe731097/trainer/checkpoint-300/tokenizer.json +0 -0
- viclickbait2025-bamibert-s44-20260827T131301Z-fe731097/trainer/checkpoint-300/tokenizer_config.json +55 -0
- viclickbait2025-bamibert-s44-20260827T131301Z-fe731097/trainer/checkpoint-300/trainer_state.json +171 -0
- viclickbait2025-bamibert-s44-20260827T131301Z-fe731097/trainer/checkpoint-525/special_tokens_map.json +51 -0
- viclickbait2025-bamibert-s44-20260827T131301Z-fe731097/trainer/checkpoint-525/tokenizer.json +0 -0
- viclickbait2025-bamibert-s44-20260827T131301Z-fe731097/trainer/checkpoint-525/tokenizer_config.json +55 -0
- viclickbait2025-bamibert-s44-20260827T131301Z-fe731097/trainer/checkpoint-525/trainer_state.json +267 -0
- vifactcheck-bamibert-s42-20260827T131409Z-6235c159/trainer/checkpoint-1141/special_tokens_map.json +51 -0
- vifactcheck-bamibert-s42-20260827T131409Z-6235c159/trainer/checkpoint-1141/tokenizer.json +0 -0
- vifactcheck-bamibert-s42-20260827T131409Z-6235c159/trainer/checkpoint-1141/tokenizer_config.json +55 -0
- vifactcheck-bamibert-s42-20260827T131409Z-6235c159/trainer/checkpoint-1141/trainer_state.json +435 -0
- vifactcheck-bamibert-s42-20260827T131409Z-6235c159/trainer/checkpoint-652/special_tokens_map.json +51 -0
- vifactcheck-bamibert-s42-20260827T131409Z-6235c159/trainer/checkpoint-652/tokenizer.json +0 -0
- vifactcheck-bamibert-s42-20260827T131409Z-6235c159/trainer/checkpoint-652/tokenizer_config.json +55 -0
- vifactcheck-bamibert-s42-20260827T131409Z-6235c159/trainer/checkpoint-652/trainer_state.json +269 -0
- vifactcheck-bamibert-s43-20260827T131648Z-3858bb98/trainer/checkpoint-1304/special_tokens_map.json +51 -0
- vifactcheck-bamibert-s43-20260827T131648Z-3858bb98/trainer/checkpoint-1304/tokenizer.json +0 -0
- vifactcheck-bamibert-s43-20260827T131648Z-3858bb98/trainer/checkpoint-1304/tokenizer_config.json +55 -0
- vifactcheck-bamibert-s43-20260827T131648Z-3858bb98/trainer/checkpoint-1304/trainer_state.json +495 -0
- vifactcheck-bamibert-s43-20260827T131648Z-3858bb98/trainer/checkpoint-815/special_tokens_map.json +51 -0
- vifactcheck-bamibert-s43-20260827T131648Z-3858bb98/trainer/checkpoint-815/tokenizer.json +0 -0
- vifactcheck-bamibert-s43-20260827T131648Z-3858bb98/trainer/checkpoint-815/tokenizer_config.json +55 -0
- vifactcheck-bamibert-s43-20260827T131648Z-3858bb98/trainer/checkpoint-815/trainer_state.json +322 -0
- vifactcheck-bamibert-s44-20260827T131947Z-496d6871/trainer/checkpoint-1304/special_tokens_map.json +51 -0
- vifactcheck-bamibert-s44-20260827T131947Z-496d6871/trainer/checkpoint-1304/tokenizer.json +0 -0
- vifactcheck-bamibert-s44-20260827T131947Z-496d6871/trainer/checkpoint-1304/tokenizer_config.json +55 -0
- vifactcheck-bamibert-s44-20260827T131947Z-496d6871/trainer/checkpoint-1304/trainer_state.json +495 -0
- vifactcheck-bamibert-s44-20260827T131947Z-496d6871/trainer/checkpoint-1793/special_tokens_map.json +51 -0
- vifactcheck-bamibert-s44-20260827T131947Z-496d6871/trainer/checkpoint-1793/tokenizer.json +0 -0
- vifactcheck-bamibert-s44-20260827T131947Z-496d6871/trainer/checkpoint-1793/tokenizer_config.json +55 -0
uit_vsmec-bamibert-s42-20260827T141858Z-45c80ccd/trainer/checkpoint-1044/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 |
+
}
|
uit_vsmec-bamibert-s42-20260827T141858Z-45c80ccd/trainer/checkpoint-1044/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
uit_vsmec-bamibert-s42-20260827T141858Z-45c80ccd/trainer/checkpoint-1044/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 |
+
}
|
uit_vsmec-bamibert-s42-20260827T141858Z-45c80ccd/trainer/checkpoint-1044/trainer_state.json
ADDED
|
@@ -0,0 +1,396 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1044,
|
| 3 |
+
"best_metric": 0.6094265089334862,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/uit_vsmec-bamibert-s42-20260827T141858Z-45c80ccd/trainer/checkpoint-1044",
|
| 5 |
+
"epoch": 6.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1044,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.1440922190201729,
|
| 14 |
+
"grad_norm": 24.969436645507812,
|
| 15 |
+
"learning_rate": 1.3793103448275862e-06,
|
| 16 |
+
"loss": 8.2303,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.2881844380403458,
|
| 21 |
+
"grad_norm": 28.5682430267334,
|
| 22 |
+
"learning_rate": 2.8160919540229887e-06,
|
| 23 |
+
"loss": 7.8085,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.4322766570605187,
|
| 28 |
+
"grad_norm": 32.34672164916992,
|
| 29 |
+
"learning_rate": 4.252873563218391e-06,
|
| 30 |
+
"loss": 7.3048,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5763688760806917,
|
| 35 |
+
"grad_norm": 55.207969665527344,
|
| 36 |
+
"learning_rate": 5.689655172413794e-06,
|
| 37 |
+
"loss": 7.1673,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.7204610951008645,
|
| 42 |
+
"grad_norm": 32.265193939208984,
|
| 43 |
+
"learning_rate": 7.126436781609196e-06,
|
| 44 |
+
"loss": 7.107,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8645533141210374,
|
| 49 |
+
"grad_norm": 23.83921241760254,
|
| 50 |
+
"learning_rate": 8.563218390804599e-06,
|
| 51 |
+
"loss": 7.1373,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 1.0,
|
| 56 |
+
"eval_accuracy": 0.36443148688046645,
|
| 57 |
+
"eval_loss": 1.6914347410202026,
|
| 58 |
+
"eval_macro_f1": 0.1628672956077675,
|
| 59 |
+
"eval_runtime": 0.4735,
|
| 60 |
+
"eval_samples_per_second": 1448.717,
|
| 61 |
+
"eval_steps_per_second": 90.809,
|
| 62 |
+
"eval_weighted_f1": 0.2614538694086085,
|
| 63 |
+
"step": 174
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 1.005763688760807,
|
| 67 |
+
"grad_norm": 24.943424224853516,
|
| 68 |
+
"learning_rate": 1e-05,
|
| 69 |
+
"loss": 6.7753,
|
| 70 |
+
"step": 175
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.1498559077809798,
|
| 74 |
+
"grad_norm": 31.301626205444336,
|
| 75 |
+
"learning_rate": 1.1436781609195405e-05,
|
| 76 |
+
"loss": 6.9149,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.2939481268011528,
|
| 81 |
+
"grad_norm": 40.84605407714844,
|
| 82 |
+
"learning_rate": 1.2873563218390805e-05,
|
| 83 |
+
"loss": 6.3964,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.4380403458213258,
|
| 88 |
+
"grad_norm": 44.23533630371094,
|
| 89 |
+
"learning_rate": 1.4310344827586209e-05,
|
| 90 |
+
"loss": 5.974,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.5821325648414986,
|
| 95 |
+
"grad_norm": 52.83032989501953,
|
| 96 |
+
"learning_rate": 1.574712643678161e-05,
|
| 97 |
+
"loss": 5.9633,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.7262247838616713,
|
| 102 |
+
"grad_norm": 38.67118835449219,
|
| 103 |
+
"learning_rate": 1.7183908045977015e-05,
|
| 104 |
+
"loss": 5.8623,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.8703170028818443,
|
| 109 |
+
"grad_norm": 50.40724182128906,
|
| 110 |
+
"learning_rate": 1.8620689655172415e-05,
|
| 111 |
+
"loss": 5.4069,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 2.0,
|
| 116 |
+
"eval_accuracy": 0.5276967930029155,
|
| 117 |
+
"eval_loss": 1.3374756574630737,
|
| 118 |
+
"eval_macro_f1": 0.40793641830715804,
|
| 119 |
+
"eval_runtime": 0.4637,
|
| 120 |
+
"eval_samples_per_second": 1479.29,
|
| 121 |
+
"eval_steps_per_second": 92.725,
|
| 122 |
+
"eval_weighted_f1": 0.5010245540069785,
|
| 123 |
+
"step": 348
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 2.011527377521614,
|
| 127 |
+
"grad_norm": 47.65451431274414,
|
| 128 |
+
"learning_rate": 1.999361430395913e-05,
|
| 129 |
+
"loss": 5.2501,
|
| 130 |
+
"step": 350
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.155619596541787,
|
| 134 |
+
"grad_norm": 47.411903381347656,
|
| 135 |
+
"learning_rate": 1.983397190293742e-05,
|
| 136 |
+
"loss": 4.9222,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.2997118155619596,
|
| 141 |
+
"grad_norm": 69.1363296508789,
|
| 142 |
+
"learning_rate": 1.967432950191571e-05,
|
| 143 |
+
"loss": 4.6496,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.443804034582133,
|
| 148 |
+
"grad_norm": 45.271759033203125,
|
| 149 |
+
"learning_rate": 1.9514687100894e-05,
|
| 150 |
+
"loss": 4.4533,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.5878962536023056,
|
| 155 |
+
"grad_norm": 51.77754211425781,
|
| 156 |
+
"learning_rate": 1.9355044699872286e-05,
|
| 157 |
+
"loss": 4.6899,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.7319884726224783,
|
| 162 |
+
"grad_norm": 55.15814208984375,
|
| 163 |
+
"learning_rate": 1.9195402298850576e-05,
|
| 164 |
+
"loss": 4.5201,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.8760806916426516,
|
| 169 |
+
"grad_norm": 71.75881958007812,
|
| 170 |
+
"learning_rate": 1.9035759897828865e-05,
|
| 171 |
+
"loss": 4.5395,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 3.0,
|
| 176 |
+
"eval_accuracy": 0.6107871720116618,
|
| 177 |
+
"eval_loss": 1.0936297178268433,
|
| 178 |
+
"eval_macro_f1": 0.5251613977734367,
|
| 179 |
+
"eval_runtime": 0.4465,
|
| 180 |
+
"eval_samples_per_second": 1536.31,
|
| 181 |
+
"eval_steps_per_second": 96.299,
|
| 182 |
+
"eval_weighted_f1": 0.6026319899168403,
|
| 183 |
+
"step": 522
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 3.0172910662824206,
|
| 187 |
+
"grad_norm": 57.012672424316406,
|
| 188 |
+
"learning_rate": 1.8876117496807155e-05,
|
| 189 |
+
"loss": 4.2391,
|
| 190 |
+
"step": 525
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.161383285302594,
|
| 194 |
+
"grad_norm": 308.2218933105469,
|
| 195 |
+
"learning_rate": 1.871647509578544e-05,
|
| 196 |
+
"loss": 3.8055,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.3054755043227666,
|
| 201 |
+
"grad_norm": 67.54936981201172,
|
| 202 |
+
"learning_rate": 1.855683269476373e-05,
|
| 203 |
+
"loss": 3.853,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.4495677233429394,
|
| 208 |
+
"grad_norm": 59.900535583496094,
|
| 209 |
+
"learning_rate": 1.839719029374202e-05,
|
| 210 |
+
"loss": 3.6998,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.5936599423631126,
|
| 215 |
+
"grad_norm": 46.4504280090332,
|
| 216 |
+
"learning_rate": 1.823754789272031e-05,
|
| 217 |
+
"loss": 3.7186,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.7377521613832854,
|
| 222 |
+
"grad_norm": 60.597259521484375,
|
| 223 |
+
"learning_rate": 1.8077905491698596e-05,
|
| 224 |
+
"loss": 3.5711,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.881844380403458,
|
| 229 |
+
"grad_norm": 54.637271881103516,
|
| 230 |
+
"learning_rate": 1.7918263090676886e-05,
|
| 231 |
+
"loss": 3.7265,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 4.0,
|
| 236 |
+
"eval_accuracy": 0.607871720116618,
|
| 237 |
+
"eval_loss": 1.0739258527755737,
|
| 238 |
+
"eval_macro_f1": 0.5137606619731833,
|
| 239 |
+
"eval_runtime": 0.4738,
|
| 240 |
+
"eval_samples_per_second": 1448.0,
|
| 241 |
+
"eval_steps_per_second": 90.764,
|
| 242 |
+
"eval_weighted_f1": 0.5961336310004652,
|
| 243 |
+
"step": 696
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 4.023054755043228,
|
| 247 |
+
"grad_norm": 51.5903434753418,
|
| 248 |
+
"learning_rate": 1.7758620689655175e-05,
|
| 249 |
+
"loss": 3.4432,
|
| 250 |
+
"step": 700
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 4.1671469740634,
|
| 254 |
+
"grad_norm": 89.19109344482422,
|
| 255 |
+
"learning_rate": 1.759897828863346e-05,
|
| 256 |
+
"loss": 3.2045,
|
| 257 |
+
"step": 725
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.311239193083574,
|
| 261 |
+
"grad_norm": 114.7201156616211,
|
| 262 |
+
"learning_rate": 1.743933588761175e-05,
|
| 263 |
+
"loss": 3.1359,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.455331412103746,
|
| 268 |
+
"grad_norm": 52.149940490722656,
|
| 269 |
+
"learning_rate": 1.7279693486590037e-05,
|
| 270 |
+
"loss": 2.8857,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.599423631123919,
|
| 275 |
+
"grad_norm": 73.68968200683594,
|
| 276 |
+
"learning_rate": 1.7120051085568327e-05,
|
| 277 |
+
"loss": 2.9043,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.743515850144092,
|
| 282 |
+
"grad_norm": 79.21112060546875,
|
| 283 |
+
"learning_rate": 1.6960408684546617e-05,
|
| 284 |
+
"loss": 3.027,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.887608069164266,
|
| 289 |
+
"grad_norm": 71.99787139892578,
|
| 290 |
+
"learning_rate": 1.6800766283524906e-05,
|
| 291 |
+
"loss": 2.9141,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 5.0,
|
| 296 |
+
"eval_accuracy": 0.6486880466472303,
|
| 297 |
+
"eval_loss": 1.0613634586334229,
|
| 298 |
+
"eval_macro_f1": 0.5759776003198281,
|
| 299 |
+
"eval_runtime": 0.4799,
|
| 300 |
+
"eval_samples_per_second": 1429.352,
|
| 301 |
+
"eval_steps_per_second": 89.595,
|
| 302 |
+
"eval_weighted_f1": 0.6342004265554475,
|
| 303 |
+
"step": 870
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 5.028818443804035,
|
| 307 |
+
"grad_norm": 81.31233215332031,
|
| 308 |
+
"learning_rate": 1.6641123882503192e-05,
|
| 309 |
+
"loss": 2.8685,
|
| 310 |
+
"step": 875
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 5.172910662824208,
|
| 314 |
+
"grad_norm": 41.58056640625,
|
| 315 |
+
"learning_rate": 1.6481481481481482e-05,
|
| 316 |
+
"loss": 2.3785,
|
| 317 |
+
"step": 900
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.31700288184438,
|
| 321 |
+
"grad_norm": 62.35469436645508,
|
| 322 |
+
"learning_rate": 1.632183908045977e-05,
|
| 323 |
+
"loss": 2.3479,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.461095100864553,
|
| 328 |
+
"grad_norm": 87.61958312988281,
|
| 329 |
+
"learning_rate": 1.616219667943806e-05,
|
| 330 |
+
"loss": 2.4226,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.605187319884726,
|
| 335 |
+
"grad_norm": 58.473052978515625,
|
| 336 |
+
"learning_rate": 1.6002554278416347e-05,
|
| 337 |
+
"loss": 2.4923,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.749279538904899,
|
| 342 |
+
"grad_norm": 81.3972396850586,
|
| 343 |
+
"learning_rate": 1.5842911877394637e-05,
|
| 344 |
+
"loss": 2.3393,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.893371757925072,
|
| 349 |
+
"grad_norm": 66.13785552978516,
|
| 350 |
+
"learning_rate": 1.5683269476372927e-05,
|
| 351 |
+
"loss": 2.7262,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 6.0,
|
| 356 |
+
"eval_accuracy": 0.6603498542274052,
|
| 357 |
+
"eval_loss": 1.062124490737915,
|
| 358 |
+
"eval_macro_f1": 0.6094265089334862,
|
| 359 |
+
"eval_runtime": 0.4753,
|
| 360 |
+
"eval_samples_per_second": 1443.323,
|
| 361 |
+
"eval_steps_per_second": 90.471,
|
| 362 |
+
"eval_weighted_f1": 0.6560771039927736,
|
| 363 |
+
"step": 1044
|
| 364 |
+
}
|
| 365 |
+
],
|
| 366 |
+
"logging_steps": 25,
|
| 367 |
+
"max_steps": 3480,
|
| 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 |
+
}
|
uit_vsmec-bamibert-s42-20260827T141858Z-45c80ccd/trainer/checkpoint-1566/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 |
+
}
|
viclickbait2025-bamibert-s42-20260827T130938Z-ee19aa8e/trainer/checkpoint-150/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
viclickbait2025-bamibert-s42-20260827T130938Z-ee19aa8e/trainer/checkpoint-150/trainer_state.json
ADDED
|
@@ -0,0 +1,107 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 150,
|
| 3 |
+
"best_metric": 0.803845611395797,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/viclickbait2025-bamibert-s42-20260827T130938Z-ee19aa8e/trainer/checkpoint-150",
|
| 5 |
+
"epoch": 2.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 150,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.33444816053511706,
|
| 14 |
+
"grad_norm": 43.914764404296875,
|
| 15 |
+
"learning_rate": 3.2000000000000003e-06,
|
| 16 |
+
"loss": 2.9851,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.6688963210702341,
|
| 21 |
+
"grad_norm": 28.770753860473633,
|
| 22 |
+
"learning_rate": 6.533333333333334e-06,
|
| 23 |
+
"loss": 2.5564,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 1.0,
|
| 28 |
+
"grad_norm": 22.380184173583984,
|
| 29 |
+
"learning_rate": 9.866666666666668e-06,
|
| 30 |
+
"loss": 2.4211,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 1.0,
|
| 35 |
+
"eval_accuracy": 0.6871345029239766,
|
| 36 |
+
"eval_loss": 0.5814430117607117,
|
| 37 |
+
"eval_macro_f1": 0.4072790294627383,
|
| 38 |
+
"eval_runtime": 0.3031,
|
| 39 |
+
"eval_samples_per_second": 1128.474,
|
| 40 |
+
"eval_steps_per_second": 72.592,
|
| 41 |
+
"eval_weighted_f1": 0.5597109469224766,
|
| 42 |
+
"step": 75
|
| 43 |
+
},
|
| 44 |
+
{
|
| 45 |
+
"epoch": 1.334448160535117,
|
| 46 |
+
"grad_norm": 41.361934661865234,
|
| 47 |
+
"learning_rate": 1.3200000000000002e-05,
|
| 48 |
+
"loss": 2.1983,
|
| 49 |
+
"step": 100
|
| 50 |
+
},
|
| 51 |
+
{
|
| 52 |
+
"epoch": 1.6688963210702341,
|
| 53 |
+
"grad_norm": 29.371566772460938,
|
| 54 |
+
"learning_rate": 1.6533333333333333e-05,
|
| 55 |
+
"loss": 1.9811,
|
| 56 |
+
"step": 125
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 2.0,
|
| 60 |
+
"grad_norm": 33.94392013549805,
|
| 61 |
+
"learning_rate": 1.9866666666666667e-05,
|
| 62 |
+
"loss": 1.7764,
|
| 63 |
+
"step": 150
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 2.0,
|
| 67 |
+
"eval_accuracy": 0.8216374269005848,
|
| 68 |
+
"eval_loss": 0.3956577181816101,
|
| 69 |
+
"eval_macro_f1": 0.803845611395797,
|
| 70 |
+
"eval_runtime": 0.2784,
|
| 71 |
+
"eval_samples_per_second": 1228.261,
|
| 72 |
+
"eval_steps_per_second": 79.011,
|
| 73 |
+
"eval_weighted_f1": 0.8259558287221352,
|
| 74 |
+
"step": 150
|
| 75 |
+
}
|
| 76 |
+
],
|
| 77 |
+
"logging_steps": 25,
|
| 78 |
+
"max_steps": 1500,
|
| 79 |
+
"num_input_tokens_seen": 0,
|
| 80 |
+
"num_train_epochs": 20,
|
| 81 |
+
"save_steps": 500,
|
| 82 |
+
"stateful_callbacks": {
|
| 83 |
+
"EarlyStoppingCallback": {
|
| 84 |
+
"args": {
|
| 85 |
+
"early_stopping_patience": 3,
|
| 86 |
+
"early_stopping_threshold": 0.0
|
| 87 |
+
},
|
| 88 |
+
"attributes": {
|
| 89 |
+
"early_stopping_patience_counter": 0
|
| 90 |
+
}
|
| 91 |
+
},
|
| 92 |
+
"TrainerControl": {
|
| 93 |
+
"args": {
|
| 94 |
+
"should_epoch_stop": false,
|
| 95 |
+
"should_evaluate": false,
|
| 96 |
+
"should_log": false,
|
| 97 |
+
"should_save": true,
|
| 98 |
+
"should_training_stop": false
|
| 99 |
+
},
|
| 100 |
+
"attributes": {}
|
| 101 |
+
}
|
| 102 |
+
},
|
| 103 |
+
"total_flos": 0.0,
|
| 104 |
+
"train_batch_size": 8,
|
| 105 |
+
"trial_name": null,
|
| 106 |
+
"trial_params": null
|
| 107 |
+
}
|
viclickbait2025-bamibert-s42-20260827T130938Z-ee19aa8e/trainer/checkpoint-375/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 |
+
}
|
viclickbait2025-bamibert-s42-20260827T130938Z-ee19aa8e/trainer/checkpoint-375/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
viclickbait2025-bamibert-s42-20260827T130938Z-ee19aa8e/trainer/checkpoint-375/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 |
+
}
|
viclickbait2025-bamibert-s42-20260827T130938Z-ee19aa8e/trainer/checkpoint-375/trainer_state.json
ADDED
|
@@ -0,0 +1,203 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 150,
|
| 3 |
+
"best_metric": 0.803845611395797,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/viclickbait2025-bamibert-s42-20260827T130938Z-ee19aa8e/trainer/checkpoint-150",
|
| 5 |
+
"epoch": 5.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 375,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.33444816053511706,
|
| 14 |
+
"grad_norm": 43.914764404296875,
|
| 15 |
+
"learning_rate": 3.2000000000000003e-06,
|
| 16 |
+
"loss": 2.9851,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.6688963210702341,
|
| 21 |
+
"grad_norm": 28.770753860473633,
|
| 22 |
+
"learning_rate": 6.533333333333334e-06,
|
| 23 |
+
"loss": 2.5564,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 1.0,
|
| 28 |
+
"grad_norm": 22.380184173583984,
|
| 29 |
+
"learning_rate": 9.866666666666668e-06,
|
| 30 |
+
"loss": 2.4211,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 1.0,
|
| 35 |
+
"eval_accuracy": 0.6871345029239766,
|
| 36 |
+
"eval_loss": 0.5814430117607117,
|
| 37 |
+
"eval_macro_f1": 0.4072790294627383,
|
| 38 |
+
"eval_runtime": 0.3031,
|
| 39 |
+
"eval_samples_per_second": 1128.474,
|
| 40 |
+
"eval_steps_per_second": 72.592,
|
| 41 |
+
"eval_weighted_f1": 0.5597109469224766,
|
| 42 |
+
"step": 75
|
| 43 |
+
},
|
| 44 |
+
{
|
| 45 |
+
"epoch": 1.334448160535117,
|
| 46 |
+
"grad_norm": 41.361934661865234,
|
| 47 |
+
"learning_rate": 1.3200000000000002e-05,
|
| 48 |
+
"loss": 2.1983,
|
| 49 |
+
"step": 100
|
| 50 |
+
},
|
| 51 |
+
{
|
| 52 |
+
"epoch": 1.6688963210702341,
|
| 53 |
+
"grad_norm": 29.371566772460938,
|
| 54 |
+
"learning_rate": 1.6533333333333333e-05,
|
| 55 |
+
"loss": 1.9811,
|
| 56 |
+
"step": 125
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 2.0,
|
| 60 |
+
"grad_norm": 33.94392013549805,
|
| 61 |
+
"learning_rate": 1.9866666666666667e-05,
|
| 62 |
+
"loss": 1.7764,
|
| 63 |
+
"step": 150
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 2.0,
|
| 67 |
+
"eval_accuracy": 0.8216374269005848,
|
| 68 |
+
"eval_loss": 0.3956577181816101,
|
| 69 |
+
"eval_macro_f1": 0.803845611395797,
|
| 70 |
+
"eval_runtime": 0.2784,
|
| 71 |
+
"eval_samples_per_second": 1228.261,
|
| 72 |
+
"eval_steps_per_second": 79.011,
|
| 73 |
+
"eval_weighted_f1": 0.8259558287221352,
|
| 74 |
+
"step": 150
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 2.334448160535117,
|
| 78 |
+
"grad_norm": 61.173194885253906,
|
| 79 |
+
"learning_rate": 1.9644444444444447e-05,
|
| 80 |
+
"loss": 1.413,
|
| 81 |
+
"step": 175
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 2.668896321070234,
|
| 85 |
+
"grad_norm": 38.87074661254883,
|
| 86 |
+
"learning_rate": 1.9274074074074076e-05,
|
| 87 |
+
"loss": 1.6102,
|
| 88 |
+
"step": 200
|
| 89 |
+
},
|
| 90 |
+
{
|
| 91 |
+
"epoch": 3.0,
|
| 92 |
+
"grad_norm": 36.45773696899414,
|
| 93 |
+
"learning_rate": 1.8903703703703705e-05,
|
| 94 |
+
"loss": 1.3993,
|
| 95 |
+
"step": 225
|
| 96 |
+
},
|
| 97 |
+
{
|
| 98 |
+
"epoch": 3.0,
|
| 99 |
+
"eval_accuracy": 0.8245614035087719,
|
| 100 |
+
"eval_loss": 0.38344475626945496,
|
| 101 |
+
"eval_macro_f1": 0.7854633656741385,
|
| 102 |
+
"eval_runtime": 0.2904,
|
| 103 |
+
"eval_samples_per_second": 1177.756,
|
| 104 |
+
"eval_steps_per_second": 75.762,
|
| 105 |
+
"eval_weighted_f1": 0.8197410974743651,
|
| 106 |
+
"step": 225
|
| 107 |
+
},
|
| 108 |
+
{
|
| 109 |
+
"epoch": 3.334448160535117,
|
| 110 |
+
"grad_norm": 41.191062927246094,
|
| 111 |
+
"learning_rate": 1.8533333333333334e-05,
|
| 112 |
+
"loss": 1.0063,
|
| 113 |
+
"step": 250
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 3.668896321070234,
|
| 117 |
+
"grad_norm": 62.208927154541016,
|
| 118 |
+
"learning_rate": 1.8162962962962963e-05,
|
| 119 |
+
"loss": 1.3154,
|
| 120 |
+
"step": 275
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"epoch": 4.0,
|
| 124 |
+
"grad_norm": 33.95317077636719,
|
| 125 |
+
"learning_rate": 1.7792592592592596e-05,
|
| 126 |
+
"loss": 1.067,
|
| 127 |
+
"step": 300
|
| 128 |
+
},
|
| 129 |
+
{
|
| 130 |
+
"epoch": 4.0,
|
| 131 |
+
"eval_accuracy": 0.8070175438596491,
|
| 132 |
+
"eval_loss": 0.38094067573547363,
|
| 133 |
+
"eval_macro_f1": 0.7897540983606557,
|
| 134 |
+
"eval_runtime": 0.3024,
|
| 135 |
+
"eval_samples_per_second": 1131.105,
|
| 136 |
+
"eval_steps_per_second": 72.761,
|
| 137 |
+
"eval_weighted_f1": 0.8123022720736267,
|
| 138 |
+
"step": 300
|
| 139 |
+
},
|
| 140 |
+
{
|
| 141 |
+
"epoch": 4.334448160535117,
|
| 142 |
+
"grad_norm": 55.01930236816406,
|
| 143 |
+
"learning_rate": 1.7422222222222222e-05,
|
| 144 |
+
"loss": 0.6486,
|
| 145 |
+
"step": 325
|
| 146 |
+
},
|
| 147 |
+
{
|
| 148 |
+
"epoch": 4.668896321070234,
|
| 149 |
+
"grad_norm": 53.08168411254883,
|
| 150 |
+
"learning_rate": 1.7051851851851854e-05,
|
| 151 |
+
"loss": 0.8595,
|
| 152 |
+
"step": 350
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 5.0,
|
| 156 |
+
"grad_norm": 77.29498291015625,
|
| 157 |
+
"learning_rate": 1.6681481481481484e-05,
|
| 158 |
+
"loss": 0.9119,
|
| 159 |
+
"step": 375
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 5.0,
|
| 163 |
+
"eval_accuracy": 0.8333333333333334,
|
| 164 |
+
"eval_loss": 0.37250635027885437,
|
| 165 |
+
"eval_macro_f1": 0.8036363636363637,
|
| 166 |
+
"eval_runtime": 0.2897,
|
| 167 |
+
"eval_samples_per_second": 1180.658,
|
| 168 |
+
"eval_steps_per_second": 75.949,
|
| 169 |
+
"eval_weighted_f1": 0.8322169059011164,
|
| 170 |
+
"step": 375
|
| 171 |
+
}
|
| 172 |
+
],
|
| 173 |
+
"logging_steps": 25,
|
| 174 |
+
"max_steps": 1500,
|
| 175 |
+
"num_input_tokens_seen": 0,
|
| 176 |
+
"num_train_epochs": 20,
|
| 177 |
+
"save_steps": 500,
|
| 178 |
+
"stateful_callbacks": {
|
| 179 |
+
"EarlyStoppingCallback": {
|
| 180 |
+
"args": {
|
| 181 |
+
"early_stopping_patience": 3,
|
| 182 |
+
"early_stopping_threshold": 0.0
|
| 183 |
+
},
|
| 184 |
+
"attributes": {
|
| 185 |
+
"early_stopping_patience_counter": 3
|
| 186 |
+
}
|
| 187 |
+
},
|
| 188 |
+
"TrainerControl": {
|
| 189 |
+
"args": {
|
| 190 |
+
"should_epoch_stop": false,
|
| 191 |
+
"should_evaluate": false,
|
| 192 |
+
"should_log": false,
|
| 193 |
+
"should_save": true,
|
| 194 |
+
"should_training_stop": true
|
| 195 |
+
},
|
| 196 |
+
"attributes": {}
|
| 197 |
+
}
|
| 198 |
+
},
|
| 199 |
+
"total_flos": 0.0,
|
| 200 |
+
"train_batch_size": 8,
|
| 201 |
+
"trial_name": null,
|
| 202 |
+
"trial_params": null
|
| 203 |
+
}
|
viclickbait2025-bamibert-s43-20260827T131028Z-32d744c6/trainer/checkpoint-1050/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 |
+
}
|
viclickbait2025-bamibert-s43-20260827T131028Z-32d744c6/trainer/checkpoint-1050/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
viclickbait2025-bamibert-s43-20260827T131028Z-32d744c6/trainer/checkpoint-1050/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 |
+
}
|
viclickbait2025-bamibert-s43-20260827T131028Z-32d744c6/trainer/checkpoint-1050/trainer_state.json
ADDED
|
@@ -0,0 +1,491 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1050,
|
| 3 |
+
"best_metric": 0.8323529411764705,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/viclickbait2025-bamibert-s43-20260827T131028Z-32d744c6/trainer/checkpoint-1050",
|
| 5 |
+
"epoch": 14.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1050,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.33444816053511706,
|
| 14 |
+
"grad_norm": 22.471839904785156,
|
| 15 |
+
"learning_rate": 3.2000000000000003e-06,
|
| 16 |
+
"loss": 2.7701,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.6688963210702341,
|
| 21 |
+
"grad_norm": 25.59295082092285,
|
| 22 |
+
"learning_rate": 6.533333333333334e-06,
|
| 23 |
+
"loss": 2.5461,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 1.0,
|
| 28 |
+
"grad_norm": 23.805505752563477,
|
| 29 |
+
"learning_rate": 9.866666666666668e-06,
|
| 30 |
+
"loss": 2.4381,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 1.0,
|
| 35 |
+
"eval_accuracy": 0.6871345029239766,
|
| 36 |
+
"eval_loss": 0.632027268409729,
|
| 37 |
+
"eval_macro_f1": 0.4072790294627383,
|
| 38 |
+
"eval_runtime": 0.2954,
|
| 39 |
+
"eval_samples_per_second": 1157.65,
|
| 40 |
+
"eval_steps_per_second": 74.469,
|
| 41 |
+
"eval_weighted_f1": 0.5597109469224766,
|
| 42 |
+
"step": 75
|
| 43 |
+
},
|
| 44 |
+
{
|
| 45 |
+
"epoch": 1.334448160535117,
|
| 46 |
+
"grad_norm": 48.98332595825195,
|
| 47 |
+
"learning_rate": 1.3200000000000002e-05,
|
| 48 |
+
"loss": 2.4488,
|
| 49 |
+
"step": 100
|
| 50 |
+
},
|
| 51 |
+
{
|
| 52 |
+
"epoch": 1.6688963210702341,
|
| 53 |
+
"grad_norm": 17.144981384277344,
|
| 54 |
+
"learning_rate": 1.6533333333333333e-05,
|
| 55 |
+
"loss": 2.4809,
|
| 56 |
+
"step": 125
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 2.0,
|
| 60 |
+
"grad_norm": 20.95052146911621,
|
| 61 |
+
"learning_rate": 1.9866666666666667e-05,
|
| 62 |
+
"loss": 2.1053,
|
| 63 |
+
"step": 150
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 2.0,
|
| 67 |
+
"eval_accuracy": 0.783625730994152,
|
| 68 |
+
"eval_loss": 0.45853108167648315,
|
| 69 |
+
"eval_macro_f1": 0.6981393129770992,
|
| 70 |
+
"eval_runtime": 0.2893,
|
| 71 |
+
"eval_samples_per_second": 1182.304,
|
| 72 |
+
"eval_steps_per_second": 76.055,
|
| 73 |
+
"eval_weighted_f1": 0.7582616289451364,
|
| 74 |
+
"step": 150
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 2.334448160535117,
|
| 78 |
+
"grad_norm": 22.72088623046875,
|
| 79 |
+
"learning_rate": 1.9644444444444447e-05,
|
| 80 |
+
"loss": 1.7726,
|
| 81 |
+
"step": 175
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 2.668896321070234,
|
| 85 |
+
"grad_norm": 39.299842834472656,
|
| 86 |
+
"learning_rate": 1.9274074074074076e-05,
|
| 87 |
+
"loss": 1.6101,
|
| 88 |
+
"step": 200
|
| 89 |
+
},
|
| 90 |
+
{
|
| 91 |
+
"epoch": 3.0,
|
| 92 |
+
"grad_norm": 42.5814208984375,
|
| 93 |
+
"learning_rate": 1.8903703703703705e-05,
|
| 94 |
+
"loss": 1.6888,
|
| 95 |
+
"step": 225
|
| 96 |
+
},
|
| 97 |
+
{
|
| 98 |
+
"epoch": 3.0,
|
| 99 |
+
"eval_accuracy": 0.7748538011695907,
|
| 100 |
+
"eval_loss": 0.4448682963848114,
|
| 101 |
+
"eval_macro_f1": 0.7628315396046291,
|
| 102 |
+
"eval_runtime": 0.2795,
|
| 103 |
+
"eval_samples_per_second": 1223.523,
|
| 104 |
+
"eval_steps_per_second": 78.706,
|
| 105 |
+
"eval_weighted_f1": 0.7828165977905391,
|
| 106 |
+
"step": 225
|
| 107 |
+
},
|
| 108 |
+
{
|
| 109 |
+
"epoch": 3.334448160535117,
|
| 110 |
+
"grad_norm": 48.096588134765625,
|
| 111 |
+
"learning_rate": 1.8533333333333334e-05,
|
| 112 |
+
"loss": 1.1655,
|
| 113 |
+
"step": 250
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 3.668896321070234,
|
| 117 |
+
"grad_norm": 30.527408599853516,
|
| 118 |
+
"learning_rate": 1.8162962962962963e-05,
|
| 119 |
+
"loss": 1.5311,
|
| 120 |
+
"step": 275
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"epoch": 4.0,
|
| 124 |
+
"grad_norm": 45.529624938964844,
|
| 125 |
+
"learning_rate": 1.7792592592592596e-05,
|
| 126 |
+
"loss": 1.3829,
|
| 127 |
+
"step": 300
|
| 128 |
+
},
|
| 129 |
+
{
|
| 130 |
+
"epoch": 4.0,
|
| 131 |
+
"eval_accuracy": 0.827485380116959,
|
| 132 |
+
"eval_loss": 0.36181196570396423,
|
| 133 |
+
"eval_macro_f1": 0.7896788584413013,
|
| 134 |
+
"eval_runtime": 0.2879,
|
| 135 |
+
"eval_samples_per_second": 1187.708,
|
| 136 |
+
"eval_steps_per_second": 76.402,
|
| 137 |
+
"eval_weighted_f1": 0.8230528913687786,
|
| 138 |
+
"step": 300
|
| 139 |
+
},
|
| 140 |
+
{
|
| 141 |
+
"epoch": 4.334448160535117,
|
| 142 |
+
"grad_norm": 52.226898193359375,
|
| 143 |
+
"learning_rate": 1.7422222222222222e-05,
|
| 144 |
+
"loss": 0.8828,
|
| 145 |
+
"step": 325
|
| 146 |
+
},
|
| 147 |
+
{
|
| 148 |
+
"epoch": 4.668896321070234,
|
| 149 |
+
"grad_norm": 73.34969329833984,
|
| 150 |
+
"learning_rate": 1.7051851851851854e-05,
|
| 151 |
+
"loss": 0.8467,
|
| 152 |
+
"step": 350
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 5.0,
|
| 156 |
+
"grad_norm": 20.253095626831055,
|
| 157 |
+
"learning_rate": 1.6681481481481484e-05,
|
| 158 |
+
"loss": 1.121,
|
| 159 |
+
"step": 375
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 5.0,
|
| 163 |
+
"eval_accuracy": 0.8216374269005848,
|
| 164 |
+
"eval_loss": 0.4209032654762268,
|
| 165 |
+
"eval_macro_f1": 0.7863441311717174,
|
| 166 |
+
"eval_runtime": 0.3263,
|
| 167 |
+
"eval_samples_per_second": 1048.09,
|
| 168 |
+
"eval_steps_per_second": 67.421,
|
| 169 |
+
"eval_weighted_f1": 0.8188444322745594,
|
| 170 |
+
"step": 375
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 5.334448160535117,
|
| 174 |
+
"grad_norm": 64.1552505493164,
|
| 175 |
+
"learning_rate": 1.6311111111111113e-05,
|
| 176 |
+
"loss": 0.69,
|
| 177 |
+
"step": 400
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 5.668896321070234,
|
| 181 |
+
"grad_norm": 37.014774322509766,
|
| 182 |
+
"learning_rate": 1.5940740740740742e-05,
|
| 183 |
+
"loss": 0.6944,
|
| 184 |
+
"step": 425
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 6.0,
|
| 188 |
+
"grad_norm": 22.31513786315918,
|
| 189 |
+
"learning_rate": 1.557037037037037e-05,
|
| 190 |
+
"loss": 0.7292,
|
| 191 |
+
"step": 450
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 6.0,
|
| 195 |
+
"eval_accuracy": 0.8216374269005848,
|
| 196 |
+
"eval_loss": 0.4374345541000366,
|
| 197 |
+
"eval_macro_f1": 0.7951391957578435,
|
| 198 |
+
"eval_runtime": 0.2944,
|
| 199 |
+
"eval_samples_per_second": 1161.754,
|
| 200 |
+
"eval_steps_per_second": 74.733,
|
| 201 |
+
"eval_weighted_f1": 0.8227145907681759,
|
| 202 |
+
"step": 450
|
| 203 |
+
},
|
| 204 |
+
{
|
| 205 |
+
"epoch": 6.334448160535117,
|
| 206 |
+
"grad_norm": 67.88407897949219,
|
| 207 |
+
"learning_rate": 1.5200000000000002e-05,
|
| 208 |
+
"loss": 0.6341,
|
| 209 |
+
"step": 475
|
| 210 |
+
},
|
| 211 |
+
{
|
| 212 |
+
"epoch": 6.668896321070234,
|
| 213 |
+
"grad_norm": 134.65679931640625,
|
| 214 |
+
"learning_rate": 1.4829629629629631e-05,
|
| 215 |
+
"loss": 0.5074,
|
| 216 |
+
"step": 500
|
| 217 |
+
},
|
| 218 |
+
{
|
| 219 |
+
"epoch": 7.0,
|
| 220 |
+
"grad_norm": 91.90208435058594,
|
| 221 |
+
"learning_rate": 1.445925925925926e-05,
|
| 222 |
+
"loss": 0.5943,
|
| 223 |
+
"step": 525
|
| 224 |
+
},
|
| 225 |
+
{
|
| 226 |
+
"epoch": 7.0,
|
| 227 |
+
"eval_accuracy": 0.8304093567251462,
|
| 228 |
+
"eval_loss": 0.72249835729599,
|
| 229 |
+
"eval_macro_f1": 0.7858530897784688,
|
| 230 |
+
"eval_runtime": 0.2773,
|
| 231 |
+
"eval_samples_per_second": 1233.452,
|
| 232 |
+
"eval_steps_per_second": 79.345,
|
| 233 |
+
"eval_weighted_f1": 0.8224120780424092,
|
| 234 |
+
"step": 525
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 7.334448160535117,
|
| 238 |
+
"grad_norm": 104.17916107177734,
|
| 239 |
+
"learning_rate": 1.408888888888889e-05,
|
| 240 |
+
"loss": 0.4513,
|
| 241 |
+
"step": 550
|
| 242 |
+
},
|
| 243 |
+
{
|
| 244 |
+
"epoch": 7.668896321070234,
|
| 245 |
+
"grad_norm": 53.6211051940918,
|
| 246 |
+
"learning_rate": 1.371851851851852e-05,
|
| 247 |
+
"loss": 0.6474,
|
| 248 |
+
"step": 575
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 8.0,
|
| 252 |
+
"grad_norm": 0.9626151919364929,
|
| 253 |
+
"learning_rate": 1.334814814814815e-05,
|
| 254 |
+
"loss": 0.3994,
|
| 255 |
+
"step": 600
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 8.0,
|
| 259 |
+
"eval_accuracy": 0.8362573099415205,
|
| 260 |
+
"eval_loss": 0.6866018176078796,
|
| 261 |
+
"eval_macro_f1": 0.8021487603305786,
|
| 262 |
+
"eval_runtime": 0.2892,
|
| 263 |
+
"eval_samples_per_second": 1182.438,
|
| 264 |
+
"eval_steps_per_second": 76.063,
|
| 265 |
+
"eval_weighted_f1": 0.8328944951911458,
|
| 266 |
+
"step": 600
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 8.334448160535118,
|
| 270 |
+
"grad_norm": 19.21327781677246,
|
| 271 |
+
"learning_rate": 1.2977777777777779e-05,
|
| 272 |
+
"loss": 0.3808,
|
| 273 |
+
"step": 625
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 8.668896321070234,
|
| 277 |
+
"grad_norm": 172.03221130371094,
|
| 278 |
+
"learning_rate": 1.2607407407407408e-05,
|
| 279 |
+
"loss": 0.3507,
|
| 280 |
+
"step": 650
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"epoch": 9.0,
|
| 284 |
+
"grad_norm": 32.312477111816406,
|
| 285 |
+
"learning_rate": 1.2237037037037039e-05,
|
| 286 |
+
"loss": 0.3982,
|
| 287 |
+
"step": 675
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 9.0,
|
| 291 |
+
"eval_accuracy": 0.8333333333333334,
|
| 292 |
+
"eval_loss": 0.6999330520629883,
|
| 293 |
+
"eval_macro_f1": 0.8103751836035913,
|
| 294 |
+
"eval_runtime": 0.2866,
|
| 295 |
+
"eval_samples_per_second": 1193.474,
|
| 296 |
+
"eval_steps_per_second": 76.773,
|
| 297 |
+
"eval_weighted_f1": 0.8350696639851626,
|
| 298 |
+
"step": 675
|
| 299 |
+
},
|
| 300 |
+
{
|
| 301 |
+
"epoch": 9.334448160535118,
|
| 302 |
+
"grad_norm": 0.25637972354888916,
|
| 303 |
+
"learning_rate": 1.186666666666667e-05,
|
| 304 |
+
"loss": 0.19,
|
| 305 |
+
"step": 700
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 9.668896321070234,
|
| 309 |
+
"grad_norm": 159.09768676757812,
|
| 310 |
+
"learning_rate": 1.1496296296296297e-05,
|
| 311 |
+
"loss": 0.3263,
|
| 312 |
+
"step": 725
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"epoch": 10.0,
|
| 316 |
+
"grad_norm": 10.676284790039062,
|
| 317 |
+
"learning_rate": 1.1125925925925926e-05,
|
| 318 |
+
"loss": 0.3227,
|
| 319 |
+
"step": 750
|
| 320 |
+
},
|
| 321 |
+
{
|
| 322 |
+
"epoch": 10.0,
|
| 323 |
+
"eval_accuracy": 0.8362573099415205,
|
| 324 |
+
"eval_loss": 0.870272159576416,
|
| 325 |
+
"eval_macro_f1": 0.8132678132678133,
|
| 326 |
+
"eval_runtime": 0.2787,
|
| 327 |
+
"eval_samples_per_second": 1227.326,
|
| 328 |
+
"eval_steps_per_second": 78.951,
|
| 329 |
+
"eval_weighted_f1": 0.8377899430531008,
|
| 330 |
+
"step": 750
|
| 331 |
+
},
|
| 332 |
+
{
|
| 333 |
+
"epoch": 10.334448160535118,
|
| 334 |
+
"grad_norm": 2.3478646278381348,
|
| 335 |
+
"learning_rate": 1.0755555555555557e-05,
|
| 336 |
+
"loss": 0.2087,
|
| 337 |
+
"step": 775
|
| 338 |
+
},
|
| 339 |
+
{
|
| 340 |
+
"epoch": 10.668896321070234,
|
| 341 |
+
"grad_norm": 0.07605490833520889,
|
| 342 |
+
"learning_rate": 1.0385185185185188e-05,
|
| 343 |
+
"loss": 0.2043,
|
| 344 |
+
"step": 800
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 11.0,
|
| 348 |
+
"grad_norm": 218.12208557128906,
|
| 349 |
+
"learning_rate": 1.0014814814814815e-05,
|
| 350 |
+
"loss": 0.1733,
|
| 351 |
+
"step": 825
|
| 352 |
+
},
|
| 353 |
+
{
|
| 354 |
+
"epoch": 11.0,
|
| 355 |
+
"eval_accuracy": 0.8450292397660819,
|
| 356 |
+
"eval_loss": 0.9843814373016357,
|
| 357 |
+
"eval_macro_f1": 0.8211350023189492,
|
| 358 |
+
"eval_runtime": 0.2991,
|
| 359 |
+
"eval_samples_per_second": 1143.255,
|
| 360 |
+
"eval_steps_per_second": 73.543,
|
| 361 |
+
"eval_weighted_f1": 0.8456027014648131,
|
| 362 |
+
"step": 825
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 11.334448160535118,
|
| 366 |
+
"grad_norm": 18.414342880249023,
|
| 367 |
+
"learning_rate": 9.644444444444444e-06,
|
| 368 |
+
"loss": 0.2544,
|
| 369 |
+
"step": 850
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 11.668896321070234,
|
| 373 |
+
"grad_norm": 0.040541332215070724,
|
| 374 |
+
"learning_rate": 9.274074074074075e-06,
|
| 375 |
+
"loss": 0.2126,
|
| 376 |
+
"step": 875
|
| 377 |
+
},
|
| 378 |
+
{
|
| 379 |
+
"epoch": 12.0,
|
| 380 |
+
"grad_norm": 67.29754638671875,
|
| 381 |
+
"learning_rate": 8.903703703703704e-06,
|
| 382 |
+
"loss": 0.1625,
|
| 383 |
+
"step": 900
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 12.0,
|
| 387 |
+
"eval_accuracy": 0.8450292397660819,
|
| 388 |
+
"eval_loss": 0.9664106965065002,
|
| 389 |
+
"eval_macro_f1": 0.8211350023189492,
|
| 390 |
+
"eval_runtime": 0.2869,
|
| 391 |
+
"eval_samples_per_second": 1191.935,
|
| 392 |
+
"eval_steps_per_second": 76.674,
|
| 393 |
+
"eval_weighted_f1": 0.8456027014648131,
|
| 394 |
+
"step": 900
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 12.334448160535118,
|
| 398 |
+
"grad_norm": 0.0316634401679039,
|
| 399 |
+
"learning_rate": 8.533333333333335e-06,
|
| 400 |
+
"loss": 0.0872,
|
| 401 |
+
"step": 925
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 12.668896321070234,
|
| 405 |
+
"grad_norm": 170.2311553955078,
|
| 406 |
+
"learning_rate": 8.162962962962964e-06,
|
| 407 |
+
"loss": 0.1634,
|
| 408 |
+
"step": 950
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 13.0,
|
| 412 |
+
"grad_norm": 0.35335856676101685,
|
| 413 |
+
"learning_rate": 7.792592592592594e-06,
|
| 414 |
+
"loss": 0.2312,
|
| 415 |
+
"step": 975
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 13.0,
|
| 419 |
+
"eval_accuracy": 0.8391812865497076,
|
| 420 |
+
"eval_loss": 1.0500155687332153,
|
| 421 |
+
"eval_macro_f1": 0.813457628799524,
|
| 422 |
+
"eval_runtime": 0.3003,
|
| 423 |
+
"eval_samples_per_second": 1139.0,
|
| 424 |
+
"eval_steps_per_second": 73.269,
|
| 425 |
+
"eval_weighted_f1": 0.8393838350359296,
|
| 426 |
+
"step": 975
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 13.334448160535118,
|
| 430 |
+
"grad_norm": 505.4784851074219,
|
| 431 |
+
"learning_rate": 7.422222222222223e-06,
|
| 432 |
+
"loss": 0.0728,
|
| 433 |
+
"step": 1000
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 13.668896321070234,
|
| 437 |
+
"grad_norm": 0.28642556071281433,
|
| 438 |
+
"learning_rate": 7.051851851851852e-06,
|
| 439 |
+
"loss": 0.0509,
|
| 440 |
+
"step": 1025
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 14.0,
|
| 444 |
+
"grad_norm": 0.024414321407675743,
|
| 445 |
+
"learning_rate": 6.681481481481482e-06,
|
| 446 |
+
"loss": 0.1217,
|
| 447 |
+
"step": 1050
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 14.0,
|
| 451 |
+
"eval_accuracy": 0.8596491228070176,
|
| 452 |
+
"eval_loss": 1.0319489240646362,
|
| 453 |
+
"eval_macro_f1": 0.8323529411764705,
|
| 454 |
+
"eval_runtime": 0.3216,
|
| 455 |
+
"eval_samples_per_second": 1063.307,
|
| 456 |
+
"eval_steps_per_second": 68.4,
|
| 457 |
+
"eval_weighted_f1": 0.857671138630891,
|
| 458 |
+
"step": 1050
|
| 459 |
+
}
|
| 460 |
+
],
|
| 461 |
+
"logging_steps": 25,
|
| 462 |
+
"max_steps": 1500,
|
| 463 |
+
"num_input_tokens_seen": 0,
|
| 464 |
+
"num_train_epochs": 20,
|
| 465 |
+
"save_steps": 500,
|
| 466 |
+
"stateful_callbacks": {
|
| 467 |
+
"EarlyStoppingCallback": {
|
| 468 |
+
"args": {
|
| 469 |
+
"early_stopping_patience": 3,
|
| 470 |
+
"early_stopping_threshold": 0.0
|
| 471 |
+
},
|
| 472 |
+
"attributes": {
|
| 473 |
+
"early_stopping_patience_counter": 0
|
| 474 |
+
}
|
| 475 |
+
},
|
| 476 |
+
"TrainerControl": {
|
| 477 |
+
"args": {
|
| 478 |
+
"should_epoch_stop": false,
|
| 479 |
+
"should_evaluate": false,
|
| 480 |
+
"should_log": false,
|
| 481 |
+
"should_save": true,
|
| 482 |
+
"should_training_stop": false
|
| 483 |
+
},
|
| 484 |
+
"attributes": {}
|
| 485 |
+
}
|
| 486 |
+
},
|
| 487 |
+
"total_flos": 0.0,
|
| 488 |
+
"train_batch_size": 8,
|
| 489 |
+
"trial_name": null,
|
| 490 |
+
"trial_params": null
|
| 491 |
+
}
|
viclickbait2025-bamibert-s43-20260827T131028Z-32d744c6/trainer/checkpoint-1275/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 |
+
}
|
viclickbait2025-bamibert-s43-20260827T131028Z-32d744c6/trainer/checkpoint-1275/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
viclickbait2025-bamibert-s43-20260827T131028Z-32d744c6/trainer/checkpoint-1275/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 |
+
}
|
viclickbait2025-bamibert-s43-20260827T131028Z-32d744c6/trainer/checkpoint-1275/trainer_state.json
ADDED
|
@@ -0,0 +1,587 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1050,
|
| 3 |
+
"best_metric": 0.8323529411764705,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/viclickbait2025-bamibert-s43-20260827T131028Z-32d744c6/trainer/checkpoint-1050",
|
| 5 |
+
"epoch": 17.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1275,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.33444816053511706,
|
| 14 |
+
"grad_norm": 22.471839904785156,
|
| 15 |
+
"learning_rate": 3.2000000000000003e-06,
|
| 16 |
+
"loss": 2.7701,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.6688963210702341,
|
| 21 |
+
"grad_norm": 25.59295082092285,
|
| 22 |
+
"learning_rate": 6.533333333333334e-06,
|
| 23 |
+
"loss": 2.5461,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 1.0,
|
| 28 |
+
"grad_norm": 23.805505752563477,
|
| 29 |
+
"learning_rate": 9.866666666666668e-06,
|
| 30 |
+
"loss": 2.4381,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 1.0,
|
| 35 |
+
"eval_accuracy": 0.6871345029239766,
|
| 36 |
+
"eval_loss": 0.632027268409729,
|
| 37 |
+
"eval_macro_f1": 0.4072790294627383,
|
| 38 |
+
"eval_runtime": 0.2954,
|
| 39 |
+
"eval_samples_per_second": 1157.65,
|
| 40 |
+
"eval_steps_per_second": 74.469,
|
| 41 |
+
"eval_weighted_f1": 0.5597109469224766,
|
| 42 |
+
"step": 75
|
| 43 |
+
},
|
| 44 |
+
{
|
| 45 |
+
"epoch": 1.334448160535117,
|
| 46 |
+
"grad_norm": 48.98332595825195,
|
| 47 |
+
"learning_rate": 1.3200000000000002e-05,
|
| 48 |
+
"loss": 2.4488,
|
| 49 |
+
"step": 100
|
| 50 |
+
},
|
| 51 |
+
{
|
| 52 |
+
"epoch": 1.6688963210702341,
|
| 53 |
+
"grad_norm": 17.144981384277344,
|
| 54 |
+
"learning_rate": 1.6533333333333333e-05,
|
| 55 |
+
"loss": 2.4809,
|
| 56 |
+
"step": 125
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 2.0,
|
| 60 |
+
"grad_norm": 20.95052146911621,
|
| 61 |
+
"learning_rate": 1.9866666666666667e-05,
|
| 62 |
+
"loss": 2.1053,
|
| 63 |
+
"step": 150
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 2.0,
|
| 67 |
+
"eval_accuracy": 0.783625730994152,
|
| 68 |
+
"eval_loss": 0.45853108167648315,
|
| 69 |
+
"eval_macro_f1": 0.6981393129770992,
|
| 70 |
+
"eval_runtime": 0.2893,
|
| 71 |
+
"eval_samples_per_second": 1182.304,
|
| 72 |
+
"eval_steps_per_second": 76.055,
|
| 73 |
+
"eval_weighted_f1": 0.7582616289451364,
|
| 74 |
+
"step": 150
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 2.334448160535117,
|
| 78 |
+
"grad_norm": 22.72088623046875,
|
| 79 |
+
"learning_rate": 1.9644444444444447e-05,
|
| 80 |
+
"loss": 1.7726,
|
| 81 |
+
"step": 175
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 2.668896321070234,
|
| 85 |
+
"grad_norm": 39.299842834472656,
|
| 86 |
+
"learning_rate": 1.9274074074074076e-05,
|
| 87 |
+
"loss": 1.6101,
|
| 88 |
+
"step": 200
|
| 89 |
+
},
|
| 90 |
+
{
|
| 91 |
+
"epoch": 3.0,
|
| 92 |
+
"grad_norm": 42.5814208984375,
|
| 93 |
+
"learning_rate": 1.8903703703703705e-05,
|
| 94 |
+
"loss": 1.6888,
|
| 95 |
+
"step": 225
|
| 96 |
+
},
|
| 97 |
+
{
|
| 98 |
+
"epoch": 3.0,
|
| 99 |
+
"eval_accuracy": 0.7748538011695907,
|
| 100 |
+
"eval_loss": 0.4448682963848114,
|
| 101 |
+
"eval_macro_f1": 0.7628315396046291,
|
| 102 |
+
"eval_runtime": 0.2795,
|
| 103 |
+
"eval_samples_per_second": 1223.523,
|
| 104 |
+
"eval_steps_per_second": 78.706,
|
| 105 |
+
"eval_weighted_f1": 0.7828165977905391,
|
| 106 |
+
"step": 225
|
| 107 |
+
},
|
| 108 |
+
{
|
| 109 |
+
"epoch": 3.334448160535117,
|
| 110 |
+
"grad_norm": 48.096588134765625,
|
| 111 |
+
"learning_rate": 1.8533333333333334e-05,
|
| 112 |
+
"loss": 1.1655,
|
| 113 |
+
"step": 250
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 3.668896321070234,
|
| 117 |
+
"grad_norm": 30.527408599853516,
|
| 118 |
+
"learning_rate": 1.8162962962962963e-05,
|
| 119 |
+
"loss": 1.5311,
|
| 120 |
+
"step": 275
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"epoch": 4.0,
|
| 124 |
+
"grad_norm": 45.529624938964844,
|
| 125 |
+
"learning_rate": 1.7792592592592596e-05,
|
| 126 |
+
"loss": 1.3829,
|
| 127 |
+
"step": 300
|
| 128 |
+
},
|
| 129 |
+
{
|
| 130 |
+
"epoch": 4.0,
|
| 131 |
+
"eval_accuracy": 0.827485380116959,
|
| 132 |
+
"eval_loss": 0.36181196570396423,
|
| 133 |
+
"eval_macro_f1": 0.7896788584413013,
|
| 134 |
+
"eval_runtime": 0.2879,
|
| 135 |
+
"eval_samples_per_second": 1187.708,
|
| 136 |
+
"eval_steps_per_second": 76.402,
|
| 137 |
+
"eval_weighted_f1": 0.8230528913687786,
|
| 138 |
+
"step": 300
|
| 139 |
+
},
|
| 140 |
+
{
|
| 141 |
+
"epoch": 4.334448160535117,
|
| 142 |
+
"grad_norm": 52.226898193359375,
|
| 143 |
+
"learning_rate": 1.7422222222222222e-05,
|
| 144 |
+
"loss": 0.8828,
|
| 145 |
+
"step": 325
|
| 146 |
+
},
|
| 147 |
+
{
|
| 148 |
+
"epoch": 4.668896321070234,
|
| 149 |
+
"grad_norm": 73.34969329833984,
|
| 150 |
+
"learning_rate": 1.7051851851851854e-05,
|
| 151 |
+
"loss": 0.8467,
|
| 152 |
+
"step": 350
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 5.0,
|
| 156 |
+
"grad_norm": 20.253095626831055,
|
| 157 |
+
"learning_rate": 1.6681481481481484e-05,
|
| 158 |
+
"loss": 1.121,
|
| 159 |
+
"step": 375
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 5.0,
|
| 163 |
+
"eval_accuracy": 0.8216374269005848,
|
| 164 |
+
"eval_loss": 0.4209032654762268,
|
| 165 |
+
"eval_macro_f1": 0.7863441311717174,
|
| 166 |
+
"eval_runtime": 0.3263,
|
| 167 |
+
"eval_samples_per_second": 1048.09,
|
| 168 |
+
"eval_steps_per_second": 67.421,
|
| 169 |
+
"eval_weighted_f1": 0.8188444322745594,
|
| 170 |
+
"step": 375
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 5.334448160535117,
|
| 174 |
+
"grad_norm": 64.1552505493164,
|
| 175 |
+
"learning_rate": 1.6311111111111113e-05,
|
| 176 |
+
"loss": 0.69,
|
| 177 |
+
"step": 400
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 5.668896321070234,
|
| 181 |
+
"grad_norm": 37.014774322509766,
|
| 182 |
+
"learning_rate": 1.5940740740740742e-05,
|
| 183 |
+
"loss": 0.6944,
|
| 184 |
+
"step": 425
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 6.0,
|
| 188 |
+
"grad_norm": 22.31513786315918,
|
| 189 |
+
"learning_rate": 1.557037037037037e-05,
|
| 190 |
+
"loss": 0.7292,
|
| 191 |
+
"step": 450
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 6.0,
|
| 195 |
+
"eval_accuracy": 0.8216374269005848,
|
| 196 |
+
"eval_loss": 0.4374345541000366,
|
| 197 |
+
"eval_macro_f1": 0.7951391957578435,
|
| 198 |
+
"eval_runtime": 0.2944,
|
| 199 |
+
"eval_samples_per_second": 1161.754,
|
| 200 |
+
"eval_steps_per_second": 74.733,
|
| 201 |
+
"eval_weighted_f1": 0.8227145907681759,
|
| 202 |
+
"step": 450
|
| 203 |
+
},
|
| 204 |
+
{
|
| 205 |
+
"epoch": 6.334448160535117,
|
| 206 |
+
"grad_norm": 67.88407897949219,
|
| 207 |
+
"learning_rate": 1.5200000000000002e-05,
|
| 208 |
+
"loss": 0.6341,
|
| 209 |
+
"step": 475
|
| 210 |
+
},
|
| 211 |
+
{
|
| 212 |
+
"epoch": 6.668896321070234,
|
| 213 |
+
"grad_norm": 134.65679931640625,
|
| 214 |
+
"learning_rate": 1.4829629629629631e-05,
|
| 215 |
+
"loss": 0.5074,
|
| 216 |
+
"step": 500
|
| 217 |
+
},
|
| 218 |
+
{
|
| 219 |
+
"epoch": 7.0,
|
| 220 |
+
"grad_norm": 91.90208435058594,
|
| 221 |
+
"learning_rate": 1.445925925925926e-05,
|
| 222 |
+
"loss": 0.5943,
|
| 223 |
+
"step": 525
|
| 224 |
+
},
|
| 225 |
+
{
|
| 226 |
+
"epoch": 7.0,
|
| 227 |
+
"eval_accuracy": 0.8304093567251462,
|
| 228 |
+
"eval_loss": 0.72249835729599,
|
| 229 |
+
"eval_macro_f1": 0.7858530897784688,
|
| 230 |
+
"eval_runtime": 0.2773,
|
| 231 |
+
"eval_samples_per_second": 1233.452,
|
| 232 |
+
"eval_steps_per_second": 79.345,
|
| 233 |
+
"eval_weighted_f1": 0.8224120780424092,
|
| 234 |
+
"step": 525
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 7.334448160535117,
|
| 238 |
+
"grad_norm": 104.17916107177734,
|
| 239 |
+
"learning_rate": 1.408888888888889e-05,
|
| 240 |
+
"loss": 0.4513,
|
| 241 |
+
"step": 550
|
| 242 |
+
},
|
| 243 |
+
{
|
| 244 |
+
"epoch": 7.668896321070234,
|
| 245 |
+
"grad_norm": 53.6211051940918,
|
| 246 |
+
"learning_rate": 1.371851851851852e-05,
|
| 247 |
+
"loss": 0.6474,
|
| 248 |
+
"step": 575
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 8.0,
|
| 252 |
+
"grad_norm": 0.9626151919364929,
|
| 253 |
+
"learning_rate": 1.334814814814815e-05,
|
| 254 |
+
"loss": 0.3994,
|
| 255 |
+
"step": 600
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 8.0,
|
| 259 |
+
"eval_accuracy": 0.8362573099415205,
|
| 260 |
+
"eval_loss": 0.6866018176078796,
|
| 261 |
+
"eval_macro_f1": 0.8021487603305786,
|
| 262 |
+
"eval_runtime": 0.2892,
|
| 263 |
+
"eval_samples_per_second": 1182.438,
|
| 264 |
+
"eval_steps_per_second": 76.063,
|
| 265 |
+
"eval_weighted_f1": 0.8328944951911458,
|
| 266 |
+
"step": 600
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 8.334448160535118,
|
| 270 |
+
"grad_norm": 19.21327781677246,
|
| 271 |
+
"learning_rate": 1.2977777777777779e-05,
|
| 272 |
+
"loss": 0.3808,
|
| 273 |
+
"step": 625
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 8.668896321070234,
|
| 277 |
+
"grad_norm": 172.03221130371094,
|
| 278 |
+
"learning_rate": 1.2607407407407408e-05,
|
| 279 |
+
"loss": 0.3507,
|
| 280 |
+
"step": 650
|
| 281 |
+
},
|
| 282 |
+
{
|
| 283 |
+
"epoch": 9.0,
|
| 284 |
+
"grad_norm": 32.312477111816406,
|
| 285 |
+
"learning_rate": 1.2237037037037039e-05,
|
| 286 |
+
"loss": 0.3982,
|
| 287 |
+
"step": 675
|
| 288 |
+
},
|
| 289 |
+
{
|
| 290 |
+
"epoch": 9.0,
|
| 291 |
+
"eval_accuracy": 0.8333333333333334,
|
| 292 |
+
"eval_loss": 0.6999330520629883,
|
| 293 |
+
"eval_macro_f1": 0.8103751836035913,
|
| 294 |
+
"eval_runtime": 0.2866,
|
| 295 |
+
"eval_samples_per_second": 1193.474,
|
| 296 |
+
"eval_steps_per_second": 76.773,
|
| 297 |
+
"eval_weighted_f1": 0.8350696639851626,
|
| 298 |
+
"step": 675
|
| 299 |
+
},
|
| 300 |
+
{
|
| 301 |
+
"epoch": 9.334448160535118,
|
| 302 |
+
"grad_norm": 0.25637972354888916,
|
| 303 |
+
"learning_rate": 1.186666666666667e-05,
|
| 304 |
+
"loss": 0.19,
|
| 305 |
+
"step": 700
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 9.668896321070234,
|
| 309 |
+
"grad_norm": 159.09768676757812,
|
| 310 |
+
"learning_rate": 1.1496296296296297e-05,
|
| 311 |
+
"loss": 0.3263,
|
| 312 |
+
"step": 725
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"epoch": 10.0,
|
| 316 |
+
"grad_norm": 10.676284790039062,
|
| 317 |
+
"learning_rate": 1.1125925925925926e-05,
|
| 318 |
+
"loss": 0.3227,
|
| 319 |
+
"step": 750
|
| 320 |
+
},
|
| 321 |
+
{
|
| 322 |
+
"epoch": 10.0,
|
| 323 |
+
"eval_accuracy": 0.8362573099415205,
|
| 324 |
+
"eval_loss": 0.870272159576416,
|
| 325 |
+
"eval_macro_f1": 0.8132678132678133,
|
| 326 |
+
"eval_runtime": 0.2787,
|
| 327 |
+
"eval_samples_per_second": 1227.326,
|
| 328 |
+
"eval_steps_per_second": 78.951,
|
| 329 |
+
"eval_weighted_f1": 0.8377899430531008,
|
| 330 |
+
"step": 750
|
| 331 |
+
},
|
| 332 |
+
{
|
| 333 |
+
"epoch": 10.334448160535118,
|
| 334 |
+
"grad_norm": 2.3478646278381348,
|
| 335 |
+
"learning_rate": 1.0755555555555557e-05,
|
| 336 |
+
"loss": 0.2087,
|
| 337 |
+
"step": 775
|
| 338 |
+
},
|
| 339 |
+
{
|
| 340 |
+
"epoch": 10.668896321070234,
|
| 341 |
+
"grad_norm": 0.07605490833520889,
|
| 342 |
+
"learning_rate": 1.0385185185185188e-05,
|
| 343 |
+
"loss": 0.2043,
|
| 344 |
+
"step": 800
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 11.0,
|
| 348 |
+
"grad_norm": 218.12208557128906,
|
| 349 |
+
"learning_rate": 1.0014814814814815e-05,
|
| 350 |
+
"loss": 0.1733,
|
| 351 |
+
"step": 825
|
| 352 |
+
},
|
| 353 |
+
{
|
| 354 |
+
"epoch": 11.0,
|
| 355 |
+
"eval_accuracy": 0.8450292397660819,
|
| 356 |
+
"eval_loss": 0.9843814373016357,
|
| 357 |
+
"eval_macro_f1": 0.8211350023189492,
|
| 358 |
+
"eval_runtime": 0.2991,
|
| 359 |
+
"eval_samples_per_second": 1143.255,
|
| 360 |
+
"eval_steps_per_second": 73.543,
|
| 361 |
+
"eval_weighted_f1": 0.8456027014648131,
|
| 362 |
+
"step": 825
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 11.334448160535118,
|
| 366 |
+
"grad_norm": 18.414342880249023,
|
| 367 |
+
"learning_rate": 9.644444444444444e-06,
|
| 368 |
+
"loss": 0.2544,
|
| 369 |
+
"step": 850
|
| 370 |
+
},
|
| 371 |
+
{
|
| 372 |
+
"epoch": 11.668896321070234,
|
| 373 |
+
"grad_norm": 0.040541332215070724,
|
| 374 |
+
"learning_rate": 9.274074074074075e-06,
|
| 375 |
+
"loss": 0.2126,
|
| 376 |
+
"step": 875
|
| 377 |
+
},
|
| 378 |
+
{
|
| 379 |
+
"epoch": 12.0,
|
| 380 |
+
"grad_norm": 67.29754638671875,
|
| 381 |
+
"learning_rate": 8.903703703703704e-06,
|
| 382 |
+
"loss": 0.1625,
|
| 383 |
+
"step": 900
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 12.0,
|
| 387 |
+
"eval_accuracy": 0.8450292397660819,
|
| 388 |
+
"eval_loss": 0.9664106965065002,
|
| 389 |
+
"eval_macro_f1": 0.8211350023189492,
|
| 390 |
+
"eval_runtime": 0.2869,
|
| 391 |
+
"eval_samples_per_second": 1191.935,
|
| 392 |
+
"eval_steps_per_second": 76.674,
|
| 393 |
+
"eval_weighted_f1": 0.8456027014648131,
|
| 394 |
+
"step": 900
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 12.334448160535118,
|
| 398 |
+
"grad_norm": 0.0316634401679039,
|
| 399 |
+
"learning_rate": 8.533333333333335e-06,
|
| 400 |
+
"loss": 0.0872,
|
| 401 |
+
"step": 925
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 12.668896321070234,
|
| 405 |
+
"grad_norm": 170.2311553955078,
|
| 406 |
+
"learning_rate": 8.162962962962964e-06,
|
| 407 |
+
"loss": 0.1634,
|
| 408 |
+
"step": 950
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 13.0,
|
| 412 |
+
"grad_norm": 0.35335856676101685,
|
| 413 |
+
"learning_rate": 7.792592592592594e-06,
|
| 414 |
+
"loss": 0.2312,
|
| 415 |
+
"step": 975
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 13.0,
|
| 419 |
+
"eval_accuracy": 0.8391812865497076,
|
| 420 |
+
"eval_loss": 1.0500155687332153,
|
| 421 |
+
"eval_macro_f1": 0.813457628799524,
|
| 422 |
+
"eval_runtime": 0.3003,
|
| 423 |
+
"eval_samples_per_second": 1139.0,
|
| 424 |
+
"eval_steps_per_second": 73.269,
|
| 425 |
+
"eval_weighted_f1": 0.8393838350359296,
|
| 426 |
+
"step": 975
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 13.334448160535118,
|
| 430 |
+
"grad_norm": 505.4784851074219,
|
| 431 |
+
"learning_rate": 7.422222222222223e-06,
|
| 432 |
+
"loss": 0.0728,
|
| 433 |
+
"step": 1000
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 13.668896321070234,
|
| 437 |
+
"grad_norm": 0.28642556071281433,
|
| 438 |
+
"learning_rate": 7.051851851851852e-06,
|
| 439 |
+
"loss": 0.0509,
|
| 440 |
+
"step": 1025
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 14.0,
|
| 444 |
+
"grad_norm": 0.024414321407675743,
|
| 445 |
+
"learning_rate": 6.681481481481482e-06,
|
| 446 |
+
"loss": 0.1217,
|
| 447 |
+
"step": 1050
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 14.0,
|
| 451 |
+
"eval_accuracy": 0.8596491228070176,
|
| 452 |
+
"eval_loss": 1.0319489240646362,
|
| 453 |
+
"eval_macro_f1": 0.8323529411764705,
|
| 454 |
+
"eval_runtime": 0.3216,
|
| 455 |
+
"eval_samples_per_second": 1063.307,
|
| 456 |
+
"eval_steps_per_second": 68.4,
|
| 457 |
+
"eval_weighted_f1": 0.857671138630891,
|
| 458 |
+
"step": 1050
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 14.334448160535118,
|
| 462 |
+
"grad_norm": 0.023380914703011513,
|
| 463 |
+
"learning_rate": 6.311111111111111e-06,
|
| 464 |
+
"loss": 0.1723,
|
| 465 |
+
"step": 1075
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 14.668896321070234,
|
| 469 |
+
"grad_norm": 0.023643262684345245,
|
| 470 |
+
"learning_rate": 5.940740740740741e-06,
|
| 471 |
+
"loss": 0.1006,
|
| 472 |
+
"step": 1100
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 15.0,
|
| 476 |
+
"grad_norm": 0.3580312132835388,
|
| 477 |
+
"learning_rate": 5.57037037037037e-06,
|
| 478 |
+
"loss": 0.0567,
|
| 479 |
+
"step": 1125
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 15.0,
|
| 483 |
+
"eval_accuracy": 0.8450292397660819,
|
| 484 |
+
"eval_loss": 1.115728497505188,
|
| 485 |
+
"eval_macro_f1": 0.8220061864781263,
|
| 486 |
+
"eval_runtime": 0.2819,
|
| 487 |
+
"eval_samples_per_second": 1213.121,
|
| 488 |
+
"eval_steps_per_second": 78.037,
|
| 489 |
+
"eval_weighted_f1": 0.845965136241202,
|
| 490 |
+
"step": 1125
|
| 491 |
+
},
|
| 492 |
+
{
|
| 493 |
+
"epoch": 15.334448160535118,
|
| 494 |
+
"grad_norm": 0.34394970536231995,
|
| 495 |
+
"learning_rate": 5.2e-06,
|
| 496 |
+
"loss": 0.1184,
|
| 497 |
+
"step": 1150
|
| 498 |
+
},
|
| 499 |
+
{
|
| 500 |
+
"epoch": 15.668896321070234,
|
| 501 |
+
"grad_norm": 0.06735589355230331,
|
| 502 |
+
"learning_rate": 4.82962962962963e-06,
|
| 503 |
+
"loss": 0.1176,
|
| 504 |
+
"step": 1175
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 16.0,
|
| 508 |
+
"grad_norm": 0.015308492816984653,
|
| 509 |
+
"learning_rate": 4.45925925925926e-06,
|
| 510 |
+
"loss": 0.0436,
|
| 511 |
+
"step": 1200
|
| 512 |
+
},
|
| 513 |
+
{
|
| 514 |
+
"epoch": 16.0,
|
| 515 |
+
"eval_accuracy": 0.8508771929824561,
|
| 516 |
+
"eval_loss": 1.1239495277404785,
|
| 517 |
+
"eval_macro_f1": 0.821369683438649,
|
| 518 |
+
"eval_runtime": 0.3064,
|
| 519 |
+
"eval_samples_per_second": 1116.203,
|
| 520 |
+
"eval_steps_per_second": 71.803,
|
| 521 |
+
"eval_weighted_f1": 0.8485420663279102,
|
| 522 |
+
"step": 1200
|
| 523 |
+
},
|
| 524 |
+
{
|
| 525 |
+
"epoch": 16.334448160535118,
|
| 526 |
+
"grad_norm": 0.03417233005166054,
|
| 527 |
+
"learning_rate": 4.088888888888889e-06,
|
| 528 |
+
"loss": 0.0997,
|
| 529 |
+
"step": 1225
|
| 530 |
+
},
|
| 531 |
+
{
|
| 532 |
+
"epoch": 16.668896321070235,
|
| 533 |
+
"grad_norm": 0.023284344002604485,
|
| 534 |
+
"learning_rate": 3.718518518518519e-06,
|
| 535 |
+
"loss": 0.0997,
|
| 536 |
+
"step": 1250
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 17.0,
|
| 540 |
+
"grad_norm": 0.036533769220113754,
|
| 541 |
+
"learning_rate": 3.348148148148148e-06,
|
| 542 |
+
"loss": 0.0792,
|
| 543 |
+
"step": 1275
|
| 544 |
+
},
|
| 545 |
+
{
|
| 546 |
+
"epoch": 17.0,
|
| 547 |
+
"eval_accuracy": 0.8508771929824561,
|
| 548 |
+
"eval_loss": 1.1051406860351562,
|
| 549 |
+
"eval_macro_f1": 0.8203392973023084,
|
| 550 |
+
"eval_runtime": 0.2979,
|
| 551 |
+
"eval_samples_per_second": 1147.996,
|
| 552 |
+
"eval_steps_per_second": 73.848,
|
| 553 |
+
"eval_weighted_f1": 0.8480616423169105,
|
| 554 |
+
"step": 1275
|
| 555 |
+
}
|
| 556 |
+
],
|
| 557 |
+
"logging_steps": 25,
|
| 558 |
+
"max_steps": 1500,
|
| 559 |
+
"num_input_tokens_seen": 0,
|
| 560 |
+
"num_train_epochs": 20,
|
| 561 |
+
"save_steps": 500,
|
| 562 |
+
"stateful_callbacks": {
|
| 563 |
+
"EarlyStoppingCallback": {
|
| 564 |
+
"args": {
|
| 565 |
+
"early_stopping_patience": 3,
|
| 566 |
+
"early_stopping_threshold": 0.0
|
| 567 |
+
},
|
| 568 |
+
"attributes": {
|
| 569 |
+
"early_stopping_patience_counter": 3
|
| 570 |
+
}
|
| 571 |
+
},
|
| 572 |
+
"TrainerControl": {
|
| 573 |
+
"args": {
|
| 574 |
+
"should_epoch_stop": false,
|
| 575 |
+
"should_evaluate": false,
|
| 576 |
+
"should_log": false,
|
| 577 |
+
"should_save": true,
|
| 578 |
+
"should_training_stop": true
|
| 579 |
+
},
|
| 580 |
+
"attributes": {}
|
| 581 |
+
}
|
| 582 |
+
},
|
| 583 |
+
"total_flos": 0.0,
|
| 584 |
+
"train_batch_size": 8,
|
| 585 |
+
"trial_name": null,
|
| 586 |
+
"trial_params": null
|
| 587 |
+
}
|
viclickbait2025-bamibert-s44-20260827T131301Z-fe731097/trainer/checkpoint-300/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 |
+
}
|
viclickbait2025-bamibert-s44-20260827T131301Z-fe731097/trainer/checkpoint-300/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
viclickbait2025-bamibert-s44-20260827T131301Z-fe731097/trainer/checkpoint-300/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 |
+
}
|
viclickbait2025-bamibert-s44-20260827T131301Z-fe731097/trainer/checkpoint-300/trainer_state.json
ADDED
|
@@ -0,0 +1,171 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 300,
|
| 3 |
+
"best_metric": 0.8304089105889534,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/viclickbait2025-bamibert-s44-20260827T131301Z-fe731097/trainer/checkpoint-300",
|
| 5 |
+
"epoch": 4.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 300,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.33444816053511706,
|
| 14 |
+
"grad_norm": 26.444822311401367,
|
| 15 |
+
"learning_rate": 3.2000000000000003e-06,
|
| 16 |
+
"loss": 2.5683,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.6688963210702341,
|
| 21 |
+
"grad_norm": 19.731224060058594,
|
| 22 |
+
"learning_rate": 6.533333333333334e-06,
|
| 23 |
+
"loss": 2.5437,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 1.0,
|
| 28 |
+
"grad_norm": 27.24931526184082,
|
| 29 |
+
"learning_rate": 9.866666666666668e-06,
|
| 30 |
+
"loss": 2.4343,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 1.0,
|
| 35 |
+
"eval_accuracy": 0.6871345029239766,
|
| 36 |
+
"eval_loss": 0.6120277643203735,
|
| 37 |
+
"eval_macro_f1": 0.4072790294627383,
|
| 38 |
+
"eval_runtime": 0.2982,
|
| 39 |
+
"eval_samples_per_second": 1147.039,
|
| 40 |
+
"eval_steps_per_second": 73.786,
|
| 41 |
+
"eval_weighted_f1": 0.5597109469224766,
|
| 42 |
+
"step": 75
|
| 43 |
+
},
|
| 44 |
+
{
|
| 45 |
+
"epoch": 1.334448160535117,
|
| 46 |
+
"grad_norm": 20.578792572021484,
|
| 47 |
+
"learning_rate": 1.3200000000000002e-05,
|
| 48 |
+
"loss": 2.35,
|
| 49 |
+
"step": 100
|
| 50 |
+
},
|
| 51 |
+
{
|
| 52 |
+
"epoch": 1.6688963210702341,
|
| 53 |
+
"grad_norm": 65.30794525146484,
|
| 54 |
+
"learning_rate": 1.6533333333333333e-05,
|
| 55 |
+
"loss": 2.0972,
|
| 56 |
+
"step": 125
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 2.0,
|
| 60 |
+
"grad_norm": 32.060848236083984,
|
| 61 |
+
"learning_rate": 1.9866666666666667e-05,
|
| 62 |
+
"loss": 1.8985,
|
| 63 |
+
"step": 150
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 2.0,
|
| 67 |
+
"eval_accuracy": 0.8128654970760234,
|
| 68 |
+
"eval_loss": 0.4043903946876526,
|
| 69 |
+
"eval_macro_f1": 0.78853808401283,
|
| 70 |
+
"eval_runtime": 0.2969,
|
| 71 |
+
"eval_samples_per_second": 1152.064,
|
| 72 |
+
"eval_steps_per_second": 74.109,
|
| 73 |
+
"eval_weighted_f1": 0.8153821260135952,
|
| 74 |
+
"step": 150
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 2.334448160535117,
|
| 78 |
+
"grad_norm": 35.3597526550293,
|
| 79 |
+
"learning_rate": 1.9644444444444447e-05,
|
| 80 |
+
"loss": 1.5521,
|
| 81 |
+
"step": 175
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 2.668896321070234,
|
| 85 |
+
"grad_norm": 56.741188049316406,
|
| 86 |
+
"learning_rate": 1.9274074074074076e-05,
|
| 87 |
+
"loss": 1.636,
|
| 88 |
+
"step": 200
|
| 89 |
+
},
|
| 90 |
+
{
|
| 91 |
+
"epoch": 3.0,
|
| 92 |
+
"grad_norm": 38.166046142578125,
|
| 93 |
+
"learning_rate": 1.8903703703703705e-05,
|
| 94 |
+
"loss": 1.6014,
|
| 95 |
+
"step": 225
|
| 96 |
+
},
|
| 97 |
+
{
|
| 98 |
+
"epoch": 3.0,
|
| 99 |
+
"eval_accuracy": 0.8070175438596491,
|
| 100 |
+
"eval_loss": 0.39469578862190247,
|
| 101 |
+
"eval_macro_f1": 0.7456389452332657,
|
| 102 |
+
"eval_runtime": 0.2959,
|
| 103 |
+
"eval_samples_per_second": 1155.836,
|
| 104 |
+
"eval_steps_per_second": 74.352,
|
| 105 |
+
"eval_weighted_f1": 0.7924035918057484,
|
| 106 |
+
"step": 225
|
| 107 |
+
},
|
| 108 |
+
{
|
| 109 |
+
"epoch": 3.334448160535117,
|
| 110 |
+
"grad_norm": 23.628355026245117,
|
| 111 |
+
"learning_rate": 1.8533333333333334e-05,
|
| 112 |
+
"loss": 1.2159,
|
| 113 |
+
"step": 250
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 3.668896321070234,
|
| 117 |
+
"grad_norm": 36.1817741394043,
|
| 118 |
+
"learning_rate": 1.8162962962962963e-05,
|
| 119 |
+
"loss": 1.0933,
|
| 120 |
+
"step": 275
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"epoch": 4.0,
|
| 124 |
+
"grad_norm": 14.19780445098877,
|
| 125 |
+
"learning_rate": 1.7792592592592596e-05,
|
| 126 |
+
"loss": 1.2817,
|
| 127 |
+
"step": 300
|
| 128 |
+
},
|
| 129 |
+
{
|
| 130 |
+
"epoch": 4.0,
|
| 131 |
+
"eval_accuracy": 0.847953216374269,
|
| 132 |
+
"eval_loss": 0.33539971709251404,
|
| 133 |
+
"eval_macro_f1": 0.8304089105889534,
|
| 134 |
+
"eval_runtime": 0.3044,
|
| 135 |
+
"eval_samples_per_second": 1123.459,
|
| 136 |
+
"eval_steps_per_second": 72.269,
|
| 137 |
+
"eval_weighted_f1": 0.8508241027755025,
|
| 138 |
+
"step": 300
|
| 139 |
+
}
|
| 140 |
+
],
|
| 141 |
+
"logging_steps": 25,
|
| 142 |
+
"max_steps": 1500,
|
| 143 |
+
"num_input_tokens_seen": 0,
|
| 144 |
+
"num_train_epochs": 20,
|
| 145 |
+
"save_steps": 500,
|
| 146 |
+
"stateful_callbacks": {
|
| 147 |
+
"EarlyStoppingCallback": {
|
| 148 |
+
"args": {
|
| 149 |
+
"early_stopping_patience": 3,
|
| 150 |
+
"early_stopping_threshold": 0.0
|
| 151 |
+
},
|
| 152 |
+
"attributes": {
|
| 153 |
+
"early_stopping_patience_counter": 0
|
| 154 |
+
}
|
| 155 |
+
},
|
| 156 |
+
"TrainerControl": {
|
| 157 |
+
"args": {
|
| 158 |
+
"should_epoch_stop": false,
|
| 159 |
+
"should_evaluate": false,
|
| 160 |
+
"should_log": false,
|
| 161 |
+
"should_save": true,
|
| 162 |
+
"should_training_stop": false
|
| 163 |
+
},
|
| 164 |
+
"attributes": {}
|
| 165 |
+
}
|
| 166 |
+
},
|
| 167 |
+
"total_flos": 0.0,
|
| 168 |
+
"train_batch_size": 8,
|
| 169 |
+
"trial_name": null,
|
| 170 |
+
"trial_params": null
|
| 171 |
+
}
|
viclickbait2025-bamibert-s44-20260827T131301Z-fe731097/trainer/checkpoint-525/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 |
+
}
|
viclickbait2025-bamibert-s44-20260827T131301Z-fe731097/trainer/checkpoint-525/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
viclickbait2025-bamibert-s44-20260827T131301Z-fe731097/trainer/checkpoint-525/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 |
+
}
|
viclickbait2025-bamibert-s44-20260827T131301Z-fe731097/trainer/checkpoint-525/trainer_state.json
ADDED
|
@@ -0,0 +1,267 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 300,
|
| 3 |
+
"best_metric": 0.8304089105889534,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/viclickbait2025-bamibert-s44-20260827T131301Z-fe731097/trainer/checkpoint-300",
|
| 5 |
+
"epoch": 7.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 525,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.33444816053511706,
|
| 14 |
+
"grad_norm": 26.444822311401367,
|
| 15 |
+
"learning_rate": 3.2000000000000003e-06,
|
| 16 |
+
"loss": 2.5683,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.6688963210702341,
|
| 21 |
+
"grad_norm": 19.731224060058594,
|
| 22 |
+
"learning_rate": 6.533333333333334e-06,
|
| 23 |
+
"loss": 2.5437,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 1.0,
|
| 28 |
+
"grad_norm": 27.24931526184082,
|
| 29 |
+
"learning_rate": 9.866666666666668e-06,
|
| 30 |
+
"loss": 2.4343,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 1.0,
|
| 35 |
+
"eval_accuracy": 0.6871345029239766,
|
| 36 |
+
"eval_loss": 0.6120277643203735,
|
| 37 |
+
"eval_macro_f1": 0.4072790294627383,
|
| 38 |
+
"eval_runtime": 0.2982,
|
| 39 |
+
"eval_samples_per_second": 1147.039,
|
| 40 |
+
"eval_steps_per_second": 73.786,
|
| 41 |
+
"eval_weighted_f1": 0.5597109469224766,
|
| 42 |
+
"step": 75
|
| 43 |
+
},
|
| 44 |
+
{
|
| 45 |
+
"epoch": 1.334448160535117,
|
| 46 |
+
"grad_norm": 20.578792572021484,
|
| 47 |
+
"learning_rate": 1.3200000000000002e-05,
|
| 48 |
+
"loss": 2.35,
|
| 49 |
+
"step": 100
|
| 50 |
+
},
|
| 51 |
+
{
|
| 52 |
+
"epoch": 1.6688963210702341,
|
| 53 |
+
"grad_norm": 65.30794525146484,
|
| 54 |
+
"learning_rate": 1.6533333333333333e-05,
|
| 55 |
+
"loss": 2.0972,
|
| 56 |
+
"step": 125
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 2.0,
|
| 60 |
+
"grad_norm": 32.060848236083984,
|
| 61 |
+
"learning_rate": 1.9866666666666667e-05,
|
| 62 |
+
"loss": 1.8985,
|
| 63 |
+
"step": 150
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 2.0,
|
| 67 |
+
"eval_accuracy": 0.8128654970760234,
|
| 68 |
+
"eval_loss": 0.4043903946876526,
|
| 69 |
+
"eval_macro_f1": 0.78853808401283,
|
| 70 |
+
"eval_runtime": 0.2969,
|
| 71 |
+
"eval_samples_per_second": 1152.064,
|
| 72 |
+
"eval_steps_per_second": 74.109,
|
| 73 |
+
"eval_weighted_f1": 0.8153821260135952,
|
| 74 |
+
"step": 150
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 2.334448160535117,
|
| 78 |
+
"grad_norm": 35.3597526550293,
|
| 79 |
+
"learning_rate": 1.9644444444444447e-05,
|
| 80 |
+
"loss": 1.5521,
|
| 81 |
+
"step": 175
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 2.668896321070234,
|
| 85 |
+
"grad_norm": 56.741188049316406,
|
| 86 |
+
"learning_rate": 1.9274074074074076e-05,
|
| 87 |
+
"loss": 1.636,
|
| 88 |
+
"step": 200
|
| 89 |
+
},
|
| 90 |
+
{
|
| 91 |
+
"epoch": 3.0,
|
| 92 |
+
"grad_norm": 38.166046142578125,
|
| 93 |
+
"learning_rate": 1.8903703703703705e-05,
|
| 94 |
+
"loss": 1.6014,
|
| 95 |
+
"step": 225
|
| 96 |
+
},
|
| 97 |
+
{
|
| 98 |
+
"epoch": 3.0,
|
| 99 |
+
"eval_accuracy": 0.8070175438596491,
|
| 100 |
+
"eval_loss": 0.39469578862190247,
|
| 101 |
+
"eval_macro_f1": 0.7456389452332657,
|
| 102 |
+
"eval_runtime": 0.2959,
|
| 103 |
+
"eval_samples_per_second": 1155.836,
|
| 104 |
+
"eval_steps_per_second": 74.352,
|
| 105 |
+
"eval_weighted_f1": 0.7924035918057484,
|
| 106 |
+
"step": 225
|
| 107 |
+
},
|
| 108 |
+
{
|
| 109 |
+
"epoch": 3.334448160535117,
|
| 110 |
+
"grad_norm": 23.628355026245117,
|
| 111 |
+
"learning_rate": 1.8533333333333334e-05,
|
| 112 |
+
"loss": 1.2159,
|
| 113 |
+
"step": 250
|
| 114 |
+
},
|
| 115 |
+
{
|
| 116 |
+
"epoch": 3.668896321070234,
|
| 117 |
+
"grad_norm": 36.1817741394043,
|
| 118 |
+
"learning_rate": 1.8162962962962963e-05,
|
| 119 |
+
"loss": 1.0933,
|
| 120 |
+
"step": 275
|
| 121 |
+
},
|
| 122 |
+
{
|
| 123 |
+
"epoch": 4.0,
|
| 124 |
+
"grad_norm": 14.19780445098877,
|
| 125 |
+
"learning_rate": 1.7792592592592596e-05,
|
| 126 |
+
"loss": 1.2817,
|
| 127 |
+
"step": 300
|
| 128 |
+
},
|
| 129 |
+
{
|
| 130 |
+
"epoch": 4.0,
|
| 131 |
+
"eval_accuracy": 0.847953216374269,
|
| 132 |
+
"eval_loss": 0.33539971709251404,
|
| 133 |
+
"eval_macro_f1": 0.8304089105889534,
|
| 134 |
+
"eval_runtime": 0.3044,
|
| 135 |
+
"eval_samples_per_second": 1123.459,
|
| 136 |
+
"eval_steps_per_second": 72.269,
|
| 137 |
+
"eval_weighted_f1": 0.8508241027755025,
|
| 138 |
+
"step": 300
|
| 139 |
+
},
|
| 140 |
+
{
|
| 141 |
+
"epoch": 4.334448160535117,
|
| 142 |
+
"grad_norm": 128.69671630859375,
|
| 143 |
+
"learning_rate": 1.7422222222222222e-05,
|
| 144 |
+
"loss": 0.7906,
|
| 145 |
+
"step": 325
|
| 146 |
+
},
|
| 147 |
+
{
|
| 148 |
+
"epoch": 4.668896321070234,
|
| 149 |
+
"grad_norm": 87.4158935546875,
|
| 150 |
+
"learning_rate": 1.7051851851851854e-05,
|
| 151 |
+
"loss": 0.8355,
|
| 152 |
+
"step": 350
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 5.0,
|
| 156 |
+
"grad_norm": 43.76517868041992,
|
| 157 |
+
"learning_rate": 1.6681481481481484e-05,
|
| 158 |
+
"loss": 0.8939,
|
| 159 |
+
"step": 375
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 5.0,
|
| 163 |
+
"eval_accuracy": 0.8508771929824561,
|
| 164 |
+
"eval_loss": 0.3892245888710022,
|
| 165 |
+
"eval_macro_f1": 0.8170835299669656,
|
| 166 |
+
"eval_runtime": 0.2966,
|
| 167 |
+
"eval_samples_per_second": 1153.242,
|
| 168 |
+
"eval_steps_per_second": 74.185,
|
| 169 |
+
"eval_weighted_f1": 0.846509304565488,
|
| 170 |
+
"step": 375
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 5.334448160535117,
|
| 174 |
+
"grad_norm": 43.355770111083984,
|
| 175 |
+
"learning_rate": 1.6311111111111113e-05,
|
| 176 |
+
"loss": 0.5254,
|
| 177 |
+
"step": 400
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 5.668896321070234,
|
| 181 |
+
"grad_norm": 64.8084487915039,
|
| 182 |
+
"learning_rate": 1.5940740740740742e-05,
|
| 183 |
+
"loss": 0.6326,
|
| 184 |
+
"step": 425
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 6.0,
|
| 188 |
+
"grad_norm": 25.248273849487305,
|
| 189 |
+
"learning_rate": 1.557037037037037e-05,
|
| 190 |
+
"loss": 0.4772,
|
| 191 |
+
"step": 450
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 6.0,
|
| 195 |
+
"eval_accuracy": 0.8596491228070176,
|
| 196 |
+
"eval_loss": 0.5135303735733032,
|
| 197 |
+
"eval_macro_f1": 0.8273090679570798,
|
| 198 |
+
"eval_runtime": 0.2867,
|
| 199 |
+
"eval_samples_per_second": 1193.074,
|
| 200 |
+
"eval_steps_per_second": 76.747,
|
| 201 |
+
"eval_weighted_f1": 0.8552788451245935,
|
| 202 |
+
"step": 450
|
| 203 |
+
},
|
| 204 |
+
{
|
| 205 |
+
"epoch": 6.334448160535117,
|
| 206 |
+
"grad_norm": 96.37144470214844,
|
| 207 |
+
"learning_rate": 1.5200000000000002e-05,
|
| 208 |
+
"loss": 0.3917,
|
| 209 |
+
"step": 475
|
| 210 |
+
},
|
| 211 |
+
{
|
| 212 |
+
"epoch": 6.668896321070234,
|
| 213 |
+
"grad_norm": 164.3301544189453,
|
| 214 |
+
"learning_rate": 1.4829629629629631e-05,
|
| 215 |
+
"loss": 0.5593,
|
| 216 |
+
"step": 500
|
| 217 |
+
},
|
| 218 |
+
{
|
| 219 |
+
"epoch": 7.0,
|
| 220 |
+
"grad_norm": 32.31126403808594,
|
| 221 |
+
"learning_rate": 1.445925925925926e-05,
|
| 222 |
+
"loss": 0.3868,
|
| 223 |
+
"step": 525
|
| 224 |
+
},
|
| 225 |
+
{
|
| 226 |
+
"epoch": 7.0,
|
| 227 |
+
"eval_accuracy": 0.8508771929824561,
|
| 228 |
+
"eval_loss": 0.6291298866271973,
|
| 229 |
+
"eval_macro_f1": 0.8135642135642136,
|
| 230 |
+
"eval_runtime": 0.2993,
|
| 231 |
+
"eval_samples_per_second": 1142.524,
|
| 232 |
+
"eval_steps_per_second": 73.496,
|
| 233 |
+
"eval_weighted_f1": 0.8447803009206518,
|
| 234 |
+
"step": 525
|
| 235 |
+
}
|
| 236 |
+
],
|
| 237 |
+
"logging_steps": 25,
|
| 238 |
+
"max_steps": 1500,
|
| 239 |
+
"num_input_tokens_seen": 0,
|
| 240 |
+
"num_train_epochs": 20,
|
| 241 |
+
"save_steps": 500,
|
| 242 |
+
"stateful_callbacks": {
|
| 243 |
+
"EarlyStoppingCallback": {
|
| 244 |
+
"args": {
|
| 245 |
+
"early_stopping_patience": 3,
|
| 246 |
+
"early_stopping_threshold": 0.0
|
| 247 |
+
},
|
| 248 |
+
"attributes": {
|
| 249 |
+
"early_stopping_patience_counter": 3
|
| 250 |
+
}
|
| 251 |
+
},
|
| 252 |
+
"TrainerControl": {
|
| 253 |
+
"args": {
|
| 254 |
+
"should_epoch_stop": false,
|
| 255 |
+
"should_evaluate": false,
|
| 256 |
+
"should_log": false,
|
| 257 |
+
"should_save": true,
|
| 258 |
+
"should_training_stop": true
|
| 259 |
+
},
|
| 260 |
+
"attributes": {}
|
| 261 |
+
}
|
| 262 |
+
},
|
| 263 |
+
"total_flos": 0.0,
|
| 264 |
+
"train_batch_size": 8,
|
| 265 |
+
"trial_name": null,
|
| 266 |
+
"trial_params": null
|
| 267 |
+
}
|
vifactcheck-bamibert-s42-20260827T131409Z-6235c159/trainer/checkpoint-1141/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 |
+
}
|
vifactcheck-bamibert-s42-20260827T131409Z-6235c159/trainer/checkpoint-1141/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
vifactcheck-bamibert-s42-20260827T131409Z-6235c159/trainer/checkpoint-1141/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 |
+
}
|
vifactcheck-bamibert-s42-20260827T131409Z-6235c159/trainer/checkpoint-1141/trainer_state.json
ADDED
|
@@ -0,0 +1,435 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 652,
|
| 3 |
+
"best_metric": 0.83379336394591,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/vifactcheck-bamibert-s42-20260827T131409Z-6235c159/trainer/checkpoint-652",
|
| 5 |
+
"epoch": 7.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1141,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.15408320493066255,
|
| 14 |
+
"grad_norm": 7.6812849044799805,
|
| 15 |
+
"learning_rate": 1.47239263803681e-06,
|
| 16 |
+
"loss": 4.4864,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.3081664098613251,
|
| 21 |
+
"grad_norm": 37.47366714477539,
|
| 22 |
+
"learning_rate": 3.0061349693251535e-06,
|
| 23 |
+
"loss": 4.4432,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.4622496147919877,
|
| 28 |
+
"grad_norm": 12.01583480834961,
|
| 29 |
+
"learning_rate": 4.539877300613497e-06,
|
| 30 |
+
"loss": 4.436,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.6163328197226502,
|
| 35 |
+
"grad_norm": 47.4721794128418,
|
| 36 |
+
"learning_rate": 6.073619631901841e-06,
|
| 37 |
+
"loss": 4.3803,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.7704160246533128,
|
| 42 |
+
"grad_norm": 20.56624984741211,
|
| 43 |
+
"learning_rate": 7.6073619631901856e-06,
|
| 44 |
+
"loss": 4.4072,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.9244992295839753,
|
| 49 |
+
"grad_norm": 10.15404987335205,
|
| 50 |
+
"learning_rate": 9.14110429447853e-06,
|
| 51 |
+
"loss": 4.3267,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 1.0,
|
| 56 |
+
"eval_accuracy": 0.4316860465116279,
|
| 57 |
+
"eval_loss": 1.0602402687072754,
|
| 58 |
+
"eval_macro_f1": 0.3536565001641662,
|
| 59 |
+
"eval_runtime": 0.7619,
|
| 60 |
+
"eval_samples_per_second": 902.969,
|
| 61 |
+
"eval_steps_per_second": 56.436,
|
| 62 |
+
"eval_weighted_f1": 0.3562171317608899,
|
| 63 |
+
"step": 163
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 1.073959938366718,
|
| 67 |
+
"grad_norm": 13.90970516204834,
|
| 68 |
+
"learning_rate": 1.0674846625766873e-05,
|
| 69 |
+
"loss": 4.1438,
|
| 70 |
+
"step": 175
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.2280431432973806,
|
| 74 |
+
"grad_norm": 22.991500854492188,
|
| 75 |
+
"learning_rate": 1.2208588957055216e-05,
|
| 76 |
+
"loss": 4.1846,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.3821263482280433,
|
| 81 |
+
"grad_norm": 29.1098690032959,
|
| 82 |
+
"learning_rate": 1.374233128834356e-05,
|
| 83 |
+
"loss": 3.8417,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.5362095531587057,
|
| 88 |
+
"grad_norm": 51.852054595947266,
|
| 89 |
+
"learning_rate": 1.5276073619631903e-05,
|
| 90 |
+
"loss": 3.3236,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.6902927580893683,
|
| 95 |
+
"grad_norm": 34.34718322753906,
|
| 96 |
+
"learning_rate": 1.6809815950920248e-05,
|
| 97 |
+
"loss": 3.2259,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.8443759630200307,
|
| 102 |
+
"grad_norm": 37.232662200927734,
|
| 103 |
+
"learning_rate": 1.834355828220859e-05,
|
| 104 |
+
"loss": 2.9019,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.9984591679506933,
|
| 109 |
+
"grad_norm": 30.793975830078125,
|
| 110 |
+
"learning_rate": 1.9877300613496935e-05,
|
| 111 |
+
"loss": 2.5158,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 2.0,
|
| 116 |
+
"eval_accuracy": 0.7761627906976745,
|
| 117 |
+
"eval_loss": 0.5281609892845154,
|
| 118 |
+
"eval_macro_f1": 0.7783366074200392,
|
| 119 |
+
"eval_runtime": 0.7763,
|
| 120 |
+
"eval_samples_per_second": 886.26,
|
| 121 |
+
"eval_steps_per_second": 55.391,
|
| 122 |
+
"eval_weighted_f1": 0.7785048894944147,
|
| 123 |
+
"step": 326
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 2.147919876733436,
|
| 127 |
+
"grad_norm": 47.142337799072266,
|
| 128 |
+
"learning_rate": 1.9843217450579417e-05,
|
| 129 |
+
"loss": 2.3995,
|
| 130 |
+
"step": 350
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.3020030816640986,
|
| 134 |
+
"grad_norm": 66.7985610961914,
|
| 135 |
+
"learning_rate": 1.967280163599182e-05,
|
| 136 |
+
"loss": 2.1824,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.4560862865947612,
|
| 141 |
+
"grad_norm": 60.10746765136719,
|
| 142 |
+
"learning_rate": 1.9502385821404228e-05,
|
| 143 |
+
"loss": 2.2218,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.610169491525424,
|
| 148 |
+
"grad_norm": 32.13690185546875,
|
| 149 |
+
"learning_rate": 1.9331970006816635e-05,
|
| 150 |
+
"loss": 2.1596,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.7642526964560865,
|
| 155 |
+
"grad_norm": 54.7959098815918,
|
| 156 |
+
"learning_rate": 1.916155419222904e-05,
|
| 157 |
+
"loss": 1.9909,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.9183359013867487,
|
| 162 |
+
"grad_norm": 55.35082244873047,
|
| 163 |
+
"learning_rate": 1.8991138377641446e-05,
|
| 164 |
+
"loss": 1.9818,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 3.0,
|
| 169 |
+
"eval_accuracy": 0.8052325581395349,
|
| 170 |
+
"eval_loss": 0.4992329776287079,
|
| 171 |
+
"eval_macro_f1": 0.8031481088868264,
|
| 172 |
+
"eval_runtime": 0.7601,
|
| 173 |
+
"eval_samples_per_second": 905.13,
|
| 174 |
+
"eval_steps_per_second": 56.571,
|
| 175 |
+
"eval_weighted_f1": 0.8035906677408563,
|
| 176 |
+
"step": 489
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 3.0677966101694913,
|
| 180 |
+
"grad_norm": 49.5398063659668,
|
| 181 |
+
"learning_rate": 1.8820722563053853e-05,
|
| 182 |
+
"loss": 1.723,
|
| 183 |
+
"step": 500
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 3.221879815100154,
|
| 187 |
+
"grad_norm": 34.983524322509766,
|
| 188 |
+
"learning_rate": 1.8650306748466257e-05,
|
| 189 |
+
"loss": 1.5617,
|
| 190 |
+
"step": 525
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.3759630200308166,
|
| 194 |
+
"grad_norm": 47.64194107055664,
|
| 195 |
+
"learning_rate": 1.8479890933878664e-05,
|
| 196 |
+
"loss": 1.4065,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.5300462249614792,
|
| 201 |
+
"grad_norm": 55.22037887573242,
|
| 202 |
+
"learning_rate": 1.8309475119291072e-05,
|
| 203 |
+
"loss": 1.4224,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.684129429892142,
|
| 208 |
+
"grad_norm": 50.603755950927734,
|
| 209 |
+
"learning_rate": 1.8139059304703476e-05,
|
| 210 |
+
"loss": 1.2816,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.8382126348228045,
|
| 215 |
+
"grad_norm": 60.967857360839844,
|
| 216 |
+
"learning_rate": 1.7968643490115886e-05,
|
| 217 |
+
"loss": 1.4546,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.992295839753467,
|
| 222 |
+
"grad_norm": 45.44284439086914,
|
| 223 |
+
"learning_rate": 1.779822767552829e-05,
|
| 224 |
+
"loss": 1.4805,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 4.0,
|
| 229 |
+
"eval_accuracy": 0.8343023255813954,
|
| 230 |
+
"eval_loss": 0.5035151839256287,
|
| 231 |
+
"eval_macro_f1": 0.83379336394591,
|
| 232 |
+
"eval_runtime": 0.7561,
|
| 233 |
+
"eval_samples_per_second": 909.982,
|
| 234 |
+
"eval_steps_per_second": 56.874,
|
| 235 |
+
"eval_weighted_f1": 0.8336651635011868,
|
| 236 |
+
"step": 652
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 4.141756548536209,
|
| 240 |
+
"grad_norm": 52.671409606933594,
|
| 241 |
+
"learning_rate": 1.7627811860940697e-05,
|
| 242 |
+
"loss": 0.9595,
|
| 243 |
+
"step": 675
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 4.295839753466872,
|
| 247 |
+
"grad_norm": 65.08079528808594,
|
| 248 |
+
"learning_rate": 1.7457396046353105e-05,
|
| 249 |
+
"loss": 1.0172,
|
| 250 |
+
"step": 700
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 4.449922958397535,
|
| 254 |
+
"grad_norm": 39.032405853271484,
|
| 255 |
+
"learning_rate": 1.728698023176551e-05,
|
| 256 |
+
"loss": 0.9625,
|
| 257 |
+
"step": 725
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.604006163328197,
|
| 261 |
+
"grad_norm": 49.76782989501953,
|
| 262 |
+
"learning_rate": 1.7116564417177916e-05,
|
| 263 |
+
"loss": 1.0306,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.75808936825886,
|
| 268 |
+
"grad_norm": 36.80021286010742,
|
| 269 |
+
"learning_rate": 1.6946148602590323e-05,
|
| 270 |
+
"loss": 1.0969,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.9121725731895225,
|
| 275 |
+
"grad_norm": 55.91025161743164,
|
| 276 |
+
"learning_rate": 1.6775732788002727e-05,
|
| 277 |
+
"loss": 0.918,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 5.0,
|
| 282 |
+
"eval_accuracy": 0.8023255813953488,
|
| 283 |
+
"eval_loss": 0.713641345500946,
|
| 284 |
+
"eval_macro_f1": 0.7963246104243034,
|
| 285 |
+
"eval_runtime": 0.7613,
|
| 286 |
+
"eval_samples_per_second": 903.695,
|
| 287 |
+
"eval_steps_per_second": 56.481,
|
| 288 |
+
"eval_weighted_f1": 0.7969736476380451,
|
| 289 |
+
"step": 815
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 5.061633281972265,
|
| 293 |
+
"grad_norm": 9.434857368469238,
|
| 294 |
+
"learning_rate": 1.6605316973415134e-05,
|
| 295 |
+
"loss": 0.9709,
|
| 296 |
+
"step": 825
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 5.215716486902927,
|
| 300 |
+
"grad_norm": 79.32661437988281,
|
| 301 |
+
"learning_rate": 1.643490115882754e-05,
|
| 302 |
+
"loss": 0.6909,
|
| 303 |
+
"step": 850
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 5.36979969183359,
|
| 307 |
+
"grad_norm": 55.69626235961914,
|
| 308 |
+
"learning_rate": 1.6264485344239945e-05,
|
| 309 |
+
"loss": 0.6665,
|
| 310 |
+
"step": 875
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 5.523882896764253,
|
| 314 |
+
"grad_norm": 71.31925964355469,
|
| 315 |
+
"learning_rate": 1.6094069529652356e-05,
|
| 316 |
+
"loss": 0.6429,
|
| 317 |
+
"step": 900
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.677966101694915,
|
| 321 |
+
"grad_norm": 22.821557998657227,
|
| 322 |
+
"learning_rate": 1.592365371506476e-05,
|
| 323 |
+
"loss": 0.8487,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.832049306625578,
|
| 328 |
+
"grad_norm": 29.089717864990234,
|
| 329 |
+
"learning_rate": 1.5753237900477167e-05,
|
| 330 |
+
"loss": 0.7974,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.9861325115562405,
|
| 335 |
+
"grad_norm": 56.60846710205078,
|
| 336 |
+
"learning_rate": 1.5582822085889574e-05,
|
| 337 |
+
"loss": 0.622,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 6.0,
|
| 342 |
+
"eval_accuracy": 0.8255813953488372,
|
| 343 |
+
"eval_loss": 0.5727497935295105,
|
| 344 |
+
"eval_macro_f1": 0.8242133526148772,
|
| 345 |
+
"eval_runtime": 0.8022,
|
| 346 |
+
"eval_samples_per_second": 857.589,
|
| 347 |
+
"eval_steps_per_second": 53.599,
|
| 348 |
+
"eval_weighted_f1": 0.8242694587048188,
|
| 349 |
+
"step": 978
|
| 350 |
+
},
|
| 351 |
+
{
|
| 352 |
+
"epoch": 6.135593220338983,
|
| 353 |
+
"grad_norm": 21.597339630126953,
|
| 354 |
+
"learning_rate": 1.5412406271301978e-05,
|
| 355 |
+
"loss": 0.484,
|
| 356 |
+
"step": 1000
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 6.289676425269645,
|
| 360 |
+
"grad_norm": 29.825008392333984,
|
| 361 |
+
"learning_rate": 1.5241990456714383e-05,
|
| 362 |
+
"loss": 0.5352,
|
| 363 |
+
"step": 1025
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 6.443759630200308,
|
| 367 |
+
"grad_norm": 140.84375,
|
| 368 |
+
"learning_rate": 1.5071574642126792e-05,
|
| 369 |
+
"loss": 0.4959,
|
| 370 |
+
"step": 1050
|
| 371 |
+
},
|
| 372 |
+
{
|
| 373 |
+
"epoch": 6.597842835130971,
|
| 374 |
+
"grad_norm": 8.233342170715332,
|
| 375 |
+
"learning_rate": 1.4901158827539198e-05,
|
| 376 |
+
"loss": 0.601,
|
| 377 |
+
"step": 1075
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 6.751926040061633,
|
| 381 |
+
"grad_norm": 12.62174129486084,
|
| 382 |
+
"learning_rate": 1.4730743012951603e-05,
|
| 383 |
+
"loss": 0.5952,
|
| 384 |
+
"step": 1100
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 6.906009244992296,
|
| 388 |
+
"grad_norm": 66.21369934082031,
|
| 389 |
+
"learning_rate": 1.456032719836401e-05,
|
| 390 |
+
"loss": 0.4856,
|
| 391 |
+
"step": 1125
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 7.0,
|
| 395 |
+
"eval_accuracy": 0.8241279069767442,
|
| 396 |
+
"eval_loss": 0.7367490530014038,
|
| 397 |
+
"eval_macro_f1": 0.8227765422268595,
|
| 398 |
+
"eval_runtime": 0.7644,
|
| 399 |
+
"eval_samples_per_second": 900.0,
|
| 400 |
+
"eval_steps_per_second": 56.25,
|
| 401 |
+
"eval_weighted_f1": 0.8231725180272306,
|
| 402 |
+
"step": 1141
|
| 403 |
+
}
|
| 404 |
+
],
|
| 405 |
+
"logging_steps": 25,
|
| 406 |
+
"max_steps": 3260,
|
| 407 |
+
"num_input_tokens_seen": 0,
|
| 408 |
+
"num_train_epochs": 20,
|
| 409 |
+
"save_steps": 500,
|
| 410 |
+
"stateful_callbacks": {
|
| 411 |
+
"EarlyStoppingCallback": {
|
| 412 |
+
"args": {
|
| 413 |
+
"early_stopping_patience": 3,
|
| 414 |
+
"early_stopping_threshold": 0.0
|
| 415 |
+
},
|
| 416 |
+
"attributes": {
|
| 417 |
+
"early_stopping_patience_counter": 3
|
| 418 |
+
}
|
| 419 |
+
},
|
| 420 |
+
"TrainerControl": {
|
| 421 |
+
"args": {
|
| 422 |
+
"should_epoch_stop": false,
|
| 423 |
+
"should_evaluate": false,
|
| 424 |
+
"should_log": false,
|
| 425 |
+
"should_save": true,
|
| 426 |
+
"should_training_stop": true
|
| 427 |
+
},
|
| 428 |
+
"attributes": {}
|
| 429 |
+
}
|
| 430 |
+
},
|
| 431 |
+
"total_flos": 0.0,
|
| 432 |
+
"train_batch_size": 8,
|
| 433 |
+
"trial_name": null,
|
| 434 |
+
"trial_params": null
|
| 435 |
+
}
|
vifactcheck-bamibert-s42-20260827T131409Z-6235c159/trainer/checkpoint-652/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 |
+
}
|
vifactcheck-bamibert-s42-20260827T131409Z-6235c159/trainer/checkpoint-652/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
vifactcheck-bamibert-s42-20260827T131409Z-6235c159/trainer/checkpoint-652/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 |
+
}
|
vifactcheck-bamibert-s42-20260827T131409Z-6235c159/trainer/checkpoint-652/trainer_state.json
ADDED
|
@@ -0,0 +1,269 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 652,
|
| 3 |
+
"best_metric": 0.83379336394591,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/vifactcheck-bamibert-s42-20260827T131409Z-6235c159/trainer/checkpoint-652",
|
| 5 |
+
"epoch": 4.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 652,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.15408320493066255,
|
| 14 |
+
"grad_norm": 7.6812849044799805,
|
| 15 |
+
"learning_rate": 1.47239263803681e-06,
|
| 16 |
+
"loss": 4.4864,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.3081664098613251,
|
| 21 |
+
"grad_norm": 37.47366714477539,
|
| 22 |
+
"learning_rate": 3.0061349693251535e-06,
|
| 23 |
+
"loss": 4.4432,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.4622496147919877,
|
| 28 |
+
"grad_norm": 12.01583480834961,
|
| 29 |
+
"learning_rate": 4.539877300613497e-06,
|
| 30 |
+
"loss": 4.436,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.6163328197226502,
|
| 35 |
+
"grad_norm": 47.4721794128418,
|
| 36 |
+
"learning_rate": 6.073619631901841e-06,
|
| 37 |
+
"loss": 4.3803,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.7704160246533128,
|
| 42 |
+
"grad_norm": 20.56624984741211,
|
| 43 |
+
"learning_rate": 7.6073619631901856e-06,
|
| 44 |
+
"loss": 4.4072,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.9244992295839753,
|
| 49 |
+
"grad_norm": 10.15404987335205,
|
| 50 |
+
"learning_rate": 9.14110429447853e-06,
|
| 51 |
+
"loss": 4.3267,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 1.0,
|
| 56 |
+
"eval_accuracy": 0.4316860465116279,
|
| 57 |
+
"eval_loss": 1.0602402687072754,
|
| 58 |
+
"eval_macro_f1": 0.3536565001641662,
|
| 59 |
+
"eval_runtime": 0.7619,
|
| 60 |
+
"eval_samples_per_second": 902.969,
|
| 61 |
+
"eval_steps_per_second": 56.436,
|
| 62 |
+
"eval_weighted_f1": 0.3562171317608899,
|
| 63 |
+
"step": 163
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 1.073959938366718,
|
| 67 |
+
"grad_norm": 13.90970516204834,
|
| 68 |
+
"learning_rate": 1.0674846625766873e-05,
|
| 69 |
+
"loss": 4.1438,
|
| 70 |
+
"step": 175
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.2280431432973806,
|
| 74 |
+
"grad_norm": 22.991500854492188,
|
| 75 |
+
"learning_rate": 1.2208588957055216e-05,
|
| 76 |
+
"loss": 4.1846,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.3821263482280433,
|
| 81 |
+
"grad_norm": 29.1098690032959,
|
| 82 |
+
"learning_rate": 1.374233128834356e-05,
|
| 83 |
+
"loss": 3.8417,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.5362095531587057,
|
| 88 |
+
"grad_norm": 51.852054595947266,
|
| 89 |
+
"learning_rate": 1.5276073619631903e-05,
|
| 90 |
+
"loss": 3.3236,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.6902927580893683,
|
| 95 |
+
"grad_norm": 34.34718322753906,
|
| 96 |
+
"learning_rate": 1.6809815950920248e-05,
|
| 97 |
+
"loss": 3.2259,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.8443759630200307,
|
| 102 |
+
"grad_norm": 37.232662200927734,
|
| 103 |
+
"learning_rate": 1.834355828220859e-05,
|
| 104 |
+
"loss": 2.9019,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.9984591679506933,
|
| 109 |
+
"grad_norm": 30.793975830078125,
|
| 110 |
+
"learning_rate": 1.9877300613496935e-05,
|
| 111 |
+
"loss": 2.5158,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 2.0,
|
| 116 |
+
"eval_accuracy": 0.7761627906976745,
|
| 117 |
+
"eval_loss": 0.5281609892845154,
|
| 118 |
+
"eval_macro_f1": 0.7783366074200392,
|
| 119 |
+
"eval_runtime": 0.7763,
|
| 120 |
+
"eval_samples_per_second": 886.26,
|
| 121 |
+
"eval_steps_per_second": 55.391,
|
| 122 |
+
"eval_weighted_f1": 0.7785048894944147,
|
| 123 |
+
"step": 326
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 2.147919876733436,
|
| 127 |
+
"grad_norm": 47.142337799072266,
|
| 128 |
+
"learning_rate": 1.9843217450579417e-05,
|
| 129 |
+
"loss": 2.3995,
|
| 130 |
+
"step": 350
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.3020030816640986,
|
| 134 |
+
"grad_norm": 66.7985610961914,
|
| 135 |
+
"learning_rate": 1.967280163599182e-05,
|
| 136 |
+
"loss": 2.1824,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.4560862865947612,
|
| 141 |
+
"grad_norm": 60.10746765136719,
|
| 142 |
+
"learning_rate": 1.9502385821404228e-05,
|
| 143 |
+
"loss": 2.2218,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.610169491525424,
|
| 148 |
+
"grad_norm": 32.13690185546875,
|
| 149 |
+
"learning_rate": 1.9331970006816635e-05,
|
| 150 |
+
"loss": 2.1596,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.7642526964560865,
|
| 155 |
+
"grad_norm": 54.7959098815918,
|
| 156 |
+
"learning_rate": 1.916155419222904e-05,
|
| 157 |
+
"loss": 1.9909,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.9183359013867487,
|
| 162 |
+
"grad_norm": 55.35082244873047,
|
| 163 |
+
"learning_rate": 1.8991138377641446e-05,
|
| 164 |
+
"loss": 1.9818,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 3.0,
|
| 169 |
+
"eval_accuracy": 0.8052325581395349,
|
| 170 |
+
"eval_loss": 0.4992329776287079,
|
| 171 |
+
"eval_macro_f1": 0.8031481088868264,
|
| 172 |
+
"eval_runtime": 0.7601,
|
| 173 |
+
"eval_samples_per_second": 905.13,
|
| 174 |
+
"eval_steps_per_second": 56.571,
|
| 175 |
+
"eval_weighted_f1": 0.8035906677408563,
|
| 176 |
+
"step": 489
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 3.0677966101694913,
|
| 180 |
+
"grad_norm": 49.5398063659668,
|
| 181 |
+
"learning_rate": 1.8820722563053853e-05,
|
| 182 |
+
"loss": 1.723,
|
| 183 |
+
"step": 500
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 3.221879815100154,
|
| 187 |
+
"grad_norm": 34.983524322509766,
|
| 188 |
+
"learning_rate": 1.8650306748466257e-05,
|
| 189 |
+
"loss": 1.5617,
|
| 190 |
+
"step": 525
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.3759630200308166,
|
| 194 |
+
"grad_norm": 47.64194107055664,
|
| 195 |
+
"learning_rate": 1.8479890933878664e-05,
|
| 196 |
+
"loss": 1.4065,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.5300462249614792,
|
| 201 |
+
"grad_norm": 55.22037887573242,
|
| 202 |
+
"learning_rate": 1.8309475119291072e-05,
|
| 203 |
+
"loss": 1.4224,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.684129429892142,
|
| 208 |
+
"grad_norm": 50.603755950927734,
|
| 209 |
+
"learning_rate": 1.8139059304703476e-05,
|
| 210 |
+
"loss": 1.2816,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.8382126348228045,
|
| 215 |
+
"grad_norm": 60.967857360839844,
|
| 216 |
+
"learning_rate": 1.7968643490115886e-05,
|
| 217 |
+
"loss": 1.4546,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.992295839753467,
|
| 222 |
+
"grad_norm": 45.44284439086914,
|
| 223 |
+
"learning_rate": 1.779822767552829e-05,
|
| 224 |
+
"loss": 1.4805,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 4.0,
|
| 229 |
+
"eval_accuracy": 0.8343023255813954,
|
| 230 |
+
"eval_loss": 0.5035151839256287,
|
| 231 |
+
"eval_macro_f1": 0.83379336394591,
|
| 232 |
+
"eval_runtime": 0.7561,
|
| 233 |
+
"eval_samples_per_second": 909.982,
|
| 234 |
+
"eval_steps_per_second": 56.874,
|
| 235 |
+
"eval_weighted_f1": 0.8336651635011868,
|
| 236 |
+
"step": 652
|
| 237 |
+
}
|
| 238 |
+
],
|
| 239 |
+
"logging_steps": 25,
|
| 240 |
+
"max_steps": 3260,
|
| 241 |
+
"num_input_tokens_seen": 0,
|
| 242 |
+
"num_train_epochs": 20,
|
| 243 |
+
"save_steps": 500,
|
| 244 |
+
"stateful_callbacks": {
|
| 245 |
+
"EarlyStoppingCallback": {
|
| 246 |
+
"args": {
|
| 247 |
+
"early_stopping_patience": 3,
|
| 248 |
+
"early_stopping_threshold": 0.0
|
| 249 |
+
},
|
| 250 |
+
"attributes": {
|
| 251 |
+
"early_stopping_patience_counter": 0
|
| 252 |
+
}
|
| 253 |
+
},
|
| 254 |
+
"TrainerControl": {
|
| 255 |
+
"args": {
|
| 256 |
+
"should_epoch_stop": false,
|
| 257 |
+
"should_evaluate": false,
|
| 258 |
+
"should_log": false,
|
| 259 |
+
"should_save": true,
|
| 260 |
+
"should_training_stop": false
|
| 261 |
+
},
|
| 262 |
+
"attributes": {}
|
| 263 |
+
}
|
| 264 |
+
},
|
| 265 |
+
"total_flos": 0.0,
|
| 266 |
+
"train_batch_size": 8,
|
| 267 |
+
"trial_name": null,
|
| 268 |
+
"trial_params": null
|
| 269 |
+
}
|
vifactcheck-bamibert-s43-20260827T131648Z-3858bb98/trainer/checkpoint-1304/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 |
+
}
|
vifactcheck-bamibert-s43-20260827T131648Z-3858bb98/trainer/checkpoint-1304/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
vifactcheck-bamibert-s43-20260827T131648Z-3858bb98/trainer/checkpoint-1304/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 |
+
}
|
vifactcheck-bamibert-s43-20260827T131648Z-3858bb98/trainer/checkpoint-1304/trainer_state.json
ADDED
|
@@ -0,0 +1,495 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 815,
|
| 3 |
+
"best_metric": 0.8539482123960574,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/vifactcheck-bamibert-s43-20260827T131648Z-3858bb98/trainer/checkpoint-815",
|
| 5 |
+
"epoch": 8.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1304,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.15408320493066255,
|
| 14 |
+
"grad_norm": 8.965160369873047,
|
| 15 |
+
"learning_rate": 1.47239263803681e-06,
|
| 16 |
+
"loss": 4.4139,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.3081664098613251,
|
| 21 |
+
"grad_norm": 13.331682205200195,
|
| 22 |
+
"learning_rate": 3.0061349693251535e-06,
|
| 23 |
+
"loss": 4.4184,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.4622496147919877,
|
| 28 |
+
"grad_norm": 103.15205383300781,
|
| 29 |
+
"learning_rate": 4.539877300613497e-06,
|
| 30 |
+
"loss": 4.4124,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.6163328197226502,
|
| 35 |
+
"grad_norm": 19.21099090576172,
|
| 36 |
+
"learning_rate": 6.073619631901841e-06,
|
| 37 |
+
"loss": 4.3879,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.7704160246533128,
|
| 42 |
+
"grad_norm": 23.29578971862793,
|
| 43 |
+
"learning_rate": 7.6073619631901856e-06,
|
| 44 |
+
"loss": 4.3548,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.9244992295839753,
|
| 49 |
+
"grad_norm": 20.717185974121094,
|
| 50 |
+
"learning_rate": 9.14110429447853e-06,
|
| 51 |
+
"loss": 4.3273,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 1.0,
|
| 56 |
+
"eval_accuracy": 0.4418604651162791,
|
| 57 |
+
"eval_loss": 1.0659151077270508,
|
| 58 |
+
"eval_macro_f1": 0.35462019711419224,
|
| 59 |
+
"eval_runtime": 0.7815,
|
| 60 |
+
"eval_samples_per_second": 880.332,
|
| 61 |
+
"eval_steps_per_second": 55.021,
|
| 62 |
+
"eval_weighted_f1": 0.35778381033049417,
|
| 63 |
+
"step": 163
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 1.073959938366718,
|
| 67 |
+
"grad_norm": 28.10257339477539,
|
| 68 |
+
"learning_rate": 1.0674846625766873e-05,
|
| 69 |
+
"loss": 4.159,
|
| 70 |
+
"step": 175
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.2280431432973806,
|
| 74 |
+
"grad_norm": 20.976919174194336,
|
| 75 |
+
"learning_rate": 1.2208588957055216e-05,
|
| 76 |
+
"loss": 4.1987,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.3821263482280433,
|
| 81 |
+
"grad_norm": 18.764131546020508,
|
| 82 |
+
"learning_rate": 1.374233128834356e-05,
|
| 83 |
+
"loss": 4.1076,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.5362095531587057,
|
| 88 |
+
"grad_norm": 85.98760223388672,
|
| 89 |
+
"learning_rate": 1.5276073619631903e-05,
|
| 90 |
+
"loss": 3.5784,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.6902927580893683,
|
| 95 |
+
"grad_norm": 70.68370819091797,
|
| 96 |
+
"learning_rate": 1.6809815950920248e-05,
|
| 97 |
+
"loss": 3.3158,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.8443759630200307,
|
| 102 |
+
"grad_norm": 87.6795654296875,
|
| 103 |
+
"learning_rate": 1.834355828220859e-05,
|
| 104 |
+
"loss": 3.0421,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.9984591679506933,
|
| 109 |
+
"grad_norm": 38.58009719848633,
|
| 110 |
+
"learning_rate": 1.9877300613496935e-05,
|
| 111 |
+
"loss": 2.8683,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 2.0,
|
| 116 |
+
"eval_accuracy": 0.7412790697674418,
|
| 117 |
+
"eval_loss": 0.6256099939346313,
|
| 118 |
+
"eval_macro_f1": 0.7366093723446947,
|
| 119 |
+
"eval_runtime": 0.7796,
|
| 120 |
+
"eval_samples_per_second": 882.459,
|
| 121 |
+
"eval_steps_per_second": 55.154,
|
| 122 |
+
"eval_weighted_f1": 0.7371250676827075,
|
| 123 |
+
"step": 326
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 2.147919876733436,
|
| 127 |
+
"grad_norm": 41.14516830444336,
|
| 128 |
+
"learning_rate": 1.9843217450579417e-05,
|
| 129 |
+
"loss": 2.3686,
|
| 130 |
+
"step": 350
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.3020030816640986,
|
| 134 |
+
"grad_norm": 32.573055267333984,
|
| 135 |
+
"learning_rate": 1.967280163599182e-05,
|
| 136 |
+
"loss": 2.3641,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.4560862865947612,
|
| 141 |
+
"grad_norm": 48.540042877197266,
|
| 142 |
+
"learning_rate": 1.9502385821404228e-05,
|
| 143 |
+
"loss": 2.2222,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.610169491525424,
|
| 148 |
+
"grad_norm": 65.02373504638672,
|
| 149 |
+
"learning_rate": 1.9331970006816635e-05,
|
| 150 |
+
"loss": 1.9679,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.7642526964560865,
|
| 155 |
+
"grad_norm": 65.37238311767578,
|
| 156 |
+
"learning_rate": 1.916155419222904e-05,
|
| 157 |
+
"loss": 2.2375,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.9183359013867487,
|
| 162 |
+
"grad_norm": 48.379146575927734,
|
| 163 |
+
"learning_rate": 1.8991138377641446e-05,
|
| 164 |
+
"loss": 2.1481,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 3.0,
|
| 169 |
+
"eval_accuracy": 0.7819767441860465,
|
| 170 |
+
"eval_loss": 0.5680427551269531,
|
| 171 |
+
"eval_macro_f1": 0.7805278799026678,
|
| 172 |
+
"eval_runtime": 0.7704,
|
| 173 |
+
"eval_samples_per_second": 893.029,
|
| 174 |
+
"eval_steps_per_second": 55.814,
|
| 175 |
+
"eval_weighted_f1": 0.78056762599925,
|
| 176 |
+
"step": 489
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 3.0677966101694913,
|
| 180 |
+
"grad_norm": 41.009098052978516,
|
| 181 |
+
"learning_rate": 1.8820722563053853e-05,
|
| 182 |
+
"loss": 1.9255,
|
| 183 |
+
"step": 500
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 3.221879815100154,
|
| 187 |
+
"grad_norm": 48.91205596923828,
|
| 188 |
+
"learning_rate": 1.8650306748466257e-05,
|
| 189 |
+
"loss": 1.6301,
|
| 190 |
+
"step": 525
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.3759630200308166,
|
| 194 |
+
"grad_norm": 74.8825912475586,
|
| 195 |
+
"learning_rate": 1.8479890933878664e-05,
|
| 196 |
+
"loss": 1.5129,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.5300462249614792,
|
| 201 |
+
"grad_norm": 76.0686264038086,
|
| 202 |
+
"learning_rate": 1.8309475119291072e-05,
|
| 203 |
+
"loss": 1.4119,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.684129429892142,
|
| 208 |
+
"grad_norm": 43.675785064697266,
|
| 209 |
+
"learning_rate": 1.8139059304703476e-05,
|
| 210 |
+
"loss": 1.3361,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.8382126348228045,
|
| 215 |
+
"grad_norm": 52.829227447509766,
|
| 216 |
+
"learning_rate": 1.7968643490115886e-05,
|
| 217 |
+
"loss": 1.5664,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.992295839753467,
|
| 222 |
+
"grad_norm": 38.033504486083984,
|
| 223 |
+
"learning_rate": 1.779822767552829e-05,
|
| 224 |
+
"loss": 1.5869,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 4.0,
|
| 229 |
+
"eval_accuracy": 0.8473837209302325,
|
| 230 |
+
"eval_loss": 0.4190322458744049,
|
| 231 |
+
"eval_macro_f1": 0.8467894155655515,
|
| 232 |
+
"eval_runtime": 0.774,
|
| 233 |
+
"eval_samples_per_second": 888.888,
|
| 234 |
+
"eval_steps_per_second": 55.555,
|
| 235 |
+
"eval_weighted_f1": 0.846585051964134,
|
| 236 |
+
"step": 652
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 4.141756548536209,
|
| 240 |
+
"grad_norm": 70.64534759521484,
|
| 241 |
+
"learning_rate": 1.7627811860940697e-05,
|
| 242 |
+
"loss": 1.0347,
|
| 243 |
+
"step": 675
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 4.295839753466872,
|
| 247 |
+
"grad_norm": 28.041074752807617,
|
| 248 |
+
"learning_rate": 1.7457396046353105e-05,
|
| 249 |
+
"loss": 0.8658,
|
| 250 |
+
"step": 700
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 4.449922958397535,
|
| 254 |
+
"grad_norm": 59.811668395996094,
|
| 255 |
+
"learning_rate": 1.728698023176551e-05,
|
| 256 |
+
"loss": 0.8912,
|
| 257 |
+
"step": 725
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.604006163328197,
|
| 261 |
+
"grad_norm": 70.92366027832031,
|
| 262 |
+
"learning_rate": 1.7116564417177916e-05,
|
| 263 |
+
"loss": 1.181,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.75808936825886,
|
| 268 |
+
"grad_norm": 35.35856246948242,
|
| 269 |
+
"learning_rate": 1.6946148602590323e-05,
|
| 270 |
+
"loss": 1.0643,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.9121725731895225,
|
| 275 |
+
"grad_norm": 90.60171508789062,
|
| 276 |
+
"learning_rate": 1.6775732788002727e-05,
|
| 277 |
+
"loss": 1.0013,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 5.0,
|
| 282 |
+
"eval_accuracy": 0.8531976744186046,
|
| 283 |
+
"eval_loss": 0.4695362150669098,
|
| 284 |
+
"eval_macro_f1": 0.8539482123960574,
|
| 285 |
+
"eval_runtime": 0.7637,
|
| 286 |
+
"eval_samples_per_second": 900.928,
|
| 287 |
+
"eval_steps_per_second": 56.308,
|
| 288 |
+
"eval_weighted_f1": 0.8538734490748382,
|
| 289 |
+
"step": 815
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 5.061633281972265,
|
| 293 |
+
"grad_norm": 44.58705139160156,
|
| 294 |
+
"learning_rate": 1.6605316973415134e-05,
|
| 295 |
+
"loss": 0.8792,
|
| 296 |
+
"step": 825
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 5.215716486902927,
|
| 300 |
+
"grad_norm": 45.49678421020508,
|
| 301 |
+
"learning_rate": 1.643490115882754e-05,
|
| 302 |
+
"loss": 0.5922,
|
| 303 |
+
"step": 850
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 5.36979969183359,
|
| 307 |
+
"grad_norm": 36.281124114990234,
|
| 308 |
+
"learning_rate": 1.6264485344239945e-05,
|
| 309 |
+
"loss": 0.6332,
|
| 310 |
+
"step": 875
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 5.523882896764253,
|
| 314 |
+
"grad_norm": 54.04136276245117,
|
| 315 |
+
"learning_rate": 1.6094069529652356e-05,
|
| 316 |
+
"loss": 0.6074,
|
| 317 |
+
"step": 900
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.677966101694915,
|
| 321 |
+
"grad_norm": 43.601593017578125,
|
| 322 |
+
"learning_rate": 1.592365371506476e-05,
|
| 323 |
+
"loss": 0.8893,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.832049306625578,
|
| 328 |
+
"grad_norm": 45.84646987915039,
|
| 329 |
+
"learning_rate": 1.5753237900477167e-05,
|
| 330 |
+
"loss": 0.822,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.9861325115562405,
|
| 335 |
+
"grad_norm": 36.80628204345703,
|
| 336 |
+
"learning_rate": 1.5582822085889574e-05,
|
| 337 |
+
"loss": 0.9505,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 6.0,
|
| 342 |
+
"eval_accuracy": 0.8444767441860465,
|
| 343 |
+
"eval_loss": 0.5237789154052734,
|
| 344 |
+
"eval_macro_f1": 0.8427500092248444,
|
| 345 |
+
"eval_runtime": 0.7891,
|
| 346 |
+
"eval_samples_per_second": 871.864,
|
| 347 |
+
"eval_steps_per_second": 54.491,
|
| 348 |
+
"eval_weighted_f1": 0.8427851941756276,
|
| 349 |
+
"step": 978
|
| 350 |
+
},
|
| 351 |
+
{
|
| 352 |
+
"epoch": 6.135593220338983,
|
| 353 |
+
"grad_norm": 124.89993286132812,
|
| 354 |
+
"learning_rate": 1.5412406271301978e-05,
|
| 355 |
+
"loss": 0.5195,
|
| 356 |
+
"step": 1000
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 6.289676425269645,
|
| 360 |
+
"grad_norm": 69.861083984375,
|
| 361 |
+
"learning_rate": 1.5241990456714383e-05,
|
| 362 |
+
"loss": 0.5254,
|
| 363 |
+
"step": 1025
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 6.443759630200308,
|
| 367 |
+
"grad_norm": 35.58173370361328,
|
| 368 |
+
"learning_rate": 1.5071574642126792e-05,
|
| 369 |
+
"loss": 0.3262,
|
| 370 |
+
"step": 1050
|
| 371 |
+
},
|
| 372 |
+
{
|
| 373 |
+
"epoch": 6.597842835130971,
|
| 374 |
+
"grad_norm": 45.6517219543457,
|
| 375 |
+
"learning_rate": 1.4901158827539198e-05,
|
| 376 |
+
"loss": 0.6022,
|
| 377 |
+
"step": 1075
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 6.751926040061633,
|
| 381 |
+
"grad_norm": 54.69720458984375,
|
| 382 |
+
"learning_rate": 1.4730743012951603e-05,
|
| 383 |
+
"loss": 0.4719,
|
| 384 |
+
"step": 1100
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 6.906009244992296,
|
| 388 |
+
"grad_norm": 20.734577178955078,
|
| 389 |
+
"learning_rate": 1.456032719836401e-05,
|
| 390 |
+
"loss": 0.4483,
|
| 391 |
+
"step": 1125
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 7.0,
|
| 395 |
+
"eval_accuracy": 0.8459302325581395,
|
| 396 |
+
"eval_loss": 0.6327059268951416,
|
| 397 |
+
"eval_macro_f1": 0.8463214195708252,
|
| 398 |
+
"eval_runtime": 0.7593,
|
| 399 |
+
"eval_samples_per_second": 906.044,
|
| 400 |
+
"eval_steps_per_second": 56.628,
|
| 401 |
+
"eval_weighted_f1": 0.8459744643849938,
|
| 402 |
+
"step": 1141
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 7.055469953775039,
|
| 406 |
+
"grad_norm": 83.32035827636719,
|
| 407 |
+
"learning_rate": 1.4389911383776416e-05,
|
| 408 |
+
"loss": 0.5188,
|
| 409 |
+
"step": 1150
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 7.209553158705701,
|
| 413 |
+
"grad_norm": 73.34212493896484,
|
| 414 |
+
"learning_rate": 1.4219495569188822e-05,
|
| 415 |
+
"loss": 0.3097,
|
| 416 |
+
"step": 1175
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 7.363636363636363,
|
| 420 |
+
"grad_norm": 104.00204467773438,
|
| 421 |
+
"learning_rate": 1.4049079754601229e-05,
|
| 422 |
+
"loss": 0.5089,
|
| 423 |
+
"step": 1200
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 7.517719568567026,
|
| 427 |
+
"grad_norm": 81.99729919433594,
|
| 428 |
+
"learning_rate": 1.3878663940013635e-05,
|
| 429 |
+
"loss": 0.519,
|
| 430 |
+
"step": 1225
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 7.671802773497689,
|
| 434 |
+
"grad_norm": 10.343159675598145,
|
| 435 |
+
"learning_rate": 1.370824812542604e-05,
|
| 436 |
+
"loss": 0.4507,
|
| 437 |
+
"step": 1250
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 7.825885978428351,
|
| 441 |
+
"grad_norm": 0.43984729051589966,
|
| 442 |
+
"learning_rate": 1.3537832310838447e-05,
|
| 443 |
+
"loss": 0.5401,
|
| 444 |
+
"step": 1275
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 7.979969183359014,
|
| 448 |
+
"grad_norm": 86.5453109741211,
|
| 449 |
+
"learning_rate": 1.3367416496250853e-05,
|
| 450 |
+
"loss": 0.5885,
|
| 451 |
+
"step": 1300
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 8.0,
|
| 455 |
+
"eval_accuracy": 0.8328488372093024,
|
| 456 |
+
"eval_loss": 0.8175472021102905,
|
| 457 |
+
"eval_macro_f1": 0.8342144806221047,
|
| 458 |
+
"eval_runtime": 0.7627,
|
| 459 |
+
"eval_samples_per_second": 902.073,
|
| 460 |
+
"eval_steps_per_second": 56.38,
|
| 461 |
+
"eval_weighted_f1": 0.8334635877603405,
|
| 462 |
+
"step": 1304
|
| 463 |
+
}
|
| 464 |
+
],
|
| 465 |
+
"logging_steps": 25,
|
| 466 |
+
"max_steps": 3260,
|
| 467 |
+
"num_input_tokens_seen": 0,
|
| 468 |
+
"num_train_epochs": 20,
|
| 469 |
+
"save_steps": 500,
|
| 470 |
+
"stateful_callbacks": {
|
| 471 |
+
"EarlyStoppingCallback": {
|
| 472 |
+
"args": {
|
| 473 |
+
"early_stopping_patience": 3,
|
| 474 |
+
"early_stopping_threshold": 0.0
|
| 475 |
+
},
|
| 476 |
+
"attributes": {
|
| 477 |
+
"early_stopping_patience_counter": 3
|
| 478 |
+
}
|
| 479 |
+
},
|
| 480 |
+
"TrainerControl": {
|
| 481 |
+
"args": {
|
| 482 |
+
"should_epoch_stop": false,
|
| 483 |
+
"should_evaluate": false,
|
| 484 |
+
"should_log": false,
|
| 485 |
+
"should_save": true,
|
| 486 |
+
"should_training_stop": true
|
| 487 |
+
},
|
| 488 |
+
"attributes": {}
|
| 489 |
+
}
|
| 490 |
+
},
|
| 491 |
+
"total_flos": 0.0,
|
| 492 |
+
"train_batch_size": 8,
|
| 493 |
+
"trial_name": null,
|
| 494 |
+
"trial_params": null
|
| 495 |
+
}
|
vifactcheck-bamibert-s43-20260827T131648Z-3858bb98/trainer/checkpoint-815/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 |
+
}
|
vifactcheck-bamibert-s43-20260827T131648Z-3858bb98/trainer/checkpoint-815/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
vifactcheck-bamibert-s43-20260827T131648Z-3858bb98/trainer/checkpoint-815/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 |
+
}
|
vifactcheck-bamibert-s43-20260827T131648Z-3858bb98/trainer/checkpoint-815/trainer_state.json
ADDED
|
@@ -0,0 +1,322 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 815,
|
| 3 |
+
"best_metric": 0.8539482123960574,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/vifactcheck-bamibert-s43-20260827T131648Z-3858bb98/trainer/checkpoint-815",
|
| 5 |
+
"epoch": 5.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 815,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.15408320493066255,
|
| 14 |
+
"grad_norm": 8.965160369873047,
|
| 15 |
+
"learning_rate": 1.47239263803681e-06,
|
| 16 |
+
"loss": 4.4139,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.3081664098613251,
|
| 21 |
+
"grad_norm": 13.331682205200195,
|
| 22 |
+
"learning_rate": 3.0061349693251535e-06,
|
| 23 |
+
"loss": 4.4184,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.4622496147919877,
|
| 28 |
+
"grad_norm": 103.15205383300781,
|
| 29 |
+
"learning_rate": 4.539877300613497e-06,
|
| 30 |
+
"loss": 4.4124,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.6163328197226502,
|
| 35 |
+
"grad_norm": 19.21099090576172,
|
| 36 |
+
"learning_rate": 6.073619631901841e-06,
|
| 37 |
+
"loss": 4.3879,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.7704160246533128,
|
| 42 |
+
"grad_norm": 23.29578971862793,
|
| 43 |
+
"learning_rate": 7.6073619631901856e-06,
|
| 44 |
+
"loss": 4.3548,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.9244992295839753,
|
| 49 |
+
"grad_norm": 20.717185974121094,
|
| 50 |
+
"learning_rate": 9.14110429447853e-06,
|
| 51 |
+
"loss": 4.3273,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 1.0,
|
| 56 |
+
"eval_accuracy": 0.4418604651162791,
|
| 57 |
+
"eval_loss": 1.0659151077270508,
|
| 58 |
+
"eval_macro_f1": 0.35462019711419224,
|
| 59 |
+
"eval_runtime": 0.7815,
|
| 60 |
+
"eval_samples_per_second": 880.332,
|
| 61 |
+
"eval_steps_per_second": 55.021,
|
| 62 |
+
"eval_weighted_f1": 0.35778381033049417,
|
| 63 |
+
"step": 163
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 1.073959938366718,
|
| 67 |
+
"grad_norm": 28.10257339477539,
|
| 68 |
+
"learning_rate": 1.0674846625766873e-05,
|
| 69 |
+
"loss": 4.159,
|
| 70 |
+
"step": 175
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.2280431432973806,
|
| 74 |
+
"grad_norm": 20.976919174194336,
|
| 75 |
+
"learning_rate": 1.2208588957055216e-05,
|
| 76 |
+
"loss": 4.1987,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.3821263482280433,
|
| 81 |
+
"grad_norm": 18.764131546020508,
|
| 82 |
+
"learning_rate": 1.374233128834356e-05,
|
| 83 |
+
"loss": 4.1076,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.5362095531587057,
|
| 88 |
+
"grad_norm": 85.98760223388672,
|
| 89 |
+
"learning_rate": 1.5276073619631903e-05,
|
| 90 |
+
"loss": 3.5784,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.6902927580893683,
|
| 95 |
+
"grad_norm": 70.68370819091797,
|
| 96 |
+
"learning_rate": 1.6809815950920248e-05,
|
| 97 |
+
"loss": 3.3158,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.8443759630200307,
|
| 102 |
+
"grad_norm": 87.6795654296875,
|
| 103 |
+
"learning_rate": 1.834355828220859e-05,
|
| 104 |
+
"loss": 3.0421,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.9984591679506933,
|
| 109 |
+
"grad_norm": 38.58009719848633,
|
| 110 |
+
"learning_rate": 1.9877300613496935e-05,
|
| 111 |
+
"loss": 2.8683,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 2.0,
|
| 116 |
+
"eval_accuracy": 0.7412790697674418,
|
| 117 |
+
"eval_loss": 0.6256099939346313,
|
| 118 |
+
"eval_macro_f1": 0.7366093723446947,
|
| 119 |
+
"eval_runtime": 0.7796,
|
| 120 |
+
"eval_samples_per_second": 882.459,
|
| 121 |
+
"eval_steps_per_second": 55.154,
|
| 122 |
+
"eval_weighted_f1": 0.7371250676827075,
|
| 123 |
+
"step": 326
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 2.147919876733436,
|
| 127 |
+
"grad_norm": 41.14516830444336,
|
| 128 |
+
"learning_rate": 1.9843217450579417e-05,
|
| 129 |
+
"loss": 2.3686,
|
| 130 |
+
"step": 350
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.3020030816640986,
|
| 134 |
+
"grad_norm": 32.573055267333984,
|
| 135 |
+
"learning_rate": 1.967280163599182e-05,
|
| 136 |
+
"loss": 2.3641,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.4560862865947612,
|
| 141 |
+
"grad_norm": 48.540042877197266,
|
| 142 |
+
"learning_rate": 1.9502385821404228e-05,
|
| 143 |
+
"loss": 2.2222,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.610169491525424,
|
| 148 |
+
"grad_norm": 65.02373504638672,
|
| 149 |
+
"learning_rate": 1.9331970006816635e-05,
|
| 150 |
+
"loss": 1.9679,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.7642526964560865,
|
| 155 |
+
"grad_norm": 65.37238311767578,
|
| 156 |
+
"learning_rate": 1.916155419222904e-05,
|
| 157 |
+
"loss": 2.2375,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.9183359013867487,
|
| 162 |
+
"grad_norm": 48.379146575927734,
|
| 163 |
+
"learning_rate": 1.8991138377641446e-05,
|
| 164 |
+
"loss": 2.1481,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 3.0,
|
| 169 |
+
"eval_accuracy": 0.7819767441860465,
|
| 170 |
+
"eval_loss": 0.5680427551269531,
|
| 171 |
+
"eval_macro_f1": 0.7805278799026678,
|
| 172 |
+
"eval_runtime": 0.7704,
|
| 173 |
+
"eval_samples_per_second": 893.029,
|
| 174 |
+
"eval_steps_per_second": 55.814,
|
| 175 |
+
"eval_weighted_f1": 0.78056762599925,
|
| 176 |
+
"step": 489
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 3.0677966101694913,
|
| 180 |
+
"grad_norm": 41.009098052978516,
|
| 181 |
+
"learning_rate": 1.8820722563053853e-05,
|
| 182 |
+
"loss": 1.9255,
|
| 183 |
+
"step": 500
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 3.221879815100154,
|
| 187 |
+
"grad_norm": 48.91205596923828,
|
| 188 |
+
"learning_rate": 1.8650306748466257e-05,
|
| 189 |
+
"loss": 1.6301,
|
| 190 |
+
"step": 525
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.3759630200308166,
|
| 194 |
+
"grad_norm": 74.8825912475586,
|
| 195 |
+
"learning_rate": 1.8479890933878664e-05,
|
| 196 |
+
"loss": 1.5129,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.5300462249614792,
|
| 201 |
+
"grad_norm": 76.0686264038086,
|
| 202 |
+
"learning_rate": 1.8309475119291072e-05,
|
| 203 |
+
"loss": 1.4119,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.684129429892142,
|
| 208 |
+
"grad_norm": 43.675785064697266,
|
| 209 |
+
"learning_rate": 1.8139059304703476e-05,
|
| 210 |
+
"loss": 1.3361,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.8382126348228045,
|
| 215 |
+
"grad_norm": 52.829227447509766,
|
| 216 |
+
"learning_rate": 1.7968643490115886e-05,
|
| 217 |
+
"loss": 1.5664,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.992295839753467,
|
| 222 |
+
"grad_norm": 38.033504486083984,
|
| 223 |
+
"learning_rate": 1.779822767552829e-05,
|
| 224 |
+
"loss": 1.5869,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 4.0,
|
| 229 |
+
"eval_accuracy": 0.8473837209302325,
|
| 230 |
+
"eval_loss": 0.4190322458744049,
|
| 231 |
+
"eval_macro_f1": 0.8467894155655515,
|
| 232 |
+
"eval_runtime": 0.774,
|
| 233 |
+
"eval_samples_per_second": 888.888,
|
| 234 |
+
"eval_steps_per_second": 55.555,
|
| 235 |
+
"eval_weighted_f1": 0.846585051964134,
|
| 236 |
+
"step": 652
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 4.141756548536209,
|
| 240 |
+
"grad_norm": 70.64534759521484,
|
| 241 |
+
"learning_rate": 1.7627811860940697e-05,
|
| 242 |
+
"loss": 1.0347,
|
| 243 |
+
"step": 675
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 4.295839753466872,
|
| 247 |
+
"grad_norm": 28.041074752807617,
|
| 248 |
+
"learning_rate": 1.7457396046353105e-05,
|
| 249 |
+
"loss": 0.8658,
|
| 250 |
+
"step": 700
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 4.449922958397535,
|
| 254 |
+
"grad_norm": 59.811668395996094,
|
| 255 |
+
"learning_rate": 1.728698023176551e-05,
|
| 256 |
+
"loss": 0.8912,
|
| 257 |
+
"step": 725
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.604006163328197,
|
| 261 |
+
"grad_norm": 70.92366027832031,
|
| 262 |
+
"learning_rate": 1.7116564417177916e-05,
|
| 263 |
+
"loss": 1.181,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.75808936825886,
|
| 268 |
+
"grad_norm": 35.35856246948242,
|
| 269 |
+
"learning_rate": 1.6946148602590323e-05,
|
| 270 |
+
"loss": 1.0643,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.9121725731895225,
|
| 275 |
+
"grad_norm": 90.60171508789062,
|
| 276 |
+
"learning_rate": 1.6775732788002727e-05,
|
| 277 |
+
"loss": 1.0013,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 5.0,
|
| 282 |
+
"eval_accuracy": 0.8531976744186046,
|
| 283 |
+
"eval_loss": 0.4695362150669098,
|
| 284 |
+
"eval_macro_f1": 0.8539482123960574,
|
| 285 |
+
"eval_runtime": 0.7637,
|
| 286 |
+
"eval_samples_per_second": 900.928,
|
| 287 |
+
"eval_steps_per_second": 56.308,
|
| 288 |
+
"eval_weighted_f1": 0.8538734490748382,
|
| 289 |
+
"step": 815
|
| 290 |
+
}
|
| 291 |
+
],
|
| 292 |
+
"logging_steps": 25,
|
| 293 |
+
"max_steps": 3260,
|
| 294 |
+
"num_input_tokens_seen": 0,
|
| 295 |
+
"num_train_epochs": 20,
|
| 296 |
+
"save_steps": 500,
|
| 297 |
+
"stateful_callbacks": {
|
| 298 |
+
"EarlyStoppingCallback": {
|
| 299 |
+
"args": {
|
| 300 |
+
"early_stopping_patience": 3,
|
| 301 |
+
"early_stopping_threshold": 0.0
|
| 302 |
+
},
|
| 303 |
+
"attributes": {
|
| 304 |
+
"early_stopping_patience_counter": 0
|
| 305 |
+
}
|
| 306 |
+
},
|
| 307 |
+
"TrainerControl": {
|
| 308 |
+
"args": {
|
| 309 |
+
"should_epoch_stop": false,
|
| 310 |
+
"should_evaluate": false,
|
| 311 |
+
"should_log": false,
|
| 312 |
+
"should_save": true,
|
| 313 |
+
"should_training_stop": false
|
| 314 |
+
},
|
| 315 |
+
"attributes": {}
|
| 316 |
+
}
|
| 317 |
+
},
|
| 318 |
+
"total_flos": 0.0,
|
| 319 |
+
"train_batch_size": 8,
|
| 320 |
+
"trial_name": null,
|
| 321 |
+
"trial_params": null
|
| 322 |
+
}
|
vifactcheck-bamibert-s44-20260827T131947Z-496d6871/trainer/checkpoint-1304/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 |
+
}
|
vifactcheck-bamibert-s44-20260827T131947Z-496d6871/trainer/checkpoint-1304/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
vifactcheck-bamibert-s44-20260827T131947Z-496d6871/trainer/checkpoint-1304/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 |
+
}
|
vifactcheck-bamibert-s44-20260827T131947Z-496d6871/trainer/checkpoint-1304/trainer_state.json
ADDED
|
@@ -0,0 +1,495 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1304,
|
| 3 |
+
"best_metric": 0.8533841176356929,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/vifactcheck-bamibert-s44-20260827T131947Z-496d6871/trainer/checkpoint-1304",
|
| 5 |
+
"epoch": 8.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1304,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.15408320493066255,
|
| 14 |
+
"grad_norm": 8.250635147094727,
|
| 15 |
+
"learning_rate": 1.47239263803681e-06,
|
| 16 |
+
"loss": 4.4124,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.3081664098613251,
|
| 21 |
+
"grad_norm": 11.264381408691406,
|
| 22 |
+
"learning_rate": 3.0061349693251535e-06,
|
| 23 |
+
"loss": 4.3947,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.4622496147919877,
|
| 28 |
+
"grad_norm": 16.838808059692383,
|
| 29 |
+
"learning_rate": 4.539877300613497e-06,
|
| 30 |
+
"loss": 4.3985,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.6163328197226502,
|
| 35 |
+
"grad_norm": 26.60426139831543,
|
| 36 |
+
"learning_rate": 6.073619631901841e-06,
|
| 37 |
+
"loss": 4.4196,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.7704160246533128,
|
| 42 |
+
"grad_norm": 10.414297103881836,
|
| 43 |
+
"learning_rate": 7.6073619631901856e-06,
|
| 44 |
+
"loss": 4.3717,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.9244992295839753,
|
| 49 |
+
"grad_norm": 12.923025131225586,
|
| 50 |
+
"learning_rate": 9.14110429447853e-06,
|
| 51 |
+
"loss": 4.3685,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 1.0,
|
| 56 |
+
"eval_accuracy": 0.4171511627906977,
|
| 57 |
+
"eval_loss": 1.0637348890304565,
|
| 58 |
+
"eval_macro_f1": 0.3552981812383649,
|
| 59 |
+
"eval_runtime": 0.7705,
|
| 60 |
+
"eval_samples_per_second": 892.9,
|
| 61 |
+
"eval_steps_per_second": 55.806,
|
| 62 |
+
"eval_weighted_f1": 0.36024714674903024,
|
| 63 |
+
"step": 163
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 1.073959938366718,
|
| 67 |
+
"grad_norm": 16.9653377532959,
|
| 68 |
+
"learning_rate": 1.0674846625766873e-05,
|
| 69 |
+
"loss": 4.1764,
|
| 70 |
+
"step": 175
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.2280431432973806,
|
| 74 |
+
"grad_norm": 16.933998107910156,
|
| 75 |
+
"learning_rate": 1.2208588957055216e-05,
|
| 76 |
+
"loss": 4.2355,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.3821263482280433,
|
| 81 |
+
"grad_norm": 22.95197296142578,
|
| 82 |
+
"learning_rate": 1.374233128834356e-05,
|
| 83 |
+
"loss": 4.0814,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.5362095531587057,
|
| 88 |
+
"grad_norm": 37.3679313659668,
|
| 89 |
+
"learning_rate": 1.5276073619631903e-05,
|
| 90 |
+
"loss": 3.5941,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.6902927580893683,
|
| 95 |
+
"grad_norm": 65.2331771850586,
|
| 96 |
+
"learning_rate": 1.6809815950920248e-05,
|
| 97 |
+
"loss": 3.1528,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.8443759630200307,
|
| 102 |
+
"grad_norm": 24.468486785888672,
|
| 103 |
+
"learning_rate": 1.834355828220859e-05,
|
| 104 |
+
"loss": 2.9046,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.9984591679506933,
|
| 109 |
+
"grad_norm": 38.351383209228516,
|
| 110 |
+
"learning_rate": 1.9877300613496935e-05,
|
| 111 |
+
"loss": 2.7848,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 2.0,
|
| 116 |
+
"eval_accuracy": 0.7877906976744186,
|
| 117 |
+
"eval_loss": 0.553124725818634,
|
| 118 |
+
"eval_macro_f1": 0.7852786849057214,
|
| 119 |
+
"eval_runtime": 0.7696,
|
| 120 |
+
"eval_samples_per_second": 893.983,
|
| 121 |
+
"eval_steps_per_second": 55.874,
|
| 122 |
+
"eval_weighted_f1": 0.785776704179733,
|
| 123 |
+
"step": 326
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 2.147919876733436,
|
| 127 |
+
"grad_norm": 45.859283447265625,
|
| 128 |
+
"learning_rate": 1.9843217450579417e-05,
|
| 129 |
+
"loss": 2.4254,
|
| 130 |
+
"step": 350
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.3020030816640986,
|
| 134 |
+
"grad_norm": 44.116729736328125,
|
| 135 |
+
"learning_rate": 1.967280163599182e-05,
|
| 136 |
+
"loss": 2.2782,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.4560862865947612,
|
| 141 |
+
"grad_norm": 37.84777069091797,
|
| 142 |
+
"learning_rate": 1.9502385821404228e-05,
|
| 143 |
+
"loss": 2.1475,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.610169491525424,
|
| 148 |
+
"grad_norm": 32.4953498840332,
|
| 149 |
+
"learning_rate": 1.9331970006816635e-05,
|
| 150 |
+
"loss": 2.1429,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.7642526964560865,
|
| 155 |
+
"grad_norm": 108.74443817138672,
|
| 156 |
+
"learning_rate": 1.916155419222904e-05,
|
| 157 |
+
"loss": 1.9236,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.9183359013867487,
|
| 162 |
+
"grad_norm": 32.53221130371094,
|
| 163 |
+
"learning_rate": 1.8991138377641446e-05,
|
| 164 |
+
"loss": 2.099,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 3.0,
|
| 169 |
+
"eval_accuracy": 0.8357558139534884,
|
| 170 |
+
"eval_loss": 0.44927799701690674,
|
| 171 |
+
"eval_macro_f1": 0.8353050219076509,
|
| 172 |
+
"eval_runtime": 0.7733,
|
| 173 |
+
"eval_samples_per_second": 889.744,
|
| 174 |
+
"eval_steps_per_second": 55.609,
|
| 175 |
+
"eval_weighted_f1": 0.835114737268516,
|
| 176 |
+
"step": 489
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 3.0677966101694913,
|
| 180 |
+
"grad_norm": 29.106243133544922,
|
| 181 |
+
"learning_rate": 1.8820722563053853e-05,
|
| 182 |
+
"loss": 1.9165,
|
| 183 |
+
"step": 500
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 3.221879815100154,
|
| 187 |
+
"grad_norm": 33.7355842590332,
|
| 188 |
+
"learning_rate": 1.8650306748466257e-05,
|
| 189 |
+
"loss": 1.4406,
|
| 190 |
+
"step": 525
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.3759630200308166,
|
| 194 |
+
"grad_norm": 50.746238708496094,
|
| 195 |
+
"learning_rate": 1.8479890933878664e-05,
|
| 196 |
+
"loss": 1.4656,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.5300462249614792,
|
| 201 |
+
"grad_norm": 46.01909255981445,
|
| 202 |
+
"learning_rate": 1.8309475119291072e-05,
|
| 203 |
+
"loss": 1.3641,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.684129429892142,
|
| 208 |
+
"grad_norm": 44.456947326660156,
|
| 209 |
+
"learning_rate": 1.8139059304703476e-05,
|
| 210 |
+
"loss": 1.5809,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.8382126348228045,
|
| 215 |
+
"grad_norm": 30.957544326782227,
|
| 216 |
+
"learning_rate": 1.7968643490115886e-05,
|
| 217 |
+
"loss": 1.4668,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.992295839753467,
|
| 222 |
+
"grad_norm": 39.679901123046875,
|
| 223 |
+
"learning_rate": 1.779822767552829e-05,
|
| 224 |
+
"loss": 1.4455,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 4.0,
|
| 229 |
+
"eval_accuracy": 0.8459302325581395,
|
| 230 |
+
"eval_loss": 0.421752393245697,
|
| 231 |
+
"eval_macro_f1": 0.8463028418185369,
|
| 232 |
+
"eval_runtime": 0.7759,
|
| 233 |
+
"eval_samples_per_second": 886.699,
|
| 234 |
+
"eval_steps_per_second": 55.419,
|
| 235 |
+
"eval_weighted_f1": 0.8465000718103232,
|
| 236 |
+
"step": 652
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 4.141756548536209,
|
| 240 |
+
"grad_norm": 54.29524612426758,
|
| 241 |
+
"learning_rate": 1.7627811860940697e-05,
|
| 242 |
+
"loss": 0.919,
|
| 243 |
+
"step": 675
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 4.295839753466872,
|
| 247 |
+
"grad_norm": 38.48719787597656,
|
| 248 |
+
"learning_rate": 1.7457396046353105e-05,
|
| 249 |
+
"loss": 1.2873,
|
| 250 |
+
"step": 700
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 4.449922958397535,
|
| 254 |
+
"grad_norm": 44.072303771972656,
|
| 255 |
+
"learning_rate": 1.728698023176551e-05,
|
| 256 |
+
"loss": 0.7732,
|
| 257 |
+
"step": 725
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.604006163328197,
|
| 261 |
+
"grad_norm": 19.41781997680664,
|
| 262 |
+
"learning_rate": 1.7116564417177916e-05,
|
| 263 |
+
"loss": 1.3187,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.75808936825886,
|
| 268 |
+
"grad_norm": 53.91307067871094,
|
| 269 |
+
"learning_rate": 1.6946148602590323e-05,
|
| 270 |
+
"loss": 1.1108,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.9121725731895225,
|
| 275 |
+
"grad_norm": 18.968048095703125,
|
| 276 |
+
"learning_rate": 1.6775732788002727e-05,
|
| 277 |
+
"loss": 1.0761,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 5.0,
|
| 282 |
+
"eval_accuracy": 0.8488372093023255,
|
| 283 |
+
"eval_loss": 0.4692850708961487,
|
| 284 |
+
"eval_macro_f1": 0.8482725048700153,
|
| 285 |
+
"eval_runtime": 0.78,
|
| 286 |
+
"eval_samples_per_second": 882.0,
|
| 287 |
+
"eval_steps_per_second": 55.125,
|
| 288 |
+
"eval_weighted_f1": 0.8486027994606591,
|
| 289 |
+
"step": 815
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 5.061633281972265,
|
| 293 |
+
"grad_norm": 35.01864242553711,
|
| 294 |
+
"learning_rate": 1.6605316973415134e-05,
|
| 295 |
+
"loss": 0.9844,
|
| 296 |
+
"step": 825
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 5.215716486902927,
|
| 300 |
+
"grad_norm": 71.26637268066406,
|
| 301 |
+
"learning_rate": 1.643490115882754e-05,
|
| 302 |
+
"loss": 0.7444,
|
| 303 |
+
"step": 850
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 5.36979969183359,
|
| 307 |
+
"grad_norm": 46.1616325378418,
|
| 308 |
+
"learning_rate": 1.6264485344239945e-05,
|
| 309 |
+
"loss": 0.7313,
|
| 310 |
+
"step": 875
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 5.523882896764253,
|
| 314 |
+
"grad_norm": 29.362157821655273,
|
| 315 |
+
"learning_rate": 1.6094069529652356e-05,
|
| 316 |
+
"loss": 0.7023,
|
| 317 |
+
"step": 900
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.677966101694915,
|
| 321 |
+
"grad_norm": 59.7331657409668,
|
| 322 |
+
"learning_rate": 1.592365371506476e-05,
|
| 323 |
+
"loss": 0.6263,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.832049306625578,
|
| 328 |
+
"grad_norm": 42.20322036743164,
|
| 329 |
+
"learning_rate": 1.5753237900477167e-05,
|
| 330 |
+
"loss": 0.6556,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.9861325115562405,
|
| 335 |
+
"grad_norm": 37.34501266479492,
|
| 336 |
+
"learning_rate": 1.5582822085889574e-05,
|
| 337 |
+
"loss": 0.7749,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 6.0,
|
| 342 |
+
"eval_accuracy": 0.8372093023255814,
|
| 343 |
+
"eval_loss": 0.6001764535903931,
|
| 344 |
+
"eval_macro_f1": 0.8369510375624243,
|
| 345 |
+
"eval_runtime": 0.776,
|
| 346 |
+
"eval_samples_per_second": 886.567,
|
| 347 |
+
"eval_steps_per_second": 55.41,
|
| 348 |
+
"eval_weighted_f1": 0.8371920026249331,
|
| 349 |
+
"step": 978
|
| 350 |
+
},
|
| 351 |
+
{
|
| 352 |
+
"epoch": 6.135593220338983,
|
| 353 |
+
"grad_norm": 82.13500213623047,
|
| 354 |
+
"learning_rate": 1.5412406271301978e-05,
|
| 355 |
+
"loss": 0.5272,
|
| 356 |
+
"step": 1000
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 6.289676425269645,
|
| 360 |
+
"grad_norm": 159.67825317382812,
|
| 361 |
+
"learning_rate": 1.5241990456714383e-05,
|
| 362 |
+
"loss": 0.4929,
|
| 363 |
+
"step": 1025
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 6.443759630200308,
|
| 367 |
+
"grad_norm": 76.81694793701172,
|
| 368 |
+
"learning_rate": 1.5071574642126792e-05,
|
| 369 |
+
"loss": 0.6785,
|
| 370 |
+
"step": 1050
|
| 371 |
+
},
|
| 372 |
+
{
|
| 373 |
+
"epoch": 6.597842835130971,
|
| 374 |
+
"grad_norm": 110.73750305175781,
|
| 375 |
+
"learning_rate": 1.4901158827539198e-05,
|
| 376 |
+
"loss": 0.6547,
|
| 377 |
+
"step": 1075
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 6.751926040061633,
|
| 381 |
+
"grad_norm": 0.8977804183959961,
|
| 382 |
+
"learning_rate": 1.4730743012951603e-05,
|
| 383 |
+
"loss": 0.5578,
|
| 384 |
+
"step": 1100
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 6.906009244992296,
|
| 388 |
+
"grad_norm": 40.81841278076172,
|
| 389 |
+
"learning_rate": 1.456032719836401e-05,
|
| 390 |
+
"loss": 0.4328,
|
| 391 |
+
"step": 1125
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 7.0,
|
| 395 |
+
"eval_accuracy": 0.8430232558139535,
|
| 396 |
+
"eval_loss": 0.7302688360214233,
|
| 397 |
+
"eval_macro_f1": 0.8424449637527144,
|
| 398 |
+
"eval_runtime": 0.788,
|
| 399 |
+
"eval_samples_per_second": 873.074,
|
| 400 |
+
"eval_steps_per_second": 54.567,
|
| 401 |
+
"eval_weighted_f1": 0.8428972602884084,
|
| 402 |
+
"step": 1141
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 7.055469953775039,
|
| 406 |
+
"grad_norm": 54.07719421386719,
|
| 407 |
+
"learning_rate": 1.4389911383776416e-05,
|
| 408 |
+
"loss": 0.6282,
|
| 409 |
+
"step": 1150
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 7.209553158705701,
|
| 413 |
+
"grad_norm": 33.80830383300781,
|
| 414 |
+
"learning_rate": 1.4219495569188822e-05,
|
| 415 |
+
"loss": 0.4269,
|
| 416 |
+
"step": 1175
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 7.363636363636363,
|
| 420 |
+
"grad_norm": 13.885575294494629,
|
| 421 |
+
"learning_rate": 1.4049079754601229e-05,
|
| 422 |
+
"loss": 0.3345,
|
| 423 |
+
"step": 1200
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 7.517719568567026,
|
| 427 |
+
"grad_norm": 103.48690032958984,
|
| 428 |
+
"learning_rate": 1.3878663940013635e-05,
|
| 429 |
+
"loss": 0.3776,
|
| 430 |
+
"step": 1225
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 7.671802773497689,
|
| 434 |
+
"grad_norm": 224.79295349121094,
|
| 435 |
+
"learning_rate": 1.370824812542604e-05,
|
| 436 |
+
"loss": 0.8548,
|
| 437 |
+
"step": 1250
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 7.825885978428351,
|
| 441 |
+
"grad_norm": 69.80616760253906,
|
| 442 |
+
"learning_rate": 1.3537832310838447e-05,
|
| 443 |
+
"loss": 0.3026,
|
| 444 |
+
"step": 1275
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 7.979969183359014,
|
| 448 |
+
"grad_norm": 19.676177978515625,
|
| 449 |
+
"learning_rate": 1.3367416496250853e-05,
|
| 450 |
+
"loss": 0.4964,
|
| 451 |
+
"step": 1300
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 8.0,
|
| 455 |
+
"eval_accuracy": 0.8531976744186046,
|
| 456 |
+
"eval_loss": 0.7939755916595459,
|
| 457 |
+
"eval_macro_f1": 0.8533841176356929,
|
| 458 |
+
"eval_runtime": 0.7761,
|
| 459 |
+
"eval_samples_per_second": 886.479,
|
| 460 |
+
"eval_steps_per_second": 55.405,
|
| 461 |
+
"eval_weighted_f1": 0.8538911806395589,
|
| 462 |
+
"step": 1304
|
| 463 |
+
}
|
| 464 |
+
],
|
| 465 |
+
"logging_steps": 25,
|
| 466 |
+
"max_steps": 3260,
|
| 467 |
+
"num_input_tokens_seen": 0,
|
| 468 |
+
"num_train_epochs": 20,
|
| 469 |
+
"save_steps": 500,
|
| 470 |
+
"stateful_callbacks": {
|
| 471 |
+
"EarlyStoppingCallback": {
|
| 472 |
+
"args": {
|
| 473 |
+
"early_stopping_patience": 3,
|
| 474 |
+
"early_stopping_threshold": 0.0
|
| 475 |
+
},
|
| 476 |
+
"attributes": {
|
| 477 |
+
"early_stopping_patience_counter": 0
|
| 478 |
+
}
|
| 479 |
+
},
|
| 480 |
+
"TrainerControl": {
|
| 481 |
+
"args": {
|
| 482 |
+
"should_epoch_stop": false,
|
| 483 |
+
"should_evaluate": false,
|
| 484 |
+
"should_log": false,
|
| 485 |
+
"should_save": true,
|
| 486 |
+
"should_training_stop": false
|
| 487 |
+
},
|
| 488 |
+
"attributes": {}
|
| 489 |
+
}
|
| 490 |
+
},
|
| 491 |
+
"total_flos": 0.0,
|
| 492 |
+
"train_batch_size": 8,
|
| 493 |
+
"trial_name": null,
|
| 494 |
+
"trial_params": null
|
| 495 |
+
}
|
vifactcheck-bamibert-s44-20260827T131947Z-496d6871/trainer/checkpoint-1793/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 |
+
}
|
vifactcheck-bamibert-s44-20260827T131947Z-496d6871/trainer/checkpoint-1793/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
vifactcheck-bamibert-s44-20260827T131947Z-496d6871/trainer/checkpoint-1793/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 |
+
}
|