Add files using upload-large-folder tool
Browse filesThis view is limited to 50 files because it contains too many changes. See raw diff
- ood-vifactcheck-phobert-topic-domain-04-s44-20260826T111516Z-ab4112e3/trainer/checkpoint-1316/trainer_state.json +484 -0
- ood-vifactcheck-phobert-topic-domain-04-s44-20260826T111516Z-ab4112e3/trainer/checkpoint-1880/added_tokens.json +3 -0
- ood-vifactcheck-phobert-topic-domain-04-s44-20260826T111516Z-ab4112e3/trainer/checkpoint-1880/bpe.codes +0 -0
- ood-vifactcheck-phobert-topic-domain-04-s44-20260826T111516Z-ab4112e3/trainer/checkpoint-1880/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-topic-domain-04-s44-20260826T111516Z-ab4112e3/trainer/checkpoint-1880/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-topic-domain-04-s44-20260826T111516Z-ab4112e3/trainer/checkpoint-1880/trainer_state.json +678 -0
- ood-vifactcheck-phobert-topic-domain-04-s44-20260826T111516Z-ab4112e3/trainer/checkpoint-1880/vocab.txt +0 -0
- ood-vifactcheck-phobert-topic-domain-05-s44-20260826T111852Z-ab4112e3/trainer/checkpoint-1225/added_tokens.json +3 -0
- ood-vifactcheck-phobert-topic-domain-05-s44-20260826T111852Z-ab4112e3/trainer/checkpoint-1225/bpe.codes +0 -0
- ood-vifactcheck-phobert-topic-domain-05-s44-20260826T111852Z-ab4112e3/trainer/checkpoint-1225/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-topic-domain-05-s44-20260826T111852Z-ab4112e3/trainer/checkpoint-1225/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-topic-domain-05-s44-20260826T111852Z-ab4112e3/trainer/checkpoint-1225/trainer_state.json +463 -0
- ood-vifactcheck-phobert-topic-domain-05-s44-20260826T111852Z-ab4112e3/trainer/checkpoint-1225/vocab.txt +0 -0
- ood-vifactcheck-phobert-topic-domain-05-s44-20260826T111852Z-ab4112e3/trainer/checkpoint-1750/added_tokens.json +3 -0
- ood-vifactcheck-phobert-topic-domain-05-s44-20260826T111852Z-ab4112e3/trainer/checkpoint-1750/bpe.codes +0 -0
- ood-vifactcheck-phobert-topic-domain-05-s44-20260826T111852Z-ab4112e3/trainer/checkpoint-1750/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-topic-domain-05-s44-20260826T111852Z-ab4112e3/trainer/checkpoint-1750/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-topic-domain-05-s44-20260826T111852Z-ab4112e3/trainer/checkpoint-1750/trainer_state.json +643 -0
- ood-vifactcheck-phobert-topic-domain-05-s44-20260826T111852Z-ab4112e3/trainer/checkpoint-1750/vocab.txt +0 -0
- ood-vifactcheck-phobert-topic-domain-06-s44-20260826T112212Z-ab4112e3/trainer/checkpoint-1104/added_tokens.json +3 -0
- ood-vifactcheck-phobert-topic-domain-06-s44-20260826T112212Z-ab4112e3/trainer/checkpoint-1104/bpe.codes +0 -0
- ood-vifactcheck-phobert-topic-domain-06-s44-20260826T112212Z-ab4112e3/trainer/checkpoint-1104/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-topic-domain-06-s44-20260826T112212Z-ab4112e3/trainer/checkpoint-1104/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-topic-domain-06-s44-20260826T112212Z-ab4112e3/trainer/checkpoint-1104/trainer_state.json +417 -0
- ood-vifactcheck-phobert-topic-domain-06-s44-20260826T112212Z-ab4112e3/trainer/checkpoint-1104/vocab.txt +0 -0
- ood-vifactcheck-phobert-topic-domain-06-s44-20260826T112212Z-ab4112e3/trainer/checkpoint-1656/added_tokens.json +3 -0
- ood-vifactcheck-phobert-topic-domain-06-s44-20260826T112212Z-ab4112e3/trainer/checkpoint-1656/bpe.codes +0 -0
- ood-vifactcheck-phobert-topic-domain-06-s44-20260826T112212Z-ab4112e3/trainer/checkpoint-1656/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-topic-domain-06-s44-20260826T112212Z-ab4112e3/trainer/checkpoint-1656/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-topic-domain-06-s44-20260826T112212Z-ab4112e3/trainer/checkpoint-1656/trainer_state.json +604 -0
- ood-vifactcheck-phobert-topic-domain-06-s44-20260826T112212Z-ab4112e3/trainer/checkpoint-1656/vocab.txt +0 -0
- ood-vifactcheck-phobert-topic-domain-07-s44-20260826T112526Z-ab4112e3/trainer/checkpoint-1379/added_tokens.json +3 -0
- ood-vifactcheck-phobert-topic-domain-07-s44-20260826T112526Z-ab4112e3/trainer/checkpoint-1379/bpe.codes +0 -0
- ood-vifactcheck-phobert-topic-domain-07-s44-20260826T112526Z-ab4112e3/trainer/checkpoint-1379/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-topic-domain-07-s44-20260826T112526Z-ab4112e3/trainer/checkpoint-1379/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-topic-domain-07-s44-20260826T112526Z-ab4112e3/trainer/checkpoint-1379/trainer_state.json +505 -0
- ood-vifactcheck-phobert-topic-domain-07-s44-20260826T112526Z-ab4112e3/trainer/checkpoint-1379/vocab.txt +0 -0
- ood-vifactcheck-phobert-topic-domain-07-s44-20260826T112526Z-ab4112e3/trainer/checkpoint-788/added_tokens.json +3 -0
- ood-vifactcheck-phobert-topic-domain-07-s44-20260826T112526Z-ab4112e3/trainer/checkpoint-788/bpe.codes +0 -0
- ood-vifactcheck-phobert-topic-domain-07-s44-20260826T112526Z-ab4112e3/trainer/checkpoint-788/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-topic-domain-07-s44-20260826T112526Z-ab4112e3/trainer/checkpoint-788/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-topic-domain-07-s44-20260826T112526Z-ab4112e3/trainer/checkpoint-788/trainer_state.json +304 -0
- ood-vifactcheck-phobert-topic-domain-07-s44-20260826T112526Z-ab4112e3/trainer/checkpoint-788/vocab.txt +0 -0
- ood-vifactcheck-phobert-topic-domain-08-s44-20260826T112804Z-ab4112e3/trainer/checkpoint-1360/added_tokens.json +3 -0
- ood-vifactcheck-phobert-topic-domain-08-s44-20260826T112804Z-ab4112e3/trainer/checkpoint-1360/bpe.codes +0 -0
- ood-vifactcheck-phobert-topic-domain-08-s44-20260826T112804Z-ab4112e3/trainer/checkpoint-1360/special_tokens_map.json +9 -0
- ood-vifactcheck-phobert-topic-domain-08-s44-20260826T112804Z-ab4112e3/trainer/checkpoint-1360/tokenizer_config.json +55 -0
- ood-vifactcheck-phobert-topic-domain-08-s44-20260826T112804Z-ab4112e3/trainer/checkpoint-1360/trainer_state.json +509 -0
- ood-vifactcheck-phobert-topic-domain-08-s44-20260826T112804Z-ab4112e3/trainer/checkpoint-1360/vocab.txt +0 -0
- ood-vifactcheck-phobert-topic-domain-08-s44-20260826T112804Z-ab4112e3/trainer/checkpoint-850/added_tokens.json +3 -0
ood-vifactcheck-phobert-topic-domain-04-s44-20260826T111516Z-ab4112e3/trainer/checkpoint-1316/trainer_state.json
ADDED
|
@@ -0,0 +1,484 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1316,
|
| 3 |
+
"best_metric": 0.842258582036095,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-topic-domain-04-s44-20260826T111516Z-ab4112e3/trainer/checkpoint-1316",
|
| 5 |
+
"epoch": 7.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1316,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.13351134846461948,
|
| 14 |
+
"grad_norm": 6.047441005706787,
|
| 15 |
+
"learning_rate": 1.276595744680851e-06,
|
| 16 |
+
"loss": 4.4062,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.26702269692923897,
|
| 21 |
+
"grad_norm": 6.4811110496521,
|
| 22 |
+
"learning_rate": 2.6063829787234047e-06,
|
| 23 |
+
"loss": 4.4013,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.40053404539385845,
|
| 28 |
+
"grad_norm": 7.357682228088379,
|
| 29 |
+
"learning_rate": 3.936170212765958e-06,
|
| 30 |
+
"loss": 4.3671,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5340453938584779,
|
| 35 |
+
"grad_norm": 7.46104621887207,
|
| 36 |
+
"learning_rate": 5.265957446808511e-06,
|
| 37 |
+
"loss": 4.3112,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6675567423230975,
|
| 42 |
+
"grad_norm": 9.131397247314453,
|
| 43 |
+
"learning_rate": 6.595744680851064e-06,
|
| 44 |
+
"loss": 4.2309,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8010680907877169,
|
| 49 |
+
"grad_norm": 14.2952880859375,
|
| 50 |
+
"learning_rate": 7.925531914893617e-06,
|
| 51 |
+
"loss": 4.1262,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9345794392523364,
|
| 56 |
+
"grad_norm": 15.645665168762207,
|
| 57 |
+
"learning_rate": 9.255319148936171e-06,
|
| 58 |
+
"loss": 3.9533,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.609181141439206,
|
| 64 |
+
"eval_loss": 0.8993740677833557,
|
| 65 |
+
"eval_macro_f1": 0.5772103729957749,
|
| 66 |
+
"eval_runtime": 0.5542,
|
| 67 |
+
"eval_samples_per_second": 1454.416,
|
| 68 |
+
"eval_steps_per_second": 92.029,
|
| 69 |
+
"eval_weighted_f1": 0.5786168029157498,
|
| 70 |
+
"step": 188
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0640854472630175,
|
| 74 |
+
"grad_norm": 34.74980926513672,
|
| 75 |
+
"learning_rate": 1.0585106382978725e-05,
|
| 76 |
+
"loss": 3.6005,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.1975967957276368,
|
| 81 |
+
"grad_norm": 28.335071563720703,
|
| 82 |
+
"learning_rate": 1.1914893617021277e-05,
|
| 83 |
+
"loss": 3.1783,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.3311081441922563,
|
| 88 |
+
"grad_norm": 26.67999839782715,
|
| 89 |
+
"learning_rate": 1.3244680851063832e-05,
|
| 90 |
+
"loss": 3.0771,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.4646194926568759,
|
| 95 |
+
"grad_norm": 51.67438507080078,
|
| 96 |
+
"learning_rate": 1.4574468085106384e-05,
|
| 97 |
+
"loss": 2.9497,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.5981308411214954,
|
| 102 |
+
"grad_norm": 28.291353225708008,
|
| 103 |
+
"learning_rate": 1.590425531914894e-05,
|
| 104 |
+
"loss": 2.6225,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7316421895861147,
|
| 109 |
+
"grad_norm": 44.70147705078125,
|
| 110 |
+
"learning_rate": 1.723404255319149e-05,
|
| 111 |
+
"loss": 2.4372,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.8651535380507345,
|
| 116 |
+
"grad_norm": 62.695106506347656,
|
| 117 |
+
"learning_rate": 1.8563829787234043e-05,
|
| 118 |
+
"loss": 2.162,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 1.9986648865153538,
|
| 123 |
+
"grad_norm": 24.98824119567871,
|
| 124 |
+
"learning_rate": 1.9893617021276595e-05,
|
| 125 |
+
"loss": 2.2262,
|
| 126 |
+
"step": 375
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"epoch": 2.0,
|
| 130 |
+
"eval_accuracy": 0.7965260545905707,
|
| 131 |
+
"eval_loss": 0.5188497304916382,
|
| 132 |
+
"eval_macro_f1": 0.7948480192243981,
|
| 133 |
+
"eval_runtime": 0.5222,
|
| 134 |
+
"eval_samples_per_second": 1543.488,
|
| 135 |
+
"eval_steps_per_second": 97.665,
|
| 136 |
+
"eval_weighted_f1": 0.7950051743295525,
|
| 137 |
+
"step": 376
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.128170894526035,
|
| 141 |
+
"grad_norm": 53.27727127075195,
|
| 142 |
+
"learning_rate": 1.986406619385343e-05,
|
| 143 |
+
"loss": 1.7151,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.2616822429906542,
|
| 148 |
+
"grad_norm": 87.19738006591797,
|
| 149 |
+
"learning_rate": 1.971631205673759e-05,
|
| 150 |
+
"loss": 1.6824,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.3951935914552736,
|
| 155 |
+
"grad_norm": 41.84999465942383,
|
| 156 |
+
"learning_rate": 1.956855791962175e-05,
|
| 157 |
+
"loss": 1.6127,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.5287049399198933,
|
| 162 |
+
"grad_norm": 30.91926383972168,
|
| 163 |
+
"learning_rate": 1.9420803782505913e-05,
|
| 164 |
+
"loss": 1.7214,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.6622162883845126,
|
| 169 |
+
"grad_norm": 56.13572311401367,
|
| 170 |
+
"learning_rate": 1.927304964539007e-05,
|
| 171 |
+
"loss": 1.6437,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.795727636849132,
|
| 176 |
+
"grad_norm": 52.92278289794922,
|
| 177 |
+
"learning_rate": 1.9125295508274234e-05,
|
| 178 |
+
"loss": 1.6039,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 2.9292389853137517,
|
| 183 |
+
"grad_norm": 43.0755500793457,
|
| 184 |
+
"learning_rate": 1.8977541371158394e-05,
|
| 185 |
+
"loss": 1.6599,
|
| 186 |
+
"step": 550
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 3.0,
|
| 190 |
+
"eval_accuracy": 0.8002481389578163,
|
| 191 |
+
"eval_loss": 0.5185101628303528,
|
| 192 |
+
"eval_macro_f1": 0.7985617142876156,
|
| 193 |
+
"eval_runtime": 0.5251,
|
| 194 |
+
"eval_samples_per_second": 1534.946,
|
| 195 |
+
"eval_steps_per_second": 97.124,
|
| 196 |
+
"eval_weighted_f1": 0.7984130121418866,
|
| 197 |
+
"step": 564
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.0587449933244324,
|
| 201 |
+
"grad_norm": 34.79388427734375,
|
| 202 |
+
"learning_rate": 1.8829787234042557e-05,
|
| 203 |
+
"loss": 1.4772,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.192256341789052,
|
| 208 |
+
"grad_norm": 22.63060188293457,
|
| 209 |
+
"learning_rate": 1.8682033096926714e-05,
|
| 210 |
+
"loss": 1.1397,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.3257676902536715,
|
| 215 |
+
"grad_norm": 37.29880142211914,
|
| 216 |
+
"learning_rate": 1.8534278959810878e-05,
|
| 217 |
+
"loss": 1.0699,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.4592790387182912,
|
| 222 |
+
"grad_norm": 18.859315872192383,
|
| 223 |
+
"learning_rate": 1.8386524822695038e-05,
|
| 224 |
+
"loss": 1.0505,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.5927903871829105,
|
| 229 |
+
"grad_norm": 63.72948455810547,
|
| 230 |
+
"learning_rate": 1.8238770685579198e-05,
|
| 231 |
+
"loss": 1.2013,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.72630173564753,
|
| 236 |
+
"grad_norm": 42.204322814941406,
|
| 237 |
+
"learning_rate": 1.8091016548463358e-05,
|
| 238 |
+
"loss": 1.1122,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.8598130841121496,
|
| 243 |
+
"grad_norm": 70.12254333496094,
|
| 244 |
+
"learning_rate": 1.7943262411347518e-05,
|
| 245 |
+
"loss": 1.1854,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 3.993324432576769,
|
| 250 |
+
"grad_norm": 85.46212768554688,
|
| 251 |
+
"learning_rate": 1.779550827423168e-05,
|
| 252 |
+
"loss": 1.3457,
|
| 253 |
+
"step": 750
|
| 254 |
+
},
|
| 255 |
+
{
|
| 256 |
+
"epoch": 4.0,
|
| 257 |
+
"eval_accuracy": 0.8089330024813896,
|
| 258 |
+
"eval_loss": 0.546148955821991,
|
| 259 |
+
"eval_macro_f1": 0.8094617405060895,
|
| 260 |
+
"eval_runtime": 0.5271,
|
| 261 |
+
"eval_samples_per_second": 1529.124,
|
| 262 |
+
"eval_steps_per_second": 96.756,
|
| 263 |
+
"eval_weighted_f1": 0.8094706960552178,
|
| 264 |
+
"step": 752
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.12283044058745,
|
| 268 |
+
"grad_norm": 27.955347061157227,
|
| 269 |
+
"learning_rate": 1.7647754137115842e-05,
|
| 270 |
+
"loss": 0.757,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.25634178905207,
|
| 275 |
+
"grad_norm": 49.59231185913086,
|
| 276 |
+
"learning_rate": 1.7500000000000002e-05,
|
| 277 |
+
"loss": 0.6709,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.389853137516689,
|
| 282 |
+
"grad_norm": 52.06037521362305,
|
| 283 |
+
"learning_rate": 1.7352245862884162e-05,
|
| 284 |
+
"loss": 0.8391,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.5233644859813085,
|
| 289 |
+
"grad_norm": 54.212562561035156,
|
| 290 |
+
"learning_rate": 1.7204491725768322e-05,
|
| 291 |
+
"loss": 0.7409,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.656875834445928,
|
| 296 |
+
"grad_norm": 38.95555877685547,
|
| 297 |
+
"learning_rate": 1.7056737588652482e-05,
|
| 298 |
+
"loss": 0.6671,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.790387182910547,
|
| 303 |
+
"grad_norm": 34.9852294921875,
|
| 304 |
+
"learning_rate": 1.6908983451536646e-05,
|
| 305 |
+
"loss": 0.7324,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 4.923898531375167,
|
| 310 |
+
"grad_norm": 28.143545150756836,
|
| 311 |
+
"learning_rate": 1.6761229314420806e-05,
|
| 312 |
+
"loss": 0.818,
|
| 313 |
+
"step": 925
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 5.0,
|
| 317 |
+
"eval_accuracy": 0.7977667493796526,
|
| 318 |
+
"eval_loss": 0.739179790019989,
|
| 319 |
+
"eval_macro_f1": 0.790721734403641,
|
| 320 |
+
"eval_runtime": 0.5778,
|
| 321 |
+
"eval_samples_per_second": 1394.872,
|
| 322 |
+
"eval_steps_per_second": 88.261,
|
| 323 |
+
"eval_weighted_f1": 0.791121932272409,
|
| 324 |
+
"step": 940
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.0534045393858475,
|
| 328 |
+
"grad_norm": 59.654903411865234,
|
| 329 |
+
"learning_rate": 1.6613475177304966e-05,
|
| 330 |
+
"loss": 0.9792,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.186915887850467,
|
| 335 |
+
"grad_norm": 42.62851333618164,
|
| 336 |
+
"learning_rate": 1.6465721040189126e-05,
|
| 337 |
+
"loss": 0.7497,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.320427236315087,
|
| 342 |
+
"grad_norm": 123.19102478027344,
|
| 343 |
+
"learning_rate": 1.6317966903073286e-05,
|
| 344 |
+
"loss": 0.5186,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.453938584779706,
|
| 349 |
+
"grad_norm": 66.54837036132812,
|
| 350 |
+
"learning_rate": 1.6170212765957446e-05,
|
| 351 |
+
"loss": 0.4953,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.587449933244326,
|
| 356 |
+
"grad_norm": 16.784555435180664,
|
| 357 |
+
"learning_rate": 1.602245862884161e-05,
|
| 358 |
+
"loss": 0.6363,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.720961281708945,
|
| 363 |
+
"grad_norm": 28.712814331054688,
|
| 364 |
+
"learning_rate": 1.587470449172577e-05,
|
| 365 |
+
"loss": 0.6987,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 5.854472630173564,
|
| 370 |
+
"grad_norm": 3.6170356273651123,
|
| 371 |
+
"learning_rate": 1.572695035460993e-05,
|
| 372 |
+
"loss": 0.5691,
|
| 373 |
+
"step": 1100
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 5.9879839786381845,
|
| 377 |
+
"grad_norm": 59.6114501953125,
|
| 378 |
+
"learning_rate": 1.557919621749409e-05,
|
| 379 |
+
"loss": 0.5618,
|
| 380 |
+
"step": 1125
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 6.0,
|
| 384 |
+
"eval_accuracy": 0.8138957816377171,
|
| 385 |
+
"eval_loss": 0.6991787552833557,
|
| 386 |
+
"eval_macro_f1": 0.8165257465569141,
|
| 387 |
+
"eval_runtime": 0.5681,
|
| 388 |
+
"eval_samples_per_second": 1418.873,
|
| 389 |
+
"eval_steps_per_second": 89.78,
|
| 390 |
+
"eval_weighted_f1": 0.8164177546029194,
|
| 391 |
+
"step": 1128
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 6.117489986648865,
|
| 395 |
+
"grad_norm": 3.2209362983703613,
|
| 396 |
+
"learning_rate": 1.543144208037825e-05,
|
| 397 |
+
"loss": 0.4893,
|
| 398 |
+
"step": 1150
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.251001335113485,
|
| 402 |
+
"grad_norm": 16.82813262939453,
|
| 403 |
+
"learning_rate": 1.5283687943262414e-05,
|
| 404 |
+
"loss": 0.5943,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.384512683578104,
|
| 409 |
+
"grad_norm": 1.2777460813522339,
|
| 410 |
+
"learning_rate": 1.5135933806146574e-05,
|
| 411 |
+
"loss": 0.3216,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.518024032042724,
|
| 416 |
+
"grad_norm": 86.51032257080078,
|
| 417 |
+
"learning_rate": 1.4988179669030734e-05,
|
| 418 |
+
"loss": 0.398,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.651535380507343,
|
| 423 |
+
"grad_norm": 31.18537712097168,
|
| 424 |
+
"learning_rate": 1.4840425531914894e-05,
|
| 425 |
+
"loss": 0.4206,
|
| 426 |
+
"step": 1250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 6.785046728971962,
|
| 430 |
+
"grad_norm": 98.30158233642578,
|
| 431 |
+
"learning_rate": 1.4692671394799056e-05,
|
| 432 |
+
"loss": 0.4636,
|
| 433 |
+
"step": 1275
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 6.9185580774365825,
|
| 437 |
+
"grad_norm": 2.0132503509521484,
|
| 438 |
+
"learning_rate": 1.4544917257683215e-05,
|
| 439 |
+
"loss": 0.3674,
|
| 440 |
+
"step": 1300
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 7.0,
|
| 444 |
+
"eval_accuracy": 0.8424317617866005,
|
| 445 |
+
"eval_loss": 0.7080403566360474,
|
| 446 |
+
"eval_macro_f1": 0.842258582036095,
|
| 447 |
+
"eval_runtime": 0.533,
|
| 448 |
+
"eval_samples_per_second": 1512.286,
|
| 449 |
+
"eval_steps_per_second": 95.691,
|
| 450 |
+
"eval_weighted_f1": 0.8421080192436209,
|
| 451 |
+
"step": 1316
|
| 452 |
+
}
|
| 453 |
+
],
|
| 454 |
+
"logging_steps": 25,
|
| 455 |
+
"max_steps": 3760,
|
| 456 |
+
"num_input_tokens_seen": 0,
|
| 457 |
+
"num_train_epochs": 20,
|
| 458 |
+
"save_steps": 500,
|
| 459 |
+
"stateful_callbacks": {
|
| 460 |
+
"EarlyStoppingCallback": {
|
| 461 |
+
"args": {
|
| 462 |
+
"early_stopping_patience": 3,
|
| 463 |
+
"early_stopping_threshold": 0.0
|
| 464 |
+
},
|
| 465 |
+
"attributes": {
|
| 466 |
+
"early_stopping_patience_counter": 0
|
| 467 |
+
}
|
| 468 |
+
},
|
| 469 |
+
"TrainerControl": {
|
| 470 |
+
"args": {
|
| 471 |
+
"should_epoch_stop": false,
|
| 472 |
+
"should_evaluate": false,
|
| 473 |
+
"should_log": false,
|
| 474 |
+
"should_save": true,
|
| 475 |
+
"should_training_stop": false
|
| 476 |
+
},
|
| 477 |
+
"attributes": {}
|
| 478 |
+
}
|
| 479 |
+
},
|
| 480 |
+
"total_flos": 0.0,
|
| 481 |
+
"train_batch_size": 8,
|
| 482 |
+
"trial_name": null,
|
| 483 |
+
"trial_params": null
|
| 484 |
+
}
|
ood-vifactcheck-phobert-topic-domain-04-s44-20260826T111516Z-ab4112e3/trainer/checkpoint-1880/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-topic-domain-04-s44-20260826T111516Z-ab4112e3/trainer/checkpoint-1880/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-04-s44-20260826T111516Z-ab4112e3/trainer/checkpoint-1880/special_tokens_map.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": "<mask>",
|
| 6 |
+
"pad_token": "<pad>",
|
| 7 |
+
"sep_token": "</s>",
|
| 8 |
+
"unk_token": "<unk>"
|
| 9 |
+
}
|
ood-vifactcheck-phobert-topic-domain-04-s44-20260826T111516Z-ab4112e3/trainer/checkpoint-1880/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"64000": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 256,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "PhobertTokenizer",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-phobert-topic-domain-04-s44-20260826T111516Z-ab4112e3/trainer/checkpoint-1880/trainer_state.json
ADDED
|
@@ -0,0 +1,678 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1316,
|
| 3 |
+
"best_metric": 0.842258582036095,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-topic-domain-04-s44-20260826T111516Z-ab4112e3/trainer/checkpoint-1316",
|
| 5 |
+
"epoch": 10.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1880,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.13351134846461948,
|
| 14 |
+
"grad_norm": 6.047441005706787,
|
| 15 |
+
"learning_rate": 1.276595744680851e-06,
|
| 16 |
+
"loss": 4.4062,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.26702269692923897,
|
| 21 |
+
"grad_norm": 6.4811110496521,
|
| 22 |
+
"learning_rate": 2.6063829787234047e-06,
|
| 23 |
+
"loss": 4.4013,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.40053404539385845,
|
| 28 |
+
"grad_norm": 7.357682228088379,
|
| 29 |
+
"learning_rate": 3.936170212765958e-06,
|
| 30 |
+
"loss": 4.3671,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5340453938584779,
|
| 35 |
+
"grad_norm": 7.46104621887207,
|
| 36 |
+
"learning_rate": 5.265957446808511e-06,
|
| 37 |
+
"loss": 4.3112,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6675567423230975,
|
| 42 |
+
"grad_norm": 9.131397247314453,
|
| 43 |
+
"learning_rate": 6.595744680851064e-06,
|
| 44 |
+
"loss": 4.2309,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8010680907877169,
|
| 49 |
+
"grad_norm": 14.2952880859375,
|
| 50 |
+
"learning_rate": 7.925531914893617e-06,
|
| 51 |
+
"loss": 4.1262,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9345794392523364,
|
| 56 |
+
"grad_norm": 15.645665168762207,
|
| 57 |
+
"learning_rate": 9.255319148936171e-06,
|
| 58 |
+
"loss": 3.9533,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.609181141439206,
|
| 64 |
+
"eval_loss": 0.8993740677833557,
|
| 65 |
+
"eval_macro_f1": 0.5772103729957749,
|
| 66 |
+
"eval_runtime": 0.5542,
|
| 67 |
+
"eval_samples_per_second": 1454.416,
|
| 68 |
+
"eval_steps_per_second": 92.029,
|
| 69 |
+
"eval_weighted_f1": 0.5786168029157498,
|
| 70 |
+
"step": 188
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0640854472630175,
|
| 74 |
+
"grad_norm": 34.74980926513672,
|
| 75 |
+
"learning_rate": 1.0585106382978725e-05,
|
| 76 |
+
"loss": 3.6005,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.1975967957276368,
|
| 81 |
+
"grad_norm": 28.335071563720703,
|
| 82 |
+
"learning_rate": 1.1914893617021277e-05,
|
| 83 |
+
"loss": 3.1783,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.3311081441922563,
|
| 88 |
+
"grad_norm": 26.67999839782715,
|
| 89 |
+
"learning_rate": 1.3244680851063832e-05,
|
| 90 |
+
"loss": 3.0771,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.4646194926568759,
|
| 95 |
+
"grad_norm": 51.67438507080078,
|
| 96 |
+
"learning_rate": 1.4574468085106384e-05,
|
| 97 |
+
"loss": 2.9497,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.5981308411214954,
|
| 102 |
+
"grad_norm": 28.291353225708008,
|
| 103 |
+
"learning_rate": 1.590425531914894e-05,
|
| 104 |
+
"loss": 2.6225,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7316421895861147,
|
| 109 |
+
"grad_norm": 44.70147705078125,
|
| 110 |
+
"learning_rate": 1.723404255319149e-05,
|
| 111 |
+
"loss": 2.4372,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.8651535380507345,
|
| 116 |
+
"grad_norm": 62.695106506347656,
|
| 117 |
+
"learning_rate": 1.8563829787234043e-05,
|
| 118 |
+
"loss": 2.162,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 1.9986648865153538,
|
| 123 |
+
"grad_norm": 24.98824119567871,
|
| 124 |
+
"learning_rate": 1.9893617021276595e-05,
|
| 125 |
+
"loss": 2.2262,
|
| 126 |
+
"step": 375
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"epoch": 2.0,
|
| 130 |
+
"eval_accuracy": 0.7965260545905707,
|
| 131 |
+
"eval_loss": 0.5188497304916382,
|
| 132 |
+
"eval_macro_f1": 0.7948480192243981,
|
| 133 |
+
"eval_runtime": 0.5222,
|
| 134 |
+
"eval_samples_per_second": 1543.488,
|
| 135 |
+
"eval_steps_per_second": 97.665,
|
| 136 |
+
"eval_weighted_f1": 0.7950051743295525,
|
| 137 |
+
"step": 376
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.128170894526035,
|
| 141 |
+
"grad_norm": 53.27727127075195,
|
| 142 |
+
"learning_rate": 1.986406619385343e-05,
|
| 143 |
+
"loss": 1.7151,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.2616822429906542,
|
| 148 |
+
"grad_norm": 87.19738006591797,
|
| 149 |
+
"learning_rate": 1.971631205673759e-05,
|
| 150 |
+
"loss": 1.6824,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.3951935914552736,
|
| 155 |
+
"grad_norm": 41.84999465942383,
|
| 156 |
+
"learning_rate": 1.956855791962175e-05,
|
| 157 |
+
"loss": 1.6127,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.5287049399198933,
|
| 162 |
+
"grad_norm": 30.91926383972168,
|
| 163 |
+
"learning_rate": 1.9420803782505913e-05,
|
| 164 |
+
"loss": 1.7214,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.6622162883845126,
|
| 169 |
+
"grad_norm": 56.13572311401367,
|
| 170 |
+
"learning_rate": 1.927304964539007e-05,
|
| 171 |
+
"loss": 1.6437,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.795727636849132,
|
| 176 |
+
"grad_norm": 52.92278289794922,
|
| 177 |
+
"learning_rate": 1.9125295508274234e-05,
|
| 178 |
+
"loss": 1.6039,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 2.9292389853137517,
|
| 183 |
+
"grad_norm": 43.0755500793457,
|
| 184 |
+
"learning_rate": 1.8977541371158394e-05,
|
| 185 |
+
"loss": 1.6599,
|
| 186 |
+
"step": 550
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 3.0,
|
| 190 |
+
"eval_accuracy": 0.8002481389578163,
|
| 191 |
+
"eval_loss": 0.5185101628303528,
|
| 192 |
+
"eval_macro_f1": 0.7985617142876156,
|
| 193 |
+
"eval_runtime": 0.5251,
|
| 194 |
+
"eval_samples_per_second": 1534.946,
|
| 195 |
+
"eval_steps_per_second": 97.124,
|
| 196 |
+
"eval_weighted_f1": 0.7984130121418866,
|
| 197 |
+
"step": 564
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.0587449933244324,
|
| 201 |
+
"grad_norm": 34.79388427734375,
|
| 202 |
+
"learning_rate": 1.8829787234042557e-05,
|
| 203 |
+
"loss": 1.4772,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.192256341789052,
|
| 208 |
+
"grad_norm": 22.63060188293457,
|
| 209 |
+
"learning_rate": 1.8682033096926714e-05,
|
| 210 |
+
"loss": 1.1397,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.3257676902536715,
|
| 215 |
+
"grad_norm": 37.29880142211914,
|
| 216 |
+
"learning_rate": 1.8534278959810878e-05,
|
| 217 |
+
"loss": 1.0699,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.4592790387182912,
|
| 222 |
+
"grad_norm": 18.859315872192383,
|
| 223 |
+
"learning_rate": 1.8386524822695038e-05,
|
| 224 |
+
"loss": 1.0505,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.5927903871829105,
|
| 229 |
+
"grad_norm": 63.72948455810547,
|
| 230 |
+
"learning_rate": 1.8238770685579198e-05,
|
| 231 |
+
"loss": 1.2013,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.72630173564753,
|
| 236 |
+
"grad_norm": 42.204322814941406,
|
| 237 |
+
"learning_rate": 1.8091016548463358e-05,
|
| 238 |
+
"loss": 1.1122,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.8598130841121496,
|
| 243 |
+
"grad_norm": 70.12254333496094,
|
| 244 |
+
"learning_rate": 1.7943262411347518e-05,
|
| 245 |
+
"loss": 1.1854,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 3.993324432576769,
|
| 250 |
+
"grad_norm": 85.46212768554688,
|
| 251 |
+
"learning_rate": 1.779550827423168e-05,
|
| 252 |
+
"loss": 1.3457,
|
| 253 |
+
"step": 750
|
| 254 |
+
},
|
| 255 |
+
{
|
| 256 |
+
"epoch": 4.0,
|
| 257 |
+
"eval_accuracy": 0.8089330024813896,
|
| 258 |
+
"eval_loss": 0.546148955821991,
|
| 259 |
+
"eval_macro_f1": 0.8094617405060895,
|
| 260 |
+
"eval_runtime": 0.5271,
|
| 261 |
+
"eval_samples_per_second": 1529.124,
|
| 262 |
+
"eval_steps_per_second": 96.756,
|
| 263 |
+
"eval_weighted_f1": 0.8094706960552178,
|
| 264 |
+
"step": 752
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.12283044058745,
|
| 268 |
+
"grad_norm": 27.955347061157227,
|
| 269 |
+
"learning_rate": 1.7647754137115842e-05,
|
| 270 |
+
"loss": 0.757,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.25634178905207,
|
| 275 |
+
"grad_norm": 49.59231185913086,
|
| 276 |
+
"learning_rate": 1.7500000000000002e-05,
|
| 277 |
+
"loss": 0.6709,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.389853137516689,
|
| 282 |
+
"grad_norm": 52.06037521362305,
|
| 283 |
+
"learning_rate": 1.7352245862884162e-05,
|
| 284 |
+
"loss": 0.8391,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.5233644859813085,
|
| 289 |
+
"grad_norm": 54.212562561035156,
|
| 290 |
+
"learning_rate": 1.7204491725768322e-05,
|
| 291 |
+
"loss": 0.7409,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.656875834445928,
|
| 296 |
+
"grad_norm": 38.95555877685547,
|
| 297 |
+
"learning_rate": 1.7056737588652482e-05,
|
| 298 |
+
"loss": 0.6671,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.790387182910547,
|
| 303 |
+
"grad_norm": 34.9852294921875,
|
| 304 |
+
"learning_rate": 1.6908983451536646e-05,
|
| 305 |
+
"loss": 0.7324,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 4.923898531375167,
|
| 310 |
+
"grad_norm": 28.143545150756836,
|
| 311 |
+
"learning_rate": 1.6761229314420806e-05,
|
| 312 |
+
"loss": 0.818,
|
| 313 |
+
"step": 925
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 5.0,
|
| 317 |
+
"eval_accuracy": 0.7977667493796526,
|
| 318 |
+
"eval_loss": 0.739179790019989,
|
| 319 |
+
"eval_macro_f1": 0.790721734403641,
|
| 320 |
+
"eval_runtime": 0.5778,
|
| 321 |
+
"eval_samples_per_second": 1394.872,
|
| 322 |
+
"eval_steps_per_second": 88.261,
|
| 323 |
+
"eval_weighted_f1": 0.791121932272409,
|
| 324 |
+
"step": 940
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.0534045393858475,
|
| 328 |
+
"grad_norm": 59.654903411865234,
|
| 329 |
+
"learning_rate": 1.6613475177304966e-05,
|
| 330 |
+
"loss": 0.9792,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.186915887850467,
|
| 335 |
+
"grad_norm": 42.62851333618164,
|
| 336 |
+
"learning_rate": 1.6465721040189126e-05,
|
| 337 |
+
"loss": 0.7497,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.320427236315087,
|
| 342 |
+
"grad_norm": 123.19102478027344,
|
| 343 |
+
"learning_rate": 1.6317966903073286e-05,
|
| 344 |
+
"loss": 0.5186,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.453938584779706,
|
| 349 |
+
"grad_norm": 66.54837036132812,
|
| 350 |
+
"learning_rate": 1.6170212765957446e-05,
|
| 351 |
+
"loss": 0.4953,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.587449933244326,
|
| 356 |
+
"grad_norm": 16.784555435180664,
|
| 357 |
+
"learning_rate": 1.602245862884161e-05,
|
| 358 |
+
"loss": 0.6363,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.720961281708945,
|
| 363 |
+
"grad_norm": 28.712814331054688,
|
| 364 |
+
"learning_rate": 1.587470449172577e-05,
|
| 365 |
+
"loss": 0.6987,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 5.854472630173564,
|
| 370 |
+
"grad_norm": 3.6170356273651123,
|
| 371 |
+
"learning_rate": 1.572695035460993e-05,
|
| 372 |
+
"loss": 0.5691,
|
| 373 |
+
"step": 1100
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 5.9879839786381845,
|
| 377 |
+
"grad_norm": 59.6114501953125,
|
| 378 |
+
"learning_rate": 1.557919621749409e-05,
|
| 379 |
+
"loss": 0.5618,
|
| 380 |
+
"step": 1125
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 6.0,
|
| 384 |
+
"eval_accuracy": 0.8138957816377171,
|
| 385 |
+
"eval_loss": 0.6991787552833557,
|
| 386 |
+
"eval_macro_f1": 0.8165257465569141,
|
| 387 |
+
"eval_runtime": 0.5681,
|
| 388 |
+
"eval_samples_per_second": 1418.873,
|
| 389 |
+
"eval_steps_per_second": 89.78,
|
| 390 |
+
"eval_weighted_f1": 0.8164177546029194,
|
| 391 |
+
"step": 1128
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 6.117489986648865,
|
| 395 |
+
"grad_norm": 3.2209362983703613,
|
| 396 |
+
"learning_rate": 1.543144208037825e-05,
|
| 397 |
+
"loss": 0.4893,
|
| 398 |
+
"step": 1150
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.251001335113485,
|
| 402 |
+
"grad_norm": 16.82813262939453,
|
| 403 |
+
"learning_rate": 1.5283687943262414e-05,
|
| 404 |
+
"loss": 0.5943,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.384512683578104,
|
| 409 |
+
"grad_norm": 1.2777460813522339,
|
| 410 |
+
"learning_rate": 1.5135933806146574e-05,
|
| 411 |
+
"loss": 0.3216,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.518024032042724,
|
| 416 |
+
"grad_norm": 86.51032257080078,
|
| 417 |
+
"learning_rate": 1.4988179669030734e-05,
|
| 418 |
+
"loss": 0.398,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.651535380507343,
|
| 423 |
+
"grad_norm": 31.18537712097168,
|
| 424 |
+
"learning_rate": 1.4840425531914894e-05,
|
| 425 |
+
"loss": 0.4206,
|
| 426 |
+
"step": 1250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 6.785046728971962,
|
| 430 |
+
"grad_norm": 98.30158233642578,
|
| 431 |
+
"learning_rate": 1.4692671394799056e-05,
|
| 432 |
+
"loss": 0.4636,
|
| 433 |
+
"step": 1275
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 6.9185580774365825,
|
| 437 |
+
"grad_norm": 2.0132503509521484,
|
| 438 |
+
"learning_rate": 1.4544917257683215e-05,
|
| 439 |
+
"loss": 0.3674,
|
| 440 |
+
"step": 1300
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 7.0,
|
| 444 |
+
"eval_accuracy": 0.8424317617866005,
|
| 445 |
+
"eval_loss": 0.7080403566360474,
|
| 446 |
+
"eval_macro_f1": 0.842258582036095,
|
| 447 |
+
"eval_runtime": 0.533,
|
| 448 |
+
"eval_samples_per_second": 1512.286,
|
| 449 |
+
"eval_steps_per_second": 95.691,
|
| 450 |
+
"eval_weighted_f1": 0.8421080192436209,
|
| 451 |
+
"step": 1316
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 7.048064085447263,
|
| 455 |
+
"grad_norm": 1.017230749130249,
|
| 456 |
+
"learning_rate": 1.4397163120567377e-05,
|
| 457 |
+
"loss": 0.3793,
|
| 458 |
+
"step": 1325
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 7.181575433911883,
|
| 462 |
+
"grad_norm": 3.634577989578247,
|
| 463 |
+
"learning_rate": 1.4249408983451538e-05,
|
| 464 |
+
"loss": 0.371,
|
| 465 |
+
"step": 1350
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 7.315086782376502,
|
| 469 |
+
"grad_norm": 4.779361724853516,
|
| 470 |
+
"learning_rate": 1.41016548463357e-05,
|
| 471 |
+
"loss": 0.3175,
|
| 472 |
+
"step": 1375
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 7.4485981308411215,
|
| 476 |
+
"grad_norm": 160.04832458496094,
|
| 477 |
+
"learning_rate": 1.3953900709219859e-05,
|
| 478 |
+
"loss": 0.2508,
|
| 479 |
+
"step": 1400
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 7.582109479305741,
|
| 483 |
+
"grad_norm": 0.7192304134368896,
|
| 484 |
+
"learning_rate": 1.380614657210402e-05,
|
| 485 |
+
"loss": 0.3544,
|
| 486 |
+
"step": 1425
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 7.71562082777036,
|
| 490 |
+
"grad_norm": 146.61080932617188,
|
| 491 |
+
"learning_rate": 1.365839243498818e-05,
|
| 492 |
+
"loss": 0.3275,
|
| 493 |
+
"step": 1450
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 7.8491321762349795,
|
| 497 |
+
"grad_norm": 0.2663750648498535,
|
| 498 |
+
"learning_rate": 1.351063829787234e-05,
|
| 499 |
+
"loss": 0.2122,
|
| 500 |
+
"step": 1475
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 7.9826435246996,
|
| 504 |
+
"grad_norm": 6.357120990753174,
|
| 505 |
+
"learning_rate": 1.3362884160756503e-05,
|
| 506 |
+
"loss": 0.2959,
|
| 507 |
+
"step": 1500
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 8.0,
|
| 511 |
+
"eval_accuracy": 0.8287841191066998,
|
| 512 |
+
"eval_loss": 0.8195311427116394,
|
| 513 |
+
"eval_macro_f1": 0.8293900733192695,
|
| 514 |
+
"eval_runtime": 0.5149,
|
| 515 |
+
"eval_samples_per_second": 1565.217,
|
| 516 |
+
"eval_steps_per_second": 99.04,
|
| 517 |
+
"eval_weighted_f1": 0.8292903291458616,
|
| 518 |
+
"step": 1504
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 8.11214953271028,
|
| 522 |
+
"grad_norm": 722.1763916015625,
|
| 523 |
+
"learning_rate": 1.3215130023640663e-05,
|
| 524 |
+
"loss": 0.2068,
|
| 525 |
+
"step": 1525
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 8.2456608811749,
|
| 529 |
+
"grad_norm": 0.22177685797214508,
|
| 530 |
+
"learning_rate": 1.3067375886524825e-05,
|
| 531 |
+
"loss": 0.1727,
|
| 532 |
+
"step": 1550
|
| 533 |
+
},
|
| 534 |
+
{
|
| 535 |
+
"epoch": 8.379172229639519,
|
| 536 |
+
"grad_norm": 0.9404691457748413,
|
| 537 |
+
"learning_rate": 1.2919621749408983e-05,
|
| 538 |
+
"loss": 0.2006,
|
| 539 |
+
"step": 1575
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 8.51268357810414,
|
| 543 |
+
"grad_norm": 2.124739408493042,
|
| 544 |
+
"learning_rate": 1.2771867612293145e-05,
|
| 545 |
+
"loss": 0.2407,
|
| 546 |
+
"step": 1600
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 8.646194926568759,
|
| 550 |
+
"grad_norm": 237.10240173339844,
|
| 551 |
+
"learning_rate": 1.2624113475177307e-05,
|
| 552 |
+
"loss": 0.1787,
|
| 553 |
+
"step": 1625
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 8.779706275033378,
|
| 557 |
+
"grad_norm": 0.17658063769340515,
|
| 558 |
+
"learning_rate": 1.2476359338061469e-05,
|
| 559 |
+
"loss": 0.1454,
|
| 560 |
+
"step": 1650
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 8.913217623497998,
|
| 564 |
+
"grad_norm": 0.3739471733570099,
|
| 565 |
+
"learning_rate": 1.2328605200945627e-05,
|
| 566 |
+
"loss": 0.1678,
|
| 567 |
+
"step": 1675
|
| 568 |
+
},
|
| 569 |
+
{
|
| 570 |
+
"epoch": 9.0,
|
| 571 |
+
"eval_accuracy": 0.8411910669975186,
|
| 572 |
+
"eval_loss": 0.8197118639945984,
|
| 573 |
+
"eval_macro_f1": 0.8406096099250835,
|
| 574 |
+
"eval_runtime": 0.5694,
|
| 575 |
+
"eval_samples_per_second": 1415.428,
|
| 576 |
+
"eval_steps_per_second": 89.562,
|
| 577 |
+
"eval_weighted_f1": 0.8404130371180706,
|
| 578 |
+
"step": 1692
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 9.042723631508679,
|
| 582 |
+
"grad_norm": 0.17199428379535675,
|
| 583 |
+
"learning_rate": 1.2180851063829789e-05,
|
| 584 |
+
"loss": 0.2372,
|
| 585 |
+
"step": 1700
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 9.176234979973298,
|
| 589 |
+
"grad_norm": 0.16588018834590912,
|
| 590 |
+
"learning_rate": 1.2033096926713949e-05,
|
| 591 |
+
"loss": 0.2007,
|
| 592 |
+
"step": 1725
|
| 593 |
+
},
|
| 594 |
+
{
|
| 595 |
+
"epoch": 9.309746328437917,
|
| 596 |
+
"grad_norm": 0.15143345296382904,
|
| 597 |
+
"learning_rate": 1.1885342789598109e-05,
|
| 598 |
+
"loss": 0.1603,
|
| 599 |
+
"step": 1750
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 9.443257676902537,
|
| 603 |
+
"grad_norm": 0.18354615569114685,
|
| 604 |
+
"learning_rate": 1.1737588652482271e-05,
|
| 605 |
+
"loss": 0.0869,
|
| 606 |
+
"step": 1775
|
| 607 |
+
},
|
| 608 |
+
{
|
| 609 |
+
"epoch": 9.576769025367156,
|
| 610 |
+
"grad_norm": 0.1277248114347458,
|
| 611 |
+
"learning_rate": 1.1589834515366431e-05,
|
| 612 |
+
"loss": 0.1494,
|
| 613 |
+
"step": 1800
|
| 614 |
+
},
|
| 615 |
+
{
|
| 616 |
+
"epoch": 9.710280373831775,
|
| 617 |
+
"grad_norm": 0.754019558429718,
|
| 618 |
+
"learning_rate": 1.1442080378250593e-05,
|
| 619 |
+
"loss": 0.1138,
|
| 620 |
+
"step": 1825
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 9.843791722296395,
|
| 624 |
+
"grad_norm": 0.12734991312026978,
|
| 625 |
+
"learning_rate": 1.1294326241134751e-05,
|
| 626 |
+
"loss": 0.2444,
|
| 627 |
+
"step": 1850
|
| 628 |
+
},
|
| 629 |
+
{
|
| 630 |
+
"epoch": 9.977303070761014,
|
| 631 |
+
"grad_norm": 0.8409662842750549,
|
| 632 |
+
"learning_rate": 1.1146572104018913e-05,
|
| 633 |
+
"loss": 0.262,
|
| 634 |
+
"step": 1875
|
| 635 |
+
},
|
| 636 |
+
{
|
| 637 |
+
"epoch": 10.0,
|
| 638 |
+
"eval_accuracy": 0.8337468982630273,
|
| 639 |
+
"eval_loss": 0.8643537759780884,
|
| 640 |
+
"eval_macro_f1": 0.833263153300989,
|
| 641 |
+
"eval_runtime": 0.5545,
|
| 642 |
+
"eval_samples_per_second": 1453.513,
|
| 643 |
+
"eval_steps_per_second": 91.972,
|
| 644 |
+
"eval_weighted_f1": 0.8330111534916994,
|
| 645 |
+
"step": 1880
|
| 646 |
+
}
|
| 647 |
+
],
|
| 648 |
+
"logging_steps": 25,
|
| 649 |
+
"max_steps": 3760,
|
| 650 |
+
"num_input_tokens_seen": 0,
|
| 651 |
+
"num_train_epochs": 20,
|
| 652 |
+
"save_steps": 500,
|
| 653 |
+
"stateful_callbacks": {
|
| 654 |
+
"EarlyStoppingCallback": {
|
| 655 |
+
"args": {
|
| 656 |
+
"early_stopping_patience": 3,
|
| 657 |
+
"early_stopping_threshold": 0.0
|
| 658 |
+
},
|
| 659 |
+
"attributes": {
|
| 660 |
+
"early_stopping_patience_counter": 3
|
| 661 |
+
}
|
| 662 |
+
},
|
| 663 |
+
"TrainerControl": {
|
| 664 |
+
"args": {
|
| 665 |
+
"should_epoch_stop": false,
|
| 666 |
+
"should_evaluate": false,
|
| 667 |
+
"should_log": false,
|
| 668 |
+
"should_save": true,
|
| 669 |
+
"should_training_stop": true
|
| 670 |
+
},
|
| 671 |
+
"attributes": {}
|
| 672 |
+
}
|
| 673 |
+
},
|
| 674 |
+
"total_flos": 0.0,
|
| 675 |
+
"train_batch_size": 8,
|
| 676 |
+
"trial_name": null,
|
| 677 |
+
"trial_params": null
|
| 678 |
+
}
|
ood-vifactcheck-phobert-topic-domain-04-s44-20260826T111516Z-ab4112e3/trainer/checkpoint-1880/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-05-s44-20260826T111852Z-ab4112e3/trainer/checkpoint-1225/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-topic-domain-05-s44-20260826T111852Z-ab4112e3/trainer/checkpoint-1225/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-05-s44-20260826T111852Z-ab4112e3/trainer/checkpoint-1225/special_tokens_map.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": "<mask>",
|
| 6 |
+
"pad_token": "<pad>",
|
| 7 |
+
"sep_token": "</s>",
|
| 8 |
+
"unk_token": "<unk>"
|
| 9 |
+
}
|
ood-vifactcheck-phobert-topic-domain-05-s44-20260826T111852Z-ab4112e3/trainer/checkpoint-1225/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"64000": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 256,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "PhobertTokenizer",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-phobert-topic-domain-05-s44-20260826T111852Z-ab4112e3/trainer/checkpoint-1225/trainer_state.json
ADDED
|
@@ -0,0 +1,463 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1225,
|
| 3 |
+
"best_metric": 0.8317774830897017,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-topic-domain-05-s44-20260826T111852Z-ab4112e3/trainer/checkpoint-1225",
|
| 5 |
+
"epoch": 7.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1225,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.1430615164520744,
|
| 14 |
+
"grad_norm": 7.753287315368652,
|
| 15 |
+
"learning_rate": 1.3714285714285717e-06,
|
| 16 |
+
"loss": 4.4065,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.2861230329041488,
|
| 21 |
+
"grad_norm": 7.016913890838623,
|
| 22 |
+
"learning_rate": 2.8000000000000003e-06,
|
| 23 |
+
"loss": 4.4101,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.4291845493562232,
|
| 28 |
+
"grad_norm": 8.370100975036621,
|
| 29 |
+
"learning_rate": 4.228571428571429e-06,
|
| 30 |
+
"loss": 4.354,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5722460658082976,
|
| 35 |
+
"grad_norm": 9.792386054992676,
|
| 36 |
+
"learning_rate": 5.6571428571428576e-06,
|
| 37 |
+
"loss": 4.3335,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.7153075822603719,
|
| 42 |
+
"grad_norm": 10.294549942016602,
|
| 43 |
+
"learning_rate": 7.085714285714286e-06,
|
| 44 |
+
"loss": 4.2785,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8583690987124464,
|
| 49 |
+
"grad_norm": 13.513267517089844,
|
| 50 |
+
"learning_rate": 8.514285714285715e-06,
|
| 51 |
+
"loss": 4.1284,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 1.0,
|
| 56 |
+
"grad_norm": 15.271583557128906,
|
| 57 |
+
"learning_rate": 9.942857142857145e-06,
|
| 58 |
+
"loss": 3.9072,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.555256064690027,
|
| 64 |
+
"eval_loss": 0.9577445983886719,
|
| 65 |
+
"eval_macro_f1": 0.5228724614420762,
|
| 66 |
+
"eval_runtime": 0.4848,
|
| 67 |
+
"eval_samples_per_second": 1530.486,
|
| 68 |
+
"eval_steps_per_second": 96.945,
|
| 69 |
+
"eval_weighted_f1": 0.5272990680247308,
|
| 70 |
+
"step": 175
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.1430615164520743,
|
| 74 |
+
"grad_norm": 25.398456573486328,
|
| 75 |
+
"learning_rate": 1.1371428571428573e-05,
|
| 76 |
+
"loss": 3.6241,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.2861230329041489,
|
| 81 |
+
"grad_norm": 28.561784744262695,
|
| 82 |
+
"learning_rate": 1.2800000000000001e-05,
|
| 83 |
+
"loss": 3.4097,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.4291845493562232,
|
| 88 |
+
"grad_norm": 48.97758865356445,
|
| 89 |
+
"learning_rate": 1.422857142857143e-05,
|
| 90 |
+
"loss": 3.255,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.5722460658082977,
|
| 95 |
+
"grad_norm": 24.396240234375,
|
| 96 |
+
"learning_rate": 1.5657142857142856e-05,
|
| 97 |
+
"loss": 2.9284,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.715307582260372,
|
| 102 |
+
"grad_norm": 64.87798309326172,
|
| 103 |
+
"learning_rate": 1.708571428571429e-05,
|
| 104 |
+
"loss": 2.4183,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.8583690987124464,
|
| 109 |
+
"grad_norm": 33.87369918823242,
|
| 110 |
+
"learning_rate": 1.8514285714285716e-05,
|
| 111 |
+
"loss": 2.2112,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 2.0,
|
| 116 |
+
"grad_norm": 58.832252502441406,
|
| 117 |
+
"learning_rate": 1.9942857142857142e-05,
|
| 118 |
+
"loss": 1.918,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.7776280323450134,
|
| 124 |
+
"eval_loss": 0.5735662579536438,
|
| 125 |
+
"eval_macro_f1": 0.7783146797451375,
|
| 126 |
+
"eval_runtime": 0.4708,
|
| 127 |
+
"eval_samples_per_second": 1575.947,
|
| 128 |
+
"eval_steps_per_second": 99.824,
|
| 129 |
+
"eval_weighted_f1": 0.778645436466508,
|
| 130 |
+
"step": 350
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.1430615164520743,
|
| 134 |
+
"grad_norm": 57.35730743408203,
|
| 135 |
+
"learning_rate": 1.984761904761905e-05,
|
| 136 |
+
"loss": 1.5529,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.2861230329041486,
|
| 141 |
+
"grad_norm": 33.88847351074219,
|
| 142 |
+
"learning_rate": 1.968888888888889e-05,
|
| 143 |
+
"loss": 1.7985,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.429184549356223,
|
| 148 |
+
"grad_norm": 31.695287704467773,
|
| 149 |
+
"learning_rate": 1.953015873015873e-05,
|
| 150 |
+
"loss": 1.5781,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.5722460658082977,
|
| 155 |
+
"grad_norm": 39.669036865234375,
|
| 156 |
+
"learning_rate": 1.937142857142857e-05,
|
| 157 |
+
"loss": 1.783,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.715307582260372,
|
| 162 |
+
"grad_norm": 40.711124420166016,
|
| 163 |
+
"learning_rate": 1.9212698412698414e-05,
|
| 164 |
+
"loss": 1.6413,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.8583690987124464,
|
| 169 |
+
"grad_norm": 34.58282470703125,
|
| 170 |
+
"learning_rate": 1.9053968253968257e-05,
|
| 171 |
+
"loss": 1.5758,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 3.0,
|
| 176 |
+
"grad_norm": 29.36798858642578,
|
| 177 |
+
"learning_rate": 1.8895238095238096e-05,
|
| 178 |
+
"loss": 1.5715,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 3.0,
|
| 183 |
+
"eval_accuracy": 0.8099730458221024,
|
| 184 |
+
"eval_loss": 0.5089641213417053,
|
| 185 |
+
"eval_macro_f1": 0.8038772377128716,
|
| 186 |
+
"eval_runtime": 0.4959,
|
| 187 |
+
"eval_samples_per_second": 1496.234,
|
| 188 |
+
"eval_steps_per_second": 94.775,
|
| 189 |
+
"eval_weighted_f1": 0.8043846279043007,
|
| 190 |
+
"step": 525
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.1430615164520743,
|
| 194 |
+
"grad_norm": 34.5817985534668,
|
| 195 |
+
"learning_rate": 1.8736507936507936e-05,
|
| 196 |
+
"loss": 1.2702,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.2861230329041486,
|
| 201 |
+
"grad_norm": 29.863290786743164,
|
| 202 |
+
"learning_rate": 1.857777777777778e-05,
|
| 203 |
+
"loss": 1.1181,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.429184549356223,
|
| 208 |
+
"grad_norm": 41.16310119628906,
|
| 209 |
+
"learning_rate": 1.8419047619047622e-05,
|
| 210 |
+
"loss": 1.0315,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.5722460658082977,
|
| 215 |
+
"grad_norm": 69.25782775878906,
|
| 216 |
+
"learning_rate": 1.8260317460317462e-05,
|
| 217 |
+
"loss": 1.0583,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.715307582260372,
|
| 222 |
+
"grad_norm": 33.97926330566406,
|
| 223 |
+
"learning_rate": 1.81015873015873e-05,
|
| 224 |
+
"loss": 0.9302,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.8583690987124464,
|
| 229 |
+
"grad_norm": 73.1862564086914,
|
| 230 |
+
"learning_rate": 1.7942857142857144e-05,
|
| 231 |
+
"loss": 1.0212,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 4.0,
|
| 236 |
+
"grad_norm": 53.455780029296875,
|
| 237 |
+
"learning_rate": 1.7784126984126987e-05,
|
| 238 |
+
"loss": 0.9405,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 4.0,
|
| 243 |
+
"eval_accuracy": 0.8288409703504043,
|
| 244 |
+
"eval_loss": 0.5120916962623596,
|
| 245 |
+
"eval_macro_f1": 0.8276280734932193,
|
| 246 |
+
"eval_runtime": 0.4936,
|
| 247 |
+
"eval_samples_per_second": 1503.147,
|
| 248 |
+
"eval_steps_per_second": 95.213,
|
| 249 |
+
"eval_weighted_f1": 0.8279505564149109,
|
| 250 |
+
"step": 700
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 4.143061516452074,
|
| 254 |
+
"grad_norm": 63.39274597167969,
|
| 255 |
+
"learning_rate": 1.7625396825396827e-05,
|
| 256 |
+
"loss": 0.6415,
|
| 257 |
+
"step": 725
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.286123032904149,
|
| 261 |
+
"grad_norm": 27.531635284423828,
|
| 262 |
+
"learning_rate": 1.7466666666666667e-05,
|
| 263 |
+
"loss": 0.9117,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.429184549356223,
|
| 268 |
+
"grad_norm": 37.4041633605957,
|
| 269 |
+
"learning_rate": 1.730793650793651e-05,
|
| 270 |
+
"loss": 0.6562,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.572246065808297,
|
| 275 |
+
"grad_norm": 20.748502731323242,
|
| 276 |
+
"learning_rate": 1.7149206349206353e-05,
|
| 277 |
+
"loss": 0.5378,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.715307582260372,
|
| 282 |
+
"grad_norm": 78.40021514892578,
|
| 283 |
+
"learning_rate": 1.6990476190476192e-05,
|
| 284 |
+
"loss": 0.8271,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.858369098712446,
|
| 289 |
+
"grad_norm": 91.58229064941406,
|
| 290 |
+
"learning_rate": 1.6831746031746032e-05,
|
| 291 |
+
"loss": 0.7786,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 5.0,
|
| 296 |
+
"grad_norm": 107.20304107666016,
|
| 297 |
+
"learning_rate": 1.6673015873015875e-05,
|
| 298 |
+
"loss": 0.79,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 5.0,
|
| 303 |
+
"eval_accuracy": 0.8288409703504043,
|
| 304 |
+
"eval_loss": 0.5741622447967529,
|
| 305 |
+
"eval_macro_f1": 0.8280824009247748,
|
| 306 |
+
"eval_runtime": 0.4978,
|
| 307 |
+
"eval_samples_per_second": 1490.575,
|
| 308 |
+
"eval_steps_per_second": 94.416,
|
| 309 |
+
"eval_weighted_f1": 0.828327014689436,
|
| 310 |
+
"step": 875
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 5.143061516452074,
|
| 314 |
+
"grad_norm": 27.98452377319336,
|
| 315 |
+
"learning_rate": 1.6514285714285714e-05,
|
| 316 |
+
"loss": 0.5552,
|
| 317 |
+
"step": 900
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.286123032904149,
|
| 321 |
+
"grad_norm": 54.50994110107422,
|
| 322 |
+
"learning_rate": 1.6355555555555557e-05,
|
| 323 |
+
"loss": 0.6091,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.429184549356223,
|
| 328 |
+
"grad_norm": 40.632144927978516,
|
| 329 |
+
"learning_rate": 1.6196825396825397e-05,
|
| 330 |
+
"loss": 0.4472,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.572246065808297,
|
| 335 |
+
"grad_norm": 58.21255111694336,
|
| 336 |
+
"learning_rate": 1.603809523809524e-05,
|
| 337 |
+
"loss": 0.4985,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.715307582260372,
|
| 342 |
+
"grad_norm": 1.093936562538147,
|
| 343 |
+
"learning_rate": 1.587936507936508e-05,
|
| 344 |
+
"loss": 0.654,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.858369098712446,
|
| 349 |
+
"grad_norm": 4.856874942779541,
|
| 350 |
+
"learning_rate": 1.5720634920634923e-05,
|
| 351 |
+
"loss": 0.6271,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 6.0,
|
| 356 |
+
"grad_norm": 53.213443756103516,
|
| 357 |
+
"learning_rate": 1.5561904761904762e-05,
|
| 358 |
+
"loss": 0.5494,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 6.0,
|
| 363 |
+
"eval_accuracy": 0.8247978436657682,
|
| 364 |
+
"eval_loss": 0.7360907196998596,
|
| 365 |
+
"eval_macro_f1": 0.823881199342193,
|
| 366 |
+
"eval_runtime": 0.526,
|
| 367 |
+
"eval_samples_per_second": 1410.676,
|
| 368 |
+
"eval_steps_per_second": 89.356,
|
| 369 |
+
"eval_weighted_f1": 0.8240906889006097,
|
| 370 |
+
"step": 1050
|
| 371 |
+
},
|
| 372 |
+
{
|
| 373 |
+
"epoch": 6.143061516452074,
|
| 374 |
+
"grad_norm": 42.2149658203125,
|
| 375 |
+
"learning_rate": 1.5403174603174605e-05,
|
| 376 |
+
"loss": 0.3972,
|
| 377 |
+
"step": 1075
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 6.286123032904149,
|
| 381 |
+
"grad_norm": 78.92060852050781,
|
| 382 |
+
"learning_rate": 1.5244444444444447e-05,
|
| 383 |
+
"loss": 0.4955,
|
| 384 |
+
"step": 1100
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 6.429184549356223,
|
| 388 |
+
"grad_norm": 1.7310129404067993,
|
| 389 |
+
"learning_rate": 1.5085714285714288e-05,
|
| 390 |
+
"loss": 0.4562,
|
| 391 |
+
"step": 1125
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 6.572246065808297,
|
| 395 |
+
"grad_norm": 3.1209323406219482,
|
| 396 |
+
"learning_rate": 1.4926984126984128e-05,
|
| 397 |
+
"loss": 0.2537,
|
| 398 |
+
"step": 1150
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.715307582260372,
|
| 402 |
+
"grad_norm": 88.41880798339844,
|
| 403 |
+
"learning_rate": 1.4768253968253969e-05,
|
| 404 |
+
"loss": 0.4628,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.858369098712446,
|
| 409 |
+
"grad_norm": 80.18390655517578,
|
| 410 |
+
"learning_rate": 1.4609523809523812e-05,
|
| 411 |
+
"loss": 0.4326,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 7.0,
|
| 416 |
+
"grad_norm": 59.84803009033203,
|
| 417 |
+
"learning_rate": 1.4450793650793653e-05,
|
| 418 |
+
"loss": 0.4724,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 7.0,
|
| 423 |
+
"eval_accuracy": 0.8315363881401617,
|
| 424 |
+
"eval_loss": 0.7438673377037048,
|
| 425 |
+
"eval_macro_f1": 0.8317774830897017,
|
| 426 |
+
"eval_runtime": 0.497,
|
| 427 |
+
"eval_samples_per_second": 1493.015,
|
| 428 |
+
"eval_steps_per_second": 94.571,
|
| 429 |
+
"eval_weighted_f1": 0.8319657318849818,
|
| 430 |
+
"step": 1225
|
| 431 |
+
}
|
| 432 |
+
],
|
| 433 |
+
"logging_steps": 25,
|
| 434 |
+
"max_steps": 3500,
|
| 435 |
+
"num_input_tokens_seen": 0,
|
| 436 |
+
"num_train_epochs": 20,
|
| 437 |
+
"save_steps": 500,
|
| 438 |
+
"stateful_callbacks": {
|
| 439 |
+
"EarlyStoppingCallback": {
|
| 440 |
+
"args": {
|
| 441 |
+
"early_stopping_patience": 3,
|
| 442 |
+
"early_stopping_threshold": 0.0
|
| 443 |
+
},
|
| 444 |
+
"attributes": {
|
| 445 |
+
"early_stopping_patience_counter": 0
|
| 446 |
+
}
|
| 447 |
+
},
|
| 448 |
+
"TrainerControl": {
|
| 449 |
+
"args": {
|
| 450 |
+
"should_epoch_stop": false,
|
| 451 |
+
"should_evaluate": false,
|
| 452 |
+
"should_log": false,
|
| 453 |
+
"should_save": true,
|
| 454 |
+
"should_training_stop": false
|
| 455 |
+
},
|
| 456 |
+
"attributes": {}
|
| 457 |
+
}
|
| 458 |
+
},
|
| 459 |
+
"total_flos": 0.0,
|
| 460 |
+
"train_batch_size": 8,
|
| 461 |
+
"trial_name": null,
|
| 462 |
+
"trial_params": null
|
| 463 |
+
}
|
ood-vifactcheck-phobert-topic-domain-05-s44-20260826T111852Z-ab4112e3/trainer/checkpoint-1225/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-05-s44-20260826T111852Z-ab4112e3/trainer/checkpoint-1750/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-topic-domain-05-s44-20260826T111852Z-ab4112e3/trainer/checkpoint-1750/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-05-s44-20260826T111852Z-ab4112e3/trainer/checkpoint-1750/special_tokens_map.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": "<mask>",
|
| 6 |
+
"pad_token": "<pad>",
|
| 7 |
+
"sep_token": "</s>",
|
| 8 |
+
"unk_token": "<unk>"
|
| 9 |
+
}
|
ood-vifactcheck-phobert-topic-domain-05-s44-20260826T111852Z-ab4112e3/trainer/checkpoint-1750/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"64000": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 256,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "PhobertTokenizer",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-phobert-topic-domain-05-s44-20260826T111852Z-ab4112e3/trainer/checkpoint-1750/trainer_state.json
ADDED
|
@@ -0,0 +1,643 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1225,
|
| 3 |
+
"best_metric": 0.8317774830897017,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-topic-domain-05-s44-20260826T111852Z-ab4112e3/trainer/checkpoint-1225",
|
| 5 |
+
"epoch": 10.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1750,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.1430615164520744,
|
| 14 |
+
"grad_norm": 7.753287315368652,
|
| 15 |
+
"learning_rate": 1.3714285714285717e-06,
|
| 16 |
+
"loss": 4.4065,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.2861230329041488,
|
| 21 |
+
"grad_norm": 7.016913890838623,
|
| 22 |
+
"learning_rate": 2.8000000000000003e-06,
|
| 23 |
+
"loss": 4.4101,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.4291845493562232,
|
| 28 |
+
"grad_norm": 8.370100975036621,
|
| 29 |
+
"learning_rate": 4.228571428571429e-06,
|
| 30 |
+
"loss": 4.354,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5722460658082976,
|
| 35 |
+
"grad_norm": 9.792386054992676,
|
| 36 |
+
"learning_rate": 5.6571428571428576e-06,
|
| 37 |
+
"loss": 4.3335,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.7153075822603719,
|
| 42 |
+
"grad_norm": 10.294549942016602,
|
| 43 |
+
"learning_rate": 7.085714285714286e-06,
|
| 44 |
+
"loss": 4.2785,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8583690987124464,
|
| 49 |
+
"grad_norm": 13.513267517089844,
|
| 50 |
+
"learning_rate": 8.514285714285715e-06,
|
| 51 |
+
"loss": 4.1284,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 1.0,
|
| 56 |
+
"grad_norm": 15.271583557128906,
|
| 57 |
+
"learning_rate": 9.942857142857145e-06,
|
| 58 |
+
"loss": 3.9072,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.555256064690027,
|
| 64 |
+
"eval_loss": 0.9577445983886719,
|
| 65 |
+
"eval_macro_f1": 0.5228724614420762,
|
| 66 |
+
"eval_runtime": 0.4848,
|
| 67 |
+
"eval_samples_per_second": 1530.486,
|
| 68 |
+
"eval_steps_per_second": 96.945,
|
| 69 |
+
"eval_weighted_f1": 0.5272990680247308,
|
| 70 |
+
"step": 175
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.1430615164520743,
|
| 74 |
+
"grad_norm": 25.398456573486328,
|
| 75 |
+
"learning_rate": 1.1371428571428573e-05,
|
| 76 |
+
"loss": 3.6241,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.2861230329041489,
|
| 81 |
+
"grad_norm": 28.561784744262695,
|
| 82 |
+
"learning_rate": 1.2800000000000001e-05,
|
| 83 |
+
"loss": 3.4097,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.4291845493562232,
|
| 88 |
+
"grad_norm": 48.97758865356445,
|
| 89 |
+
"learning_rate": 1.422857142857143e-05,
|
| 90 |
+
"loss": 3.255,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.5722460658082977,
|
| 95 |
+
"grad_norm": 24.396240234375,
|
| 96 |
+
"learning_rate": 1.5657142857142856e-05,
|
| 97 |
+
"loss": 2.9284,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.715307582260372,
|
| 102 |
+
"grad_norm": 64.87798309326172,
|
| 103 |
+
"learning_rate": 1.708571428571429e-05,
|
| 104 |
+
"loss": 2.4183,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.8583690987124464,
|
| 109 |
+
"grad_norm": 33.87369918823242,
|
| 110 |
+
"learning_rate": 1.8514285714285716e-05,
|
| 111 |
+
"loss": 2.2112,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 2.0,
|
| 116 |
+
"grad_norm": 58.832252502441406,
|
| 117 |
+
"learning_rate": 1.9942857142857142e-05,
|
| 118 |
+
"loss": 1.918,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.7776280323450134,
|
| 124 |
+
"eval_loss": 0.5735662579536438,
|
| 125 |
+
"eval_macro_f1": 0.7783146797451375,
|
| 126 |
+
"eval_runtime": 0.4708,
|
| 127 |
+
"eval_samples_per_second": 1575.947,
|
| 128 |
+
"eval_steps_per_second": 99.824,
|
| 129 |
+
"eval_weighted_f1": 0.778645436466508,
|
| 130 |
+
"step": 350
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.1430615164520743,
|
| 134 |
+
"grad_norm": 57.35730743408203,
|
| 135 |
+
"learning_rate": 1.984761904761905e-05,
|
| 136 |
+
"loss": 1.5529,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.2861230329041486,
|
| 141 |
+
"grad_norm": 33.88847351074219,
|
| 142 |
+
"learning_rate": 1.968888888888889e-05,
|
| 143 |
+
"loss": 1.7985,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.429184549356223,
|
| 148 |
+
"grad_norm": 31.695287704467773,
|
| 149 |
+
"learning_rate": 1.953015873015873e-05,
|
| 150 |
+
"loss": 1.5781,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.5722460658082977,
|
| 155 |
+
"grad_norm": 39.669036865234375,
|
| 156 |
+
"learning_rate": 1.937142857142857e-05,
|
| 157 |
+
"loss": 1.783,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.715307582260372,
|
| 162 |
+
"grad_norm": 40.711124420166016,
|
| 163 |
+
"learning_rate": 1.9212698412698414e-05,
|
| 164 |
+
"loss": 1.6413,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.8583690987124464,
|
| 169 |
+
"grad_norm": 34.58282470703125,
|
| 170 |
+
"learning_rate": 1.9053968253968257e-05,
|
| 171 |
+
"loss": 1.5758,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 3.0,
|
| 176 |
+
"grad_norm": 29.36798858642578,
|
| 177 |
+
"learning_rate": 1.8895238095238096e-05,
|
| 178 |
+
"loss": 1.5715,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 3.0,
|
| 183 |
+
"eval_accuracy": 0.8099730458221024,
|
| 184 |
+
"eval_loss": 0.5089641213417053,
|
| 185 |
+
"eval_macro_f1": 0.8038772377128716,
|
| 186 |
+
"eval_runtime": 0.4959,
|
| 187 |
+
"eval_samples_per_second": 1496.234,
|
| 188 |
+
"eval_steps_per_second": 94.775,
|
| 189 |
+
"eval_weighted_f1": 0.8043846279043007,
|
| 190 |
+
"step": 525
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.1430615164520743,
|
| 194 |
+
"grad_norm": 34.5817985534668,
|
| 195 |
+
"learning_rate": 1.8736507936507936e-05,
|
| 196 |
+
"loss": 1.2702,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.2861230329041486,
|
| 201 |
+
"grad_norm": 29.863290786743164,
|
| 202 |
+
"learning_rate": 1.857777777777778e-05,
|
| 203 |
+
"loss": 1.1181,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.429184549356223,
|
| 208 |
+
"grad_norm": 41.16310119628906,
|
| 209 |
+
"learning_rate": 1.8419047619047622e-05,
|
| 210 |
+
"loss": 1.0315,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.5722460658082977,
|
| 215 |
+
"grad_norm": 69.25782775878906,
|
| 216 |
+
"learning_rate": 1.8260317460317462e-05,
|
| 217 |
+
"loss": 1.0583,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.715307582260372,
|
| 222 |
+
"grad_norm": 33.97926330566406,
|
| 223 |
+
"learning_rate": 1.81015873015873e-05,
|
| 224 |
+
"loss": 0.9302,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.8583690987124464,
|
| 229 |
+
"grad_norm": 73.1862564086914,
|
| 230 |
+
"learning_rate": 1.7942857142857144e-05,
|
| 231 |
+
"loss": 1.0212,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 4.0,
|
| 236 |
+
"grad_norm": 53.455780029296875,
|
| 237 |
+
"learning_rate": 1.7784126984126987e-05,
|
| 238 |
+
"loss": 0.9405,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 4.0,
|
| 243 |
+
"eval_accuracy": 0.8288409703504043,
|
| 244 |
+
"eval_loss": 0.5120916962623596,
|
| 245 |
+
"eval_macro_f1": 0.8276280734932193,
|
| 246 |
+
"eval_runtime": 0.4936,
|
| 247 |
+
"eval_samples_per_second": 1503.147,
|
| 248 |
+
"eval_steps_per_second": 95.213,
|
| 249 |
+
"eval_weighted_f1": 0.8279505564149109,
|
| 250 |
+
"step": 700
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 4.143061516452074,
|
| 254 |
+
"grad_norm": 63.39274597167969,
|
| 255 |
+
"learning_rate": 1.7625396825396827e-05,
|
| 256 |
+
"loss": 0.6415,
|
| 257 |
+
"step": 725
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.286123032904149,
|
| 261 |
+
"grad_norm": 27.531635284423828,
|
| 262 |
+
"learning_rate": 1.7466666666666667e-05,
|
| 263 |
+
"loss": 0.9117,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.429184549356223,
|
| 268 |
+
"grad_norm": 37.4041633605957,
|
| 269 |
+
"learning_rate": 1.730793650793651e-05,
|
| 270 |
+
"loss": 0.6562,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.572246065808297,
|
| 275 |
+
"grad_norm": 20.748502731323242,
|
| 276 |
+
"learning_rate": 1.7149206349206353e-05,
|
| 277 |
+
"loss": 0.5378,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.715307582260372,
|
| 282 |
+
"grad_norm": 78.40021514892578,
|
| 283 |
+
"learning_rate": 1.6990476190476192e-05,
|
| 284 |
+
"loss": 0.8271,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.858369098712446,
|
| 289 |
+
"grad_norm": 91.58229064941406,
|
| 290 |
+
"learning_rate": 1.6831746031746032e-05,
|
| 291 |
+
"loss": 0.7786,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 5.0,
|
| 296 |
+
"grad_norm": 107.20304107666016,
|
| 297 |
+
"learning_rate": 1.6673015873015875e-05,
|
| 298 |
+
"loss": 0.79,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 5.0,
|
| 303 |
+
"eval_accuracy": 0.8288409703504043,
|
| 304 |
+
"eval_loss": 0.5741622447967529,
|
| 305 |
+
"eval_macro_f1": 0.8280824009247748,
|
| 306 |
+
"eval_runtime": 0.4978,
|
| 307 |
+
"eval_samples_per_second": 1490.575,
|
| 308 |
+
"eval_steps_per_second": 94.416,
|
| 309 |
+
"eval_weighted_f1": 0.828327014689436,
|
| 310 |
+
"step": 875
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 5.143061516452074,
|
| 314 |
+
"grad_norm": 27.98452377319336,
|
| 315 |
+
"learning_rate": 1.6514285714285714e-05,
|
| 316 |
+
"loss": 0.5552,
|
| 317 |
+
"step": 900
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.286123032904149,
|
| 321 |
+
"grad_norm": 54.50994110107422,
|
| 322 |
+
"learning_rate": 1.6355555555555557e-05,
|
| 323 |
+
"loss": 0.6091,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.429184549356223,
|
| 328 |
+
"grad_norm": 40.632144927978516,
|
| 329 |
+
"learning_rate": 1.6196825396825397e-05,
|
| 330 |
+
"loss": 0.4472,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.572246065808297,
|
| 335 |
+
"grad_norm": 58.21255111694336,
|
| 336 |
+
"learning_rate": 1.603809523809524e-05,
|
| 337 |
+
"loss": 0.4985,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.715307582260372,
|
| 342 |
+
"grad_norm": 1.093936562538147,
|
| 343 |
+
"learning_rate": 1.587936507936508e-05,
|
| 344 |
+
"loss": 0.654,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.858369098712446,
|
| 349 |
+
"grad_norm": 4.856874942779541,
|
| 350 |
+
"learning_rate": 1.5720634920634923e-05,
|
| 351 |
+
"loss": 0.6271,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 6.0,
|
| 356 |
+
"grad_norm": 53.213443756103516,
|
| 357 |
+
"learning_rate": 1.5561904761904762e-05,
|
| 358 |
+
"loss": 0.5494,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 6.0,
|
| 363 |
+
"eval_accuracy": 0.8247978436657682,
|
| 364 |
+
"eval_loss": 0.7360907196998596,
|
| 365 |
+
"eval_macro_f1": 0.823881199342193,
|
| 366 |
+
"eval_runtime": 0.526,
|
| 367 |
+
"eval_samples_per_second": 1410.676,
|
| 368 |
+
"eval_steps_per_second": 89.356,
|
| 369 |
+
"eval_weighted_f1": 0.8240906889006097,
|
| 370 |
+
"step": 1050
|
| 371 |
+
},
|
| 372 |
+
{
|
| 373 |
+
"epoch": 6.143061516452074,
|
| 374 |
+
"grad_norm": 42.2149658203125,
|
| 375 |
+
"learning_rate": 1.5403174603174605e-05,
|
| 376 |
+
"loss": 0.3972,
|
| 377 |
+
"step": 1075
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 6.286123032904149,
|
| 381 |
+
"grad_norm": 78.92060852050781,
|
| 382 |
+
"learning_rate": 1.5244444444444447e-05,
|
| 383 |
+
"loss": 0.4955,
|
| 384 |
+
"step": 1100
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 6.429184549356223,
|
| 388 |
+
"grad_norm": 1.7310129404067993,
|
| 389 |
+
"learning_rate": 1.5085714285714288e-05,
|
| 390 |
+
"loss": 0.4562,
|
| 391 |
+
"step": 1125
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 6.572246065808297,
|
| 395 |
+
"grad_norm": 3.1209323406219482,
|
| 396 |
+
"learning_rate": 1.4926984126984128e-05,
|
| 397 |
+
"loss": 0.2537,
|
| 398 |
+
"step": 1150
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.715307582260372,
|
| 402 |
+
"grad_norm": 88.41880798339844,
|
| 403 |
+
"learning_rate": 1.4768253968253969e-05,
|
| 404 |
+
"loss": 0.4628,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.858369098712446,
|
| 409 |
+
"grad_norm": 80.18390655517578,
|
| 410 |
+
"learning_rate": 1.4609523809523812e-05,
|
| 411 |
+
"loss": 0.4326,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 7.0,
|
| 416 |
+
"grad_norm": 59.84803009033203,
|
| 417 |
+
"learning_rate": 1.4450793650793653e-05,
|
| 418 |
+
"loss": 0.4724,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 7.0,
|
| 423 |
+
"eval_accuracy": 0.8315363881401617,
|
| 424 |
+
"eval_loss": 0.7438673377037048,
|
| 425 |
+
"eval_macro_f1": 0.8317774830897017,
|
| 426 |
+
"eval_runtime": 0.497,
|
| 427 |
+
"eval_samples_per_second": 1493.015,
|
| 428 |
+
"eval_steps_per_second": 94.571,
|
| 429 |
+
"eval_weighted_f1": 0.8319657318849818,
|
| 430 |
+
"step": 1225
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 7.143061516452074,
|
| 434 |
+
"grad_norm": 2.8599753379821777,
|
| 435 |
+
"learning_rate": 1.4292063492063493e-05,
|
| 436 |
+
"loss": 0.2574,
|
| 437 |
+
"step": 1250
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 7.286123032904149,
|
| 441 |
+
"grad_norm": 137.70645141601562,
|
| 442 |
+
"learning_rate": 1.4133333333333334e-05,
|
| 443 |
+
"loss": 0.3395,
|
| 444 |
+
"step": 1275
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 7.429184549356223,
|
| 448 |
+
"grad_norm": 72.82587432861328,
|
| 449 |
+
"learning_rate": 1.3974603174603177e-05,
|
| 450 |
+
"loss": 0.3047,
|
| 451 |
+
"step": 1300
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 7.572246065808297,
|
| 455 |
+
"grad_norm": 7.936881065368652,
|
| 456 |
+
"learning_rate": 1.3815873015873018e-05,
|
| 457 |
+
"loss": 0.3111,
|
| 458 |
+
"step": 1325
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 7.715307582260372,
|
| 462 |
+
"grad_norm": 0.33879590034484863,
|
| 463 |
+
"learning_rate": 1.3657142857142858e-05,
|
| 464 |
+
"loss": 0.279,
|
| 465 |
+
"step": 1350
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 7.858369098712446,
|
| 469 |
+
"grad_norm": 0.24516333639621735,
|
| 470 |
+
"learning_rate": 1.34984126984127e-05,
|
| 471 |
+
"loss": 0.2971,
|
| 472 |
+
"step": 1375
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 8.0,
|
| 476 |
+
"grad_norm": 73.0232162475586,
|
| 477 |
+
"learning_rate": 1.333968253968254e-05,
|
| 478 |
+
"loss": 0.2414,
|
| 479 |
+
"step": 1400
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 8.0,
|
| 483 |
+
"eval_accuracy": 0.8301886792452831,
|
| 484 |
+
"eval_loss": 0.8327235579490662,
|
| 485 |
+
"eval_macro_f1": 0.828648354681861,
|
| 486 |
+
"eval_runtime": 0.4941,
|
| 487 |
+
"eval_samples_per_second": 1501.829,
|
| 488 |
+
"eval_steps_per_second": 95.129,
|
| 489 |
+
"eval_weighted_f1": 0.8289762272185157,
|
| 490 |
+
"step": 1400
|
| 491 |
+
},
|
| 492 |
+
{
|
| 493 |
+
"epoch": 8.143061516452075,
|
| 494 |
+
"grad_norm": 35.001617431640625,
|
| 495 |
+
"learning_rate": 1.318095238095238e-05,
|
| 496 |
+
"loss": 0.065,
|
| 497 |
+
"step": 1425
|
| 498 |
+
},
|
| 499 |
+
{
|
| 500 |
+
"epoch": 8.286123032904149,
|
| 501 |
+
"grad_norm": 0.2105819284915924,
|
| 502 |
+
"learning_rate": 1.3022222222222223e-05,
|
| 503 |
+
"loss": 0.1762,
|
| 504 |
+
"step": 1450
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 8.429184549356224,
|
| 508 |
+
"grad_norm": 0.47483184933662415,
|
| 509 |
+
"learning_rate": 1.2863492063492065e-05,
|
| 510 |
+
"loss": 0.3143,
|
| 511 |
+
"step": 1475
|
| 512 |
+
},
|
| 513 |
+
{
|
| 514 |
+
"epoch": 8.572246065808297,
|
| 515 |
+
"grad_norm": 6.644261360168457,
|
| 516 |
+
"learning_rate": 1.2704761904761906e-05,
|
| 517 |
+
"loss": 0.2549,
|
| 518 |
+
"step": 1500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 8.715307582260372,
|
| 522 |
+
"grad_norm": 0.18914642930030823,
|
| 523 |
+
"learning_rate": 1.2546031746031746e-05,
|
| 524 |
+
"loss": 0.2006,
|
| 525 |
+
"step": 1525
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 8.858369098712446,
|
| 529 |
+
"grad_norm": 265.77154541015625,
|
| 530 |
+
"learning_rate": 1.2387301587301589e-05,
|
| 531 |
+
"loss": 0.3159,
|
| 532 |
+
"step": 1550
|
| 533 |
+
},
|
| 534 |
+
{
|
| 535 |
+
"epoch": 9.0,
|
| 536 |
+
"grad_norm": 0.12892097234725952,
|
| 537 |
+
"learning_rate": 1.222857142857143e-05,
|
| 538 |
+
"loss": 0.2356,
|
| 539 |
+
"step": 1575
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 9.0,
|
| 543 |
+
"eval_accuracy": 0.8153638814016172,
|
| 544 |
+
"eval_loss": 0.972222089767456,
|
| 545 |
+
"eval_macro_f1": 0.8141311321158206,
|
| 546 |
+
"eval_runtime": 0.5328,
|
| 547 |
+
"eval_samples_per_second": 1392.664,
|
| 548 |
+
"eval_steps_per_second": 88.215,
|
| 549 |
+
"eval_weighted_f1": 0.8141618202906478,
|
| 550 |
+
"step": 1575
|
| 551 |
+
},
|
| 552 |
+
{
|
| 553 |
+
"epoch": 9.143061516452075,
|
| 554 |
+
"grad_norm": 0.3004032075405121,
|
| 555 |
+
"learning_rate": 1.2069841269841271e-05,
|
| 556 |
+
"loss": 0.1211,
|
| 557 |
+
"step": 1600
|
| 558 |
+
},
|
| 559 |
+
{
|
| 560 |
+
"epoch": 9.286123032904149,
|
| 561 |
+
"grad_norm": 0.409688800573349,
|
| 562 |
+
"learning_rate": 1.191111111111111e-05,
|
| 563 |
+
"loss": 0.2187,
|
| 564 |
+
"step": 1625
|
| 565 |
+
},
|
| 566 |
+
{
|
| 567 |
+
"epoch": 9.429184549356224,
|
| 568 |
+
"grad_norm": 0.24715174734592438,
|
| 569 |
+
"learning_rate": 1.1752380952380952e-05,
|
| 570 |
+
"loss": 0.2348,
|
| 571 |
+
"step": 1650
|
| 572 |
+
},
|
| 573 |
+
{
|
| 574 |
+
"epoch": 9.572246065808297,
|
| 575 |
+
"grad_norm": 0.18835674226284027,
|
| 576 |
+
"learning_rate": 1.1593650793650795e-05,
|
| 577 |
+
"loss": 0.1877,
|
| 578 |
+
"step": 1675
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 9.715307582260372,
|
| 582 |
+
"grad_norm": 14.859042167663574,
|
| 583 |
+
"learning_rate": 1.1434920634920636e-05,
|
| 584 |
+
"loss": 0.1498,
|
| 585 |
+
"step": 1700
|
| 586 |
+
},
|
| 587 |
+
{
|
| 588 |
+
"epoch": 9.858369098712446,
|
| 589 |
+
"grad_norm": 13.43192195892334,
|
| 590 |
+
"learning_rate": 1.1276190476190476e-05,
|
| 591 |
+
"loss": 0.1118,
|
| 592 |
+
"step": 1725
|
| 593 |
+
},
|
| 594 |
+
{
|
| 595 |
+
"epoch": 10.0,
|
| 596 |
+
"grad_norm": 0.12201019376516342,
|
| 597 |
+
"learning_rate": 1.1117460317460317e-05,
|
| 598 |
+
"loss": 0.1416,
|
| 599 |
+
"step": 1750
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 10.0,
|
| 603 |
+
"eval_accuracy": 0.8288409703504043,
|
| 604 |
+
"eval_loss": 0.9273040294647217,
|
| 605 |
+
"eval_macro_f1": 0.8279250116936684,
|
| 606 |
+
"eval_runtime": 0.4893,
|
| 607 |
+
"eval_samples_per_second": 1516.421,
|
| 608 |
+
"eval_steps_per_second": 96.054,
|
| 609 |
+
"eval_weighted_f1": 0.8282596215308831,
|
| 610 |
+
"step": 1750
|
| 611 |
+
}
|
| 612 |
+
],
|
| 613 |
+
"logging_steps": 25,
|
| 614 |
+
"max_steps": 3500,
|
| 615 |
+
"num_input_tokens_seen": 0,
|
| 616 |
+
"num_train_epochs": 20,
|
| 617 |
+
"save_steps": 500,
|
| 618 |
+
"stateful_callbacks": {
|
| 619 |
+
"EarlyStoppingCallback": {
|
| 620 |
+
"args": {
|
| 621 |
+
"early_stopping_patience": 3,
|
| 622 |
+
"early_stopping_threshold": 0.0
|
| 623 |
+
},
|
| 624 |
+
"attributes": {
|
| 625 |
+
"early_stopping_patience_counter": 3
|
| 626 |
+
}
|
| 627 |
+
},
|
| 628 |
+
"TrainerControl": {
|
| 629 |
+
"args": {
|
| 630 |
+
"should_epoch_stop": false,
|
| 631 |
+
"should_evaluate": false,
|
| 632 |
+
"should_log": false,
|
| 633 |
+
"should_save": true,
|
| 634 |
+
"should_training_stop": true
|
| 635 |
+
},
|
| 636 |
+
"attributes": {}
|
| 637 |
+
}
|
| 638 |
+
},
|
| 639 |
+
"total_flos": 0.0,
|
| 640 |
+
"train_batch_size": 8,
|
| 641 |
+
"trial_name": null,
|
| 642 |
+
"trial_params": null
|
| 643 |
+
}
|
ood-vifactcheck-phobert-topic-domain-05-s44-20260826T111852Z-ab4112e3/trainer/checkpoint-1750/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-06-s44-20260826T112212Z-ab4112e3/trainer/checkpoint-1104/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-topic-domain-06-s44-20260826T112212Z-ab4112e3/trainer/checkpoint-1104/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-06-s44-20260826T112212Z-ab4112e3/trainer/checkpoint-1104/special_tokens_map.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": "<mask>",
|
| 6 |
+
"pad_token": "<pad>",
|
| 7 |
+
"sep_token": "</s>",
|
| 8 |
+
"unk_token": "<unk>"
|
| 9 |
+
}
|
ood-vifactcheck-phobert-topic-domain-06-s44-20260826T112212Z-ab4112e3/trainer/checkpoint-1104/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"64000": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 256,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "PhobertTokenizer",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-phobert-topic-domain-06-s44-20260826T112212Z-ab4112e3/trainer/checkpoint-1104/trainer_state.json
ADDED
|
@@ -0,0 +1,417 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1104,
|
| 3 |
+
"best_metric": 0.8418743152235338,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-topic-domain-06-s44-20260826T112212Z-ab4112e3/trainer/checkpoint-1104",
|
| 5 |
+
"epoch": 6.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1104,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.1364256480218281,
|
| 14 |
+
"grad_norm": 7.466602325439453,
|
| 15 |
+
"learning_rate": 1.3043478260869566e-06,
|
| 16 |
+
"loss": 4.4258,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.2728512960436562,
|
| 21 |
+
"grad_norm": 6.493659019470215,
|
| 22 |
+
"learning_rate": 2.6630434782608698e-06,
|
| 23 |
+
"loss": 4.3773,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.4092769440654843,
|
| 28 |
+
"grad_norm": 9.546185493469238,
|
| 29 |
+
"learning_rate": 4.021739130434783e-06,
|
| 30 |
+
"loss": 4.3508,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5457025920873124,
|
| 35 |
+
"grad_norm": 6.903392791748047,
|
| 36 |
+
"learning_rate": 5.380434782608695e-06,
|
| 37 |
+
"loss": 4.3749,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6821282401091405,
|
| 42 |
+
"grad_norm": 8.428988456726074,
|
| 43 |
+
"learning_rate": 6.739130434782609e-06,
|
| 44 |
+
"loss": 4.2909,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8185538881309686,
|
| 49 |
+
"grad_norm": 9.652042388916016,
|
| 50 |
+
"learning_rate": 8.097826086956523e-06,
|
| 51 |
+
"loss": 4.1908,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9549795361527967,
|
| 56 |
+
"grad_norm": 20.115737915039062,
|
| 57 |
+
"learning_rate": 9.456521739130436e-06,
|
| 58 |
+
"loss": 3.9099,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.5454545454545454,
|
| 64 |
+
"eval_loss": 0.969987690448761,
|
| 65 |
+
"eval_macro_f1": 0.49590549589676675,
|
| 66 |
+
"eval_runtime": 0.5994,
|
| 67 |
+
"eval_samples_per_second": 1247.89,
|
| 68 |
+
"eval_steps_per_second": 78.41,
|
| 69 |
+
"eval_weighted_f1": 0.4999070192835386,
|
| 70 |
+
"step": 184
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.08731241473397,
|
| 74 |
+
"grad_norm": 31.854272842407227,
|
| 75 |
+
"learning_rate": 1.0815217391304349e-05,
|
| 76 |
+
"loss": 3.4745,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.223738062755798,
|
| 81 |
+
"grad_norm": 24.532371520996094,
|
| 82 |
+
"learning_rate": 1.2173913043478263e-05,
|
| 83 |
+
"loss": 3.3423,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.360163710777626,
|
| 88 |
+
"grad_norm": 24.3089656829834,
|
| 89 |
+
"learning_rate": 1.3532608695652174e-05,
|
| 90 |
+
"loss": 2.9605,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.4965893587994543,
|
| 95 |
+
"grad_norm": 36.238922119140625,
|
| 96 |
+
"learning_rate": 1.4891304347826087e-05,
|
| 97 |
+
"loss": 2.7379,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.6330150068212825,
|
| 102 |
+
"grad_norm": 20.342714309692383,
|
| 103 |
+
"learning_rate": 1.6250000000000002e-05,
|
| 104 |
+
"loss": 2.5024,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7694406548431105,
|
| 109 |
+
"grad_norm": 60.8023796081543,
|
| 110 |
+
"learning_rate": 1.7608695652173915e-05,
|
| 111 |
+
"loss": 2.2671,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.9058663028649385,
|
| 116 |
+
"grad_norm": 39.70579528808594,
|
| 117 |
+
"learning_rate": 1.8967391304347828e-05,
|
| 118 |
+
"loss": 2.3222,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.7954545454545454,
|
| 124 |
+
"eval_loss": 0.5298647880554199,
|
| 125 |
+
"eval_macro_f1": 0.7890388609768451,
|
| 126 |
+
"eval_runtime": 0.5012,
|
| 127 |
+
"eval_samples_per_second": 1492.31,
|
| 128 |
+
"eval_steps_per_second": 93.768,
|
| 129 |
+
"eval_weighted_f1": 0.7902885029951275,
|
| 130 |
+
"step": 368
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.038199181446112,
|
| 134 |
+
"grad_norm": 29.71407699584961,
|
| 135 |
+
"learning_rate": 1.9963768115942032e-05,
|
| 136 |
+
"loss": 2.1494,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.17462482946794,
|
| 141 |
+
"grad_norm": 73.53707885742188,
|
| 142 |
+
"learning_rate": 1.9812801932367153e-05,
|
| 143 |
+
"loss": 1.6693,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.311050477489768,
|
| 148 |
+
"grad_norm": 23.069602966308594,
|
| 149 |
+
"learning_rate": 1.966183574879227e-05,
|
| 150 |
+
"loss": 1.5568,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.447476125511596,
|
| 155 |
+
"grad_norm": 15.098575592041016,
|
| 156 |
+
"learning_rate": 1.9510869565217393e-05,
|
| 157 |
+
"loss": 1.664,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.5839017735334244,
|
| 162 |
+
"grad_norm": 73.36605072021484,
|
| 163 |
+
"learning_rate": 1.9359903381642515e-05,
|
| 164 |
+
"loss": 1.6234,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.720327421555252,
|
| 169 |
+
"grad_norm": 30.50521469116211,
|
| 170 |
+
"learning_rate": 1.9208937198067633e-05,
|
| 171 |
+
"loss": 1.7241,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.8567530695770804,
|
| 176 |
+
"grad_norm": 110.73439025878906,
|
| 177 |
+
"learning_rate": 1.9057971014492754e-05,
|
| 178 |
+
"loss": 1.6371,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 2.9931787175989086,
|
| 183 |
+
"grad_norm": 38.11772155761719,
|
| 184 |
+
"learning_rate": 1.8907004830917876e-05,
|
| 185 |
+
"loss": 1.5924,
|
| 186 |
+
"step": 550
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 3.0,
|
| 190 |
+
"eval_accuracy": 0.8315508021390374,
|
| 191 |
+
"eval_loss": 0.4641593098640442,
|
| 192 |
+
"eval_macro_f1": 0.8322598076895886,
|
| 193 |
+
"eval_runtime": 0.5035,
|
| 194 |
+
"eval_samples_per_second": 1485.595,
|
| 195 |
+
"eval_steps_per_second": 93.346,
|
| 196 |
+
"eval_weighted_f1": 0.8323204111259968,
|
| 197 |
+
"step": 552
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.1255115961800817,
|
| 201 |
+
"grad_norm": 67.04457092285156,
|
| 202 |
+
"learning_rate": 1.8756038647342997e-05,
|
| 203 |
+
"loss": 1.1025,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.26193724420191,
|
| 208 |
+
"grad_norm": 53.298683166503906,
|
| 209 |
+
"learning_rate": 1.8605072463768115e-05,
|
| 210 |
+
"loss": 1.3097,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.398362892223738,
|
| 215 |
+
"grad_norm": 60.064125061035156,
|
| 216 |
+
"learning_rate": 1.8454106280193237e-05,
|
| 217 |
+
"loss": 1.1808,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.5347885402455663,
|
| 222 |
+
"grad_norm": 18.580032348632812,
|
| 223 |
+
"learning_rate": 1.830314009661836e-05,
|
| 224 |
+
"loss": 0.9484,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.6712141882673945,
|
| 229 |
+
"grad_norm": 15.898066520690918,
|
| 230 |
+
"learning_rate": 1.815217391304348e-05,
|
| 231 |
+
"loss": 1.0441,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.8076398362892223,
|
| 236 |
+
"grad_norm": 39.26723098754883,
|
| 237 |
+
"learning_rate": 1.8001207729468598e-05,
|
| 238 |
+
"loss": 1.175,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.9440654843110505,
|
| 243 |
+
"grad_norm": 128.85169982910156,
|
| 244 |
+
"learning_rate": 1.785024154589372e-05,
|
| 245 |
+
"loss": 1.0105,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 4.0,
|
| 250 |
+
"eval_accuracy": 0.839572192513369,
|
| 251 |
+
"eval_loss": 0.4994402527809143,
|
| 252 |
+
"eval_macro_f1": 0.8400012995451592,
|
| 253 |
+
"eval_runtime": 0.5155,
|
| 254 |
+
"eval_samples_per_second": 1450.994,
|
| 255 |
+
"eval_steps_per_second": 91.172,
|
| 256 |
+
"eval_weighted_f1": 0.8400200821423732,
|
| 257 |
+
"step": 736
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.076398362892224,
|
| 261 |
+
"grad_norm": 60.4824333190918,
|
| 262 |
+
"learning_rate": 1.769927536231884e-05,
|
| 263 |
+
"loss": 0.7984,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.212824010914052,
|
| 268 |
+
"grad_norm": 37.35578155517578,
|
| 269 |
+
"learning_rate": 1.7548309178743963e-05,
|
| 270 |
+
"loss": 0.8523,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.34924965893588,
|
| 275 |
+
"grad_norm": 83.08707427978516,
|
| 276 |
+
"learning_rate": 1.7397342995169084e-05,
|
| 277 |
+
"loss": 0.7283,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.485675306957708,
|
| 282 |
+
"grad_norm": 59.59986877441406,
|
| 283 |
+
"learning_rate": 1.7246376811594206e-05,
|
| 284 |
+
"loss": 0.8062,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.622100954979536,
|
| 289 |
+
"grad_norm": 19.21984100341797,
|
| 290 |
+
"learning_rate": 1.7095410628019324e-05,
|
| 291 |
+
"loss": 0.635,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.758526603001364,
|
| 296 |
+
"grad_norm": 40.22753143310547,
|
| 297 |
+
"learning_rate": 1.6944444444444446e-05,
|
| 298 |
+
"loss": 0.7494,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.894952251023192,
|
| 303 |
+
"grad_norm": 7.336869239807129,
|
| 304 |
+
"learning_rate": 1.6793478260869567e-05,
|
| 305 |
+
"loss": 0.6429,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 5.0,
|
| 310 |
+
"eval_accuracy": 0.8262032085561497,
|
| 311 |
+
"eval_loss": 0.5925671458244324,
|
| 312 |
+
"eval_macro_f1": 0.8263055625509157,
|
| 313 |
+
"eval_runtime": 0.4912,
|
| 314 |
+
"eval_samples_per_second": 1522.85,
|
| 315 |
+
"eval_steps_per_second": 95.687,
|
| 316 |
+
"eval_weighted_f1": 0.8266571930192887,
|
| 317 |
+
"step": 920
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.0272851296043655,
|
| 321 |
+
"grad_norm": 13.629725456237793,
|
| 322 |
+
"learning_rate": 1.664251207729469e-05,
|
| 323 |
+
"loss": 0.6556,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.163710777626194,
|
| 328 |
+
"grad_norm": 69.7869873046875,
|
| 329 |
+
"learning_rate": 1.649154589371981e-05,
|
| 330 |
+
"loss": 0.5352,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.300136425648022,
|
| 335 |
+
"grad_norm": 39.57054138183594,
|
| 336 |
+
"learning_rate": 1.634057971014493e-05,
|
| 337 |
+
"loss": 0.5037,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.43656207366985,
|
| 342 |
+
"grad_norm": 73.22761535644531,
|
| 343 |
+
"learning_rate": 1.618961352657005e-05,
|
| 344 |
+
"loss": 0.5501,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.572987721691678,
|
| 349 |
+
"grad_norm": 21.088153839111328,
|
| 350 |
+
"learning_rate": 1.603864734299517e-05,
|
| 351 |
+
"loss": 0.5329,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.709413369713506,
|
| 356 |
+
"grad_norm": 29.600107192993164,
|
| 357 |
+
"learning_rate": 1.5887681159420293e-05,
|
| 358 |
+
"loss": 0.4821,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.845839017735334,
|
| 363 |
+
"grad_norm": 2.194171190261841,
|
| 364 |
+
"learning_rate": 1.5736714975845415e-05,
|
| 365 |
+
"loss": 0.5719,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 5.982264665757162,
|
| 370 |
+
"grad_norm": 80.75299072265625,
|
| 371 |
+
"learning_rate": 1.5585748792270533e-05,
|
| 372 |
+
"loss": 0.5306,
|
| 373 |
+
"step": 1100
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 6.0,
|
| 377 |
+
"eval_accuracy": 0.8422459893048129,
|
| 378 |
+
"eval_loss": 0.620465874671936,
|
| 379 |
+
"eval_macro_f1": 0.8418743152235338,
|
| 380 |
+
"eval_runtime": 0.5309,
|
| 381 |
+
"eval_samples_per_second": 1408.861,
|
| 382 |
+
"eval_steps_per_second": 88.525,
|
| 383 |
+
"eval_weighted_f1": 0.8419501500622384,
|
| 384 |
+
"step": 1104
|
| 385 |
+
}
|
| 386 |
+
],
|
| 387 |
+
"logging_steps": 25,
|
| 388 |
+
"max_steps": 3680,
|
| 389 |
+
"num_input_tokens_seen": 0,
|
| 390 |
+
"num_train_epochs": 20,
|
| 391 |
+
"save_steps": 500,
|
| 392 |
+
"stateful_callbacks": {
|
| 393 |
+
"EarlyStoppingCallback": {
|
| 394 |
+
"args": {
|
| 395 |
+
"early_stopping_patience": 3,
|
| 396 |
+
"early_stopping_threshold": 0.0
|
| 397 |
+
},
|
| 398 |
+
"attributes": {
|
| 399 |
+
"early_stopping_patience_counter": 0
|
| 400 |
+
}
|
| 401 |
+
},
|
| 402 |
+
"TrainerControl": {
|
| 403 |
+
"args": {
|
| 404 |
+
"should_epoch_stop": false,
|
| 405 |
+
"should_evaluate": false,
|
| 406 |
+
"should_log": false,
|
| 407 |
+
"should_save": true,
|
| 408 |
+
"should_training_stop": false
|
| 409 |
+
},
|
| 410 |
+
"attributes": {}
|
| 411 |
+
}
|
| 412 |
+
},
|
| 413 |
+
"total_flos": 0.0,
|
| 414 |
+
"train_batch_size": 8,
|
| 415 |
+
"trial_name": null,
|
| 416 |
+
"trial_params": null
|
| 417 |
+
}
|
ood-vifactcheck-phobert-topic-domain-06-s44-20260826T112212Z-ab4112e3/trainer/checkpoint-1104/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-06-s44-20260826T112212Z-ab4112e3/trainer/checkpoint-1656/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-topic-domain-06-s44-20260826T112212Z-ab4112e3/trainer/checkpoint-1656/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-06-s44-20260826T112212Z-ab4112e3/trainer/checkpoint-1656/special_tokens_map.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": "<mask>",
|
| 6 |
+
"pad_token": "<pad>",
|
| 7 |
+
"sep_token": "</s>",
|
| 8 |
+
"unk_token": "<unk>"
|
| 9 |
+
}
|
ood-vifactcheck-phobert-topic-domain-06-s44-20260826T112212Z-ab4112e3/trainer/checkpoint-1656/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"64000": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 256,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "PhobertTokenizer",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-phobert-topic-domain-06-s44-20260826T112212Z-ab4112e3/trainer/checkpoint-1656/trainer_state.json
ADDED
|
@@ -0,0 +1,604 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1104,
|
| 3 |
+
"best_metric": 0.8418743152235338,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-topic-domain-06-s44-20260826T112212Z-ab4112e3/trainer/checkpoint-1104",
|
| 5 |
+
"epoch": 9.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1656,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.1364256480218281,
|
| 14 |
+
"grad_norm": 7.466602325439453,
|
| 15 |
+
"learning_rate": 1.3043478260869566e-06,
|
| 16 |
+
"loss": 4.4258,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.2728512960436562,
|
| 21 |
+
"grad_norm": 6.493659019470215,
|
| 22 |
+
"learning_rate": 2.6630434782608698e-06,
|
| 23 |
+
"loss": 4.3773,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.4092769440654843,
|
| 28 |
+
"grad_norm": 9.546185493469238,
|
| 29 |
+
"learning_rate": 4.021739130434783e-06,
|
| 30 |
+
"loss": 4.3508,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5457025920873124,
|
| 35 |
+
"grad_norm": 6.903392791748047,
|
| 36 |
+
"learning_rate": 5.380434782608695e-06,
|
| 37 |
+
"loss": 4.3749,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6821282401091405,
|
| 42 |
+
"grad_norm": 8.428988456726074,
|
| 43 |
+
"learning_rate": 6.739130434782609e-06,
|
| 44 |
+
"loss": 4.2909,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8185538881309686,
|
| 49 |
+
"grad_norm": 9.652042388916016,
|
| 50 |
+
"learning_rate": 8.097826086956523e-06,
|
| 51 |
+
"loss": 4.1908,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.9549795361527967,
|
| 56 |
+
"grad_norm": 20.115737915039062,
|
| 57 |
+
"learning_rate": 9.456521739130436e-06,
|
| 58 |
+
"loss": 3.9099,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.5454545454545454,
|
| 64 |
+
"eval_loss": 0.969987690448761,
|
| 65 |
+
"eval_macro_f1": 0.49590549589676675,
|
| 66 |
+
"eval_runtime": 0.5994,
|
| 67 |
+
"eval_samples_per_second": 1247.89,
|
| 68 |
+
"eval_steps_per_second": 78.41,
|
| 69 |
+
"eval_weighted_f1": 0.4999070192835386,
|
| 70 |
+
"step": 184
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.08731241473397,
|
| 74 |
+
"grad_norm": 31.854272842407227,
|
| 75 |
+
"learning_rate": 1.0815217391304349e-05,
|
| 76 |
+
"loss": 3.4745,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.223738062755798,
|
| 81 |
+
"grad_norm": 24.532371520996094,
|
| 82 |
+
"learning_rate": 1.2173913043478263e-05,
|
| 83 |
+
"loss": 3.3423,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.360163710777626,
|
| 88 |
+
"grad_norm": 24.3089656829834,
|
| 89 |
+
"learning_rate": 1.3532608695652174e-05,
|
| 90 |
+
"loss": 2.9605,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.4965893587994543,
|
| 95 |
+
"grad_norm": 36.238922119140625,
|
| 96 |
+
"learning_rate": 1.4891304347826087e-05,
|
| 97 |
+
"loss": 2.7379,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.6330150068212825,
|
| 102 |
+
"grad_norm": 20.342714309692383,
|
| 103 |
+
"learning_rate": 1.6250000000000002e-05,
|
| 104 |
+
"loss": 2.5024,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.7694406548431105,
|
| 109 |
+
"grad_norm": 60.8023796081543,
|
| 110 |
+
"learning_rate": 1.7608695652173915e-05,
|
| 111 |
+
"loss": 2.2671,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.9058663028649385,
|
| 116 |
+
"grad_norm": 39.70579528808594,
|
| 117 |
+
"learning_rate": 1.8967391304347828e-05,
|
| 118 |
+
"loss": 2.3222,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 2.0,
|
| 123 |
+
"eval_accuracy": 0.7954545454545454,
|
| 124 |
+
"eval_loss": 0.5298647880554199,
|
| 125 |
+
"eval_macro_f1": 0.7890388609768451,
|
| 126 |
+
"eval_runtime": 0.5012,
|
| 127 |
+
"eval_samples_per_second": 1492.31,
|
| 128 |
+
"eval_steps_per_second": 93.768,
|
| 129 |
+
"eval_weighted_f1": 0.7902885029951275,
|
| 130 |
+
"step": 368
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.038199181446112,
|
| 134 |
+
"grad_norm": 29.71407699584961,
|
| 135 |
+
"learning_rate": 1.9963768115942032e-05,
|
| 136 |
+
"loss": 2.1494,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.17462482946794,
|
| 141 |
+
"grad_norm": 73.53707885742188,
|
| 142 |
+
"learning_rate": 1.9812801932367153e-05,
|
| 143 |
+
"loss": 1.6693,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.311050477489768,
|
| 148 |
+
"grad_norm": 23.069602966308594,
|
| 149 |
+
"learning_rate": 1.966183574879227e-05,
|
| 150 |
+
"loss": 1.5568,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.447476125511596,
|
| 155 |
+
"grad_norm": 15.098575592041016,
|
| 156 |
+
"learning_rate": 1.9510869565217393e-05,
|
| 157 |
+
"loss": 1.664,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.5839017735334244,
|
| 162 |
+
"grad_norm": 73.36605072021484,
|
| 163 |
+
"learning_rate": 1.9359903381642515e-05,
|
| 164 |
+
"loss": 1.6234,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.720327421555252,
|
| 169 |
+
"grad_norm": 30.50521469116211,
|
| 170 |
+
"learning_rate": 1.9208937198067633e-05,
|
| 171 |
+
"loss": 1.7241,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.8567530695770804,
|
| 176 |
+
"grad_norm": 110.73439025878906,
|
| 177 |
+
"learning_rate": 1.9057971014492754e-05,
|
| 178 |
+
"loss": 1.6371,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 2.9931787175989086,
|
| 183 |
+
"grad_norm": 38.11772155761719,
|
| 184 |
+
"learning_rate": 1.8907004830917876e-05,
|
| 185 |
+
"loss": 1.5924,
|
| 186 |
+
"step": 550
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 3.0,
|
| 190 |
+
"eval_accuracy": 0.8315508021390374,
|
| 191 |
+
"eval_loss": 0.4641593098640442,
|
| 192 |
+
"eval_macro_f1": 0.8322598076895886,
|
| 193 |
+
"eval_runtime": 0.5035,
|
| 194 |
+
"eval_samples_per_second": 1485.595,
|
| 195 |
+
"eval_steps_per_second": 93.346,
|
| 196 |
+
"eval_weighted_f1": 0.8323204111259968,
|
| 197 |
+
"step": 552
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.1255115961800817,
|
| 201 |
+
"grad_norm": 67.04457092285156,
|
| 202 |
+
"learning_rate": 1.8756038647342997e-05,
|
| 203 |
+
"loss": 1.1025,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.26193724420191,
|
| 208 |
+
"grad_norm": 53.298683166503906,
|
| 209 |
+
"learning_rate": 1.8605072463768115e-05,
|
| 210 |
+
"loss": 1.3097,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.398362892223738,
|
| 215 |
+
"grad_norm": 60.064125061035156,
|
| 216 |
+
"learning_rate": 1.8454106280193237e-05,
|
| 217 |
+
"loss": 1.1808,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.5347885402455663,
|
| 222 |
+
"grad_norm": 18.580032348632812,
|
| 223 |
+
"learning_rate": 1.830314009661836e-05,
|
| 224 |
+
"loss": 0.9484,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.6712141882673945,
|
| 229 |
+
"grad_norm": 15.898066520690918,
|
| 230 |
+
"learning_rate": 1.815217391304348e-05,
|
| 231 |
+
"loss": 1.0441,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.8076398362892223,
|
| 236 |
+
"grad_norm": 39.26723098754883,
|
| 237 |
+
"learning_rate": 1.8001207729468598e-05,
|
| 238 |
+
"loss": 1.175,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.9440654843110505,
|
| 243 |
+
"grad_norm": 128.85169982910156,
|
| 244 |
+
"learning_rate": 1.785024154589372e-05,
|
| 245 |
+
"loss": 1.0105,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 4.0,
|
| 250 |
+
"eval_accuracy": 0.839572192513369,
|
| 251 |
+
"eval_loss": 0.4994402527809143,
|
| 252 |
+
"eval_macro_f1": 0.8400012995451592,
|
| 253 |
+
"eval_runtime": 0.5155,
|
| 254 |
+
"eval_samples_per_second": 1450.994,
|
| 255 |
+
"eval_steps_per_second": 91.172,
|
| 256 |
+
"eval_weighted_f1": 0.8400200821423732,
|
| 257 |
+
"step": 736
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.076398362892224,
|
| 261 |
+
"grad_norm": 60.4824333190918,
|
| 262 |
+
"learning_rate": 1.769927536231884e-05,
|
| 263 |
+
"loss": 0.7984,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.212824010914052,
|
| 268 |
+
"grad_norm": 37.35578155517578,
|
| 269 |
+
"learning_rate": 1.7548309178743963e-05,
|
| 270 |
+
"loss": 0.8523,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.34924965893588,
|
| 275 |
+
"grad_norm": 83.08707427978516,
|
| 276 |
+
"learning_rate": 1.7397342995169084e-05,
|
| 277 |
+
"loss": 0.7283,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.485675306957708,
|
| 282 |
+
"grad_norm": 59.59986877441406,
|
| 283 |
+
"learning_rate": 1.7246376811594206e-05,
|
| 284 |
+
"loss": 0.8062,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.622100954979536,
|
| 289 |
+
"grad_norm": 19.21984100341797,
|
| 290 |
+
"learning_rate": 1.7095410628019324e-05,
|
| 291 |
+
"loss": 0.635,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.758526603001364,
|
| 296 |
+
"grad_norm": 40.22753143310547,
|
| 297 |
+
"learning_rate": 1.6944444444444446e-05,
|
| 298 |
+
"loss": 0.7494,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.894952251023192,
|
| 303 |
+
"grad_norm": 7.336869239807129,
|
| 304 |
+
"learning_rate": 1.6793478260869567e-05,
|
| 305 |
+
"loss": 0.6429,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 5.0,
|
| 310 |
+
"eval_accuracy": 0.8262032085561497,
|
| 311 |
+
"eval_loss": 0.5925671458244324,
|
| 312 |
+
"eval_macro_f1": 0.8263055625509157,
|
| 313 |
+
"eval_runtime": 0.4912,
|
| 314 |
+
"eval_samples_per_second": 1522.85,
|
| 315 |
+
"eval_steps_per_second": 95.687,
|
| 316 |
+
"eval_weighted_f1": 0.8266571930192887,
|
| 317 |
+
"step": 920
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.0272851296043655,
|
| 321 |
+
"grad_norm": 13.629725456237793,
|
| 322 |
+
"learning_rate": 1.664251207729469e-05,
|
| 323 |
+
"loss": 0.6556,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.163710777626194,
|
| 328 |
+
"grad_norm": 69.7869873046875,
|
| 329 |
+
"learning_rate": 1.649154589371981e-05,
|
| 330 |
+
"loss": 0.5352,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.300136425648022,
|
| 335 |
+
"grad_norm": 39.57054138183594,
|
| 336 |
+
"learning_rate": 1.634057971014493e-05,
|
| 337 |
+
"loss": 0.5037,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.43656207366985,
|
| 342 |
+
"grad_norm": 73.22761535644531,
|
| 343 |
+
"learning_rate": 1.618961352657005e-05,
|
| 344 |
+
"loss": 0.5501,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.572987721691678,
|
| 349 |
+
"grad_norm": 21.088153839111328,
|
| 350 |
+
"learning_rate": 1.603864734299517e-05,
|
| 351 |
+
"loss": 0.5329,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.709413369713506,
|
| 356 |
+
"grad_norm": 29.600107192993164,
|
| 357 |
+
"learning_rate": 1.5887681159420293e-05,
|
| 358 |
+
"loss": 0.4821,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.845839017735334,
|
| 363 |
+
"grad_norm": 2.194171190261841,
|
| 364 |
+
"learning_rate": 1.5736714975845415e-05,
|
| 365 |
+
"loss": 0.5719,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 5.982264665757162,
|
| 370 |
+
"grad_norm": 80.75299072265625,
|
| 371 |
+
"learning_rate": 1.5585748792270533e-05,
|
| 372 |
+
"loss": 0.5306,
|
| 373 |
+
"step": 1100
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 6.0,
|
| 377 |
+
"eval_accuracy": 0.8422459893048129,
|
| 378 |
+
"eval_loss": 0.620465874671936,
|
| 379 |
+
"eval_macro_f1": 0.8418743152235338,
|
| 380 |
+
"eval_runtime": 0.5309,
|
| 381 |
+
"eval_samples_per_second": 1408.861,
|
| 382 |
+
"eval_steps_per_second": 88.525,
|
| 383 |
+
"eval_weighted_f1": 0.8419501500622384,
|
| 384 |
+
"step": 1104
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 6.114597544338336,
|
| 388 |
+
"grad_norm": 77.83263397216797,
|
| 389 |
+
"learning_rate": 1.5434782608695654e-05,
|
| 390 |
+
"loss": 0.347,
|
| 391 |
+
"step": 1125
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 6.251023192360163,
|
| 395 |
+
"grad_norm": 75.55965423583984,
|
| 396 |
+
"learning_rate": 1.5283816425120776e-05,
|
| 397 |
+
"loss": 0.4206,
|
| 398 |
+
"step": 1150
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.387448840381992,
|
| 402 |
+
"grad_norm": 200.16627502441406,
|
| 403 |
+
"learning_rate": 1.5132850241545896e-05,
|
| 404 |
+
"loss": 0.3207,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.52387448840382,
|
| 409 |
+
"grad_norm": 14.267569541931152,
|
| 410 |
+
"learning_rate": 1.4981884057971017e-05,
|
| 411 |
+
"loss": 0.5526,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.660300136425648,
|
| 416 |
+
"grad_norm": 96.6065902709961,
|
| 417 |
+
"learning_rate": 1.4830917874396135e-05,
|
| 418 |
+
"loss": 0.3227,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.796725784447476,
|
| 423 |
+
"grad_norm": 0.4383222460746765,
|
| 424 |
+
"learning_rate": 1.4679951690821257e-05,
|
| 425 |
+
"loss": 0.6142,
|
| 426 |
+
"step": 1250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 6.933151432469304,
|
| 430 |
+
"grad_norm": 5.978261947631836,
|
| 431 |
+
"learning_rate": 1.4528985507246378e-05,
|
| 432 |
+
"loss": 0.3543,
|
| 433 |
+
"step": 1275
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 7.0,
|
| 437 |
+
"eval_accuracy": 0.8368983957219251,
|
| 438 |
+
"eval_loss": 0.7540048360824585,
|
| 439 |
+
"eval_macro_f1": 0.8367259571155384,
|
| 440 |
+
"eval_runtime": 0.5216,
|
| 441 |
+
"eval_samples_per_second": 1434.162,
|
| 442 |
+
"eval_steps_per_second": 90.114,
|
| 443 |
+
"eval_weighted_f1": 0.8370218667262496,
|
| 444 |
+
"step": 1288
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 7.0654843110504775,
|
| 448 |
+
"grad_norm": 1.8376160860061646,
|
| 449 |
+
"learning_rate": 1.43780193236715e-05,
|
| 450 |
+
"loss": 0.2413,
|
| 451 |
+
"step": 1300
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 7.201909959072306,
|
| 455 |
+
"grad_norm": 0.33565554022789,
|
| 456 |
+
"learning_rate": 1.422705314009662e-05,
|
| 457 |
+
"loss": 0.3808,
|
| 458 |
+
"step": 1325
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 7.338335607094134,
|
| 462 |
+
"grad_norm": 12.636874198913574,
|
| 463 |
+
"learning_rate": 1.407608695652174e-05,
|
| 464 |
+
"loss": 0.3391,
|
| 465 |
+
"step": 1350
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 7.474761255115962,
|
| 469 |
+
"grad_norm": 14.967347145080566,
|
| 470 |
+
"learning_rate": 1.3925120772946861e-05,
|
| 471 |
+
"loss": 0.3196,
|
| 472 |
+
"step": 1375
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 7.61118690313779,
|
| 476 |
+
"grad_norm": 68.00273132324219,
|
| 477 |
+
"learning_rate": 1.3774154589371983e-05,
|
| 478 |
+
"loss": 0.3138,
|
| 479 |
+
"step": 1400
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 7.747612551159618,
|
| 483 |
+
"grad_norm": 139.94955444335938,
|
| 484 |
+
"learning_rate": 1.3623188405797103e-05,
|
| 485 |
+
"loss": 0.2326,
|
| 486 |
+
"step": 1425
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 7.884038199181446,
|
| 490 |
+
"grad_norm": 1.7836552858352661,
|
| 491 |
+
"learning_rate": 1.3472222222222224e-05,
|
| 492 |
+
"loss": 0.2864,
|
| 493 |
+
"step": 1450
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 8.0,
|
| 497 |
+
"eval_accuracy": 0.8382352941176471,
|
| 498 |
+
"eval_loss": 0.7864946722984314,
|
| 499 |
+
"eval_macro_f1": 0.8386784778531645,
|
| 500 |
+
"eval_runtime": 0.5908,
|
| 501 |
+
"eval_samples_per_second": 1265.993,
|
| 502 |
+
"eval_steps_per_second": 79.548,
|
| 503 |
+
"eval_weighted_f1": 0.8386013878044508,
|
| 504 |
+
"step": 1472
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 8.016371077762619,
|
| 508 |
+
"grad_norm": 37.00727844238281,
|
| 509 |
+
"learning_rate": 1.3321256038647344e-05,
|
| 510 |
+
"loss": 0.3295,
|
| 511 |
+
"step": 1475
|
| 512 |
+
},
|
| 513 |
+
{
|
| 514 |
+
"epoch": 8.152796725784448,
|
| 515 |
+
"grad_norm": 0.24529293179512024,
|
| 516 |
+
"learning_rate": 1.3170289855072464e-05,
|
| 517 |
+
"loss": 0.3555,
|
| 518 |
+
"step": 1500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 8.289222373806275,
|
| 522 |
+
"grad_norm": 3.803839921951294,
|
| 523 |
+
"learning_rate": 1.3019323671497585e-05,
|
| 524 |
+
"loss": 0.2143,
|
| 525 |
+
"step": 1525
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 8.425648021828104,
|
| 529 |
+
"grad_norm": 1.8090559244155884,
|
| 530 |
+
"learning_rate": 1.2868357487922707e-05,
|
| 531 |
+
"loss": 0.1898,
|
| 532 |
+
"step": 1550
|
| 533 |
+
},
|
| 534 |
+
{
|
| 535 |
+
"epoch": 8.562073669849932,
|
| 536 |
+
"grad_norm": 0.2629372775554657,
|
| 537 |
+
"learning_rate": 1.2717391304347828e-05,
|
| 538 |
+
"loss": 0.1968,
|
| 539 |
+
"step": 1575
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 8.69849931787176,
|
| 543 |
+
"grad_norm": 0.17688652873039246,
|
| 544 |
+
"learning_rate": 1.2566425120772947e-05,
|
| 545 |
+
"loss": 0.1035,
|
| 546 |
+
"step": 1600
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 8.834924965893588,
|
| 550 |
+
"grad_norm": 0.20628862082958221,
|
| 551 |
+
"learning_rate": 1.2415458937198068e-05,
|
| 552 |
+
"loss": 0.2244,
|
| 553 |
+
"step": 1625
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 8.971350613915416,
|
| 557 |
+
"grad_norm": 24.174407958984375,
|
| 558 |
+
"learning_rate": 1.226449275362319e-05,
|
| 559 |
+
"loss": 0.3238,
|
| 560 |
+
"step": 1650
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 9.0,
|
| 564 |
+
"eval_accuracy": 0.8355614973262032,
|
| 565 |
+
"eval_loss": 0.8632557988166809,
|
| 566 |
+
"eval_macro_f1": 0.835163029087906,
|
| 567 |
+
"eval_runtime": 0.5176,
|
| 568 |
+
"eval_samples_per_second": 1445.224,
|
| 569 |
+
"eval_steps_per_second": 90.81,
|
| 570 |
+
"eval_weighted_f1": 0.8352410679055065,
|
| 571 |
+
"step": 1656
|
| 572 |
+
}
|
| 573 |
+
],
|
| 574 |
+
"logging_steps": 25,
|
| 575 |
+
"max_steps": 3680,
|
| 576 |
+
"num_input_tokens_seen": 0,
|
| 577 |
+
"num_train_epochs": 20,
|
| 578 |
+
"save_steps": 500,
|
| 579 |
+
"stateful_callbacks": {
|
| 580 |
+
"EarlyStoppingCallback": {
|
| 581 |
+
"args": {
|
| 582 |
+
"early_stopping_patience": 3,
|
| 583 |
+
"early_stopping_threshold": 0.0
|
| 584 |
+
},
|
| 585 |
+
"attributes": {
|
| 586 |
+
"early_stopping_patience_counter": 3
|
| 587 |
+
}
|
| 588 |
+
},
|
| 589 |
+
"TrainerControl": {
|
| 590 |
+
"args": {
|
| 591 |
+
"should_epoch_stop": false,
|
| 592 |
+
"should_evaluate": false,
|
| 593 |
+
"should_log": false,
|
| 594 |
+
"should_save": true,
|
| 595 |
+
"should_training_stop": true
|
| 596 |
+
},
|
| 597 |
+
"attributes": {}
|
| 598 |
+
}
|
| 599 |
+
},
|
| 600 |
+
"total_flos": 0.0,
|
| 601 |
+
"train_batch_size": 8,
|
| 602 |
+
"trial_name": null,
|
| 603 |
+
"trial_params": null
|
| 604 |
+
}
|
ood-vifactcheck-phobert-topic-domain-06-s44-20260826T112212Z-ab4112e3/trainer/checkpoint-1656/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-07-s44-20260826T112526Z-ab4112e3/trainer/checkpoint-1379/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-topic-domain-07-s44-20260826T112526Z-ab4112e3/trainer/checkpoint-1379/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-07-s44-20260826T112526Z-ab4112e3/trainer/checkpoint-1379/special_tokens_map.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": "<mask>",
|
| 6 |
+
"pad_token": "<pad>",
|
| 7 |
+
"sep_token": "</s>",
|
| 8 |
+
"unk_token": "<unk>"
|
| 9 |
+
}
|
ood-vifactcheck-phobert-topic-domain-07-s44-20260826T112526Z-ab4112e3/trainer/checkpoint-1379/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"64000": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 256,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "PhobertTokenizer",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-phobert-topic-domain-07-s44-20260826T112526Z-ab4112e3/trainer/checkpoint-1379/trainer_state.json
ADDED
|
@@ -0,0 +1,505 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 788,
|
| 3 |
+
"best_metric": 0.8617025526281172,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-topic-domain-07-s44-20260826T112526Z-ab4112e3/trainer/checkpoint-788",
|
| 5 |
+
"epoch": 7.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1379,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.12706480304955528,
|
| 14 |
+
"grad_norm": 7.947079658508301,
|
| 15 |
+
"learning_rate": 1.218274111675127e-06,
|
| 16 |
+
"loss": 4.4434,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.25412960609911056,
|
| 21 |
+
"grad_norm": 9.109493255615234,
|
| 22 |
+
"learning_rate": 2.487309644670051e-06,
|
| 23 |
+
"loss": 4.4237,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.3811944091486658,
|
| 28 |
+
"grad_norm": 8.6504545211792,
|
| 29 |
+
"learning_rate": 3.756345177664975e-06,
|
| 30 |
+
"loss": 4.3967,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5082592121982211,
|
| 35 |
+
"grad_norm": 7.3916473388671875,
|
| 36 |
+
"learning_rate": 5.025380710659899e-06,
|
| 37 |
+
"loss": 4.3643,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6353240152477764,
|
| 42 |
+
"grad_norm": 8.624367713928223,
|
| 43 |
+
"learning_rate": 6.2944162436548225e-06,
|
| 44 |
+
"loss": 4.2858,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.7623888182973316,
|
| 49 |
+
"grad_norm": 11.439811706542969,
|
| 50 |
+
"learning_rate": 7.563451776649747e-06,
|
| 51 |
+
"loss": 4.1555,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.8894536213468869,
|
| 56 |
+
"grad_norm": 33.69758605957031,
|
| 57 |
+
"learning_rate": 8.832487309644671e-06,
|
| 58 |
+
"loss": 4.0057,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.6068759342301944,
|
| 64 |
+
"eval_loss": 0.8771998286247253,
|
| 65 |
+
"eval_macro_f1": 0.5903004195538256,
|
| 66 |
+
"eval_runtime": 0.4748,
|
| 67 |
+
"eval_samples_per_second": 1408.918,
|
| 68 |
+
"eval_steps_per_second": 88.452,
|
| 69 |
+
"eval_weighted_f1": 0.5916977740373158,
|
| 70 |
+
"step": 197
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0152477763659467,
|
| 74 |
+
"grad_norm": 31.929121017456055,
|
| 75 |
+
"learning_rate": 1.0101522842639596e-05,
|
| 76 |
+
"loss": 3.6638,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.1423125794155018,
|
| 81 |
+
"grad_norm": 43.645084381103516,
|
| 82 |
+
"learning_rate": 1.1370558375634519e-05,
|
| 83 |
+
"loss": 3.4842,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.2693773824650572,
|
| 88 |
+
"grad_norm": 48.55276107788086,
|
| 89 |
+
"learning_rate": 1.2639593908629442e-05,
|
| 90 |
+
"loss": 3.2856,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.3964421855146125,
|
| 95 |
+
"grad_norm": 50.12979507446289,
|
| 96 |
+
"learning_rate": 1.3908629441624366e-05,
|
| 97 |
+
"loss": 3.0123,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.5235069885641677,
|
| 102 |
+
"grad_norm": 37.894901275634766,
|
| 103 |
+
"learning_rate": 1.5177664974619291e-05,
|
| 104 |
+
"loss": 2.6343,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.650571791613723,
|
| 109 |
+
"grad_norm": 33.85085678100586,
|
| 110 |
+
"learning_rate": 1.6446700507614214e-05,
|
| 111 |
+
"loss": 2.4439,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.7776365946632784,
|
| 116 |
+
"grad_norm": 40.96009063720703,
|
| 117 |
+
"learning_rate": 1.771573604060914e-05,
|
| 118 |
+
"loss": 2.2235,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 1.9047013977128335,
|
| 123 |
+
"grad_norm": 49.43701171875,
|
| 124 |
+
"learning_rate": 1.8984771573604063e-05,
|
| 125 |
+
"loss": 2.2721,
|
| 126 |
+
"step": 375
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"epoch": 2.0,
|
| 130 |
+
"eval_accuracy": 0.8370702541106129,
|
| 131 |
+
"eval_loss": 0.4577271640300751,
|
| 132 |
+
"eval_macro_f1": 0.8367230629781583,
|
| 133 |
+
"eval_runtime": 0.4445,
|
| 134 |
+
"eval_samples_per_second": 1505.054,
|
| 135 |
+
"eval_steps_per_second": 94.488,
|
| 136 |
+
"eval_weighted_f1": 0.8369150576357447,
|
| 137 |
+
"step": 394
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.0304955527318933,
|
| 141 |
+
"grad_norm": 52.79537582397461,
|
| 142 |
+
"learning_rate": 1.997179921037789e-05,
|
| 143 |
+
"loss": 2.0769,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.1575603557814484,
|
| 148 |
+
"grad_norm": 61.200138092041016,
|
| 149 |
+
"learning_rate": 1.9830795262267346e-05,
|
| 150 |
+
"loss": 1.7633,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.2846251588310036,
|
| 155 |
+
"grad_norm": 36.30815887451172,
|
| 156 |
+
"learning_rate": 1.96897913141568e-05,
|
| 157 |
+
"loss": 1.8138,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.411689961880559,
|
| 162 |
+
"grad_norm": 62.584781646728516,
|
| 163 |
+
"learning_rate": 1.954878736604625e-05,
|
| 164 |
+
"loss": 1.615,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.5387547649301143,
|
| 169 |
+
"grad_norm": 15.579201698303223,
|
| 170 |
+
"learning_rate": 1.9407783417935703e-05,
|
| 171 |
+
"loss": 1.6278,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.6658195679796695,
|
| 176 |
+
"grad_norm": 32.65045928955078,
|
| 177 |
+
"learning_rate": 1.9266779469825156e-05,
|
| 178 |
+
"loss": 1.4847,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 2.792884371029225,
|
| 183 |
+
"grad_norm": 34.87415313720703,
|
| 184 |
+
"learning_rate": 1.912577552171461e-05,
|
| 185 |
+
"loss": 1.7246,
|
| 186 |
+
"step": 550
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 2.91994917407878,
|
| 190 |
+
"grad_norm": 42.98828125,
|
| 191 |
+
"learning_rate": 1.8984771573604063e-05,
|
| 192 |
+
"loss": 1.6823,
|
| 193 |
+
"step": 575
|
| 194 |
+
},
|
| 195 |
+
{
|
| 196 |
+
"epoch": 3.0,
|
| 197 |
+
"eval_accuracy": 0.8146487294469357,
|
| 198 |
+
"eval_loss": 0.534042477607727,
|
| 199 |
+
"eval_macro_f1": 0.8102402234971793,
|
| 200 |
+
"eval_runtime": 0.4423,
|
| 201 |
+
"eval_samples_per_second": 1512.702,
|
| 202 |
+
"eval_steps_per_second": 94.968,
|
| 203 |
+
"eval_weighted_f1": 0.8107100454608398,
|
| 204 |
+
"step": 591
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.04574332909784,
|
| 208 |
+
"grad_norm": 41.373844146728516,
|
| 209 |
+
"learning_rate": 1.8843767625493517e-05,
|
| 210 |
+
"loss": 1.585,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.172808132147395,
|
| 215 |
+
"grad_norm": 49.36249923706055,
|
| 216 |
+
"learning_rate": 1.870276367738297e-05,
|
| 217 |
+
"loss": 1.138,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.2998729351969507,
|
| 222 |
+
"grad_norm": 18.925535202026367,
|
| 223 |
+
"learning_rate": 1.856175972927242e-05,
|
| 224 |
+
"loss": 1.009,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.426937738246506,
|
| 229 |
+
"grad_norm": 38.49395751953125,
|
| 230 |
+
"learning_rate": 1.8420755781161873e-05,
|
| 231 |
+
"loss": 1.047,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.554002541296061,
|
| 236 |
+
"grad_norm": 30.579927444458008,
|
| 237 |
+
"learning_rate": 1.8279751833051327e-05,
|
| 238 |
+
"loss": 1.1788,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.681067344345616,
|
| 243 |
+
"grad_norm": 47.26395797729492,
|
| 244 |
+
"learning_rate": 1.813874788494078e-05,
|
| 245 |
+
"loss": 0.9307,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 3.8081321473951717,
|
| 250 |
+
"grad_norm": 54.2235107421875,
|
| 251 |
+
"learning_rate": 1.7997743936830234e-05,
|
| 252 |
+
"loss": 1.0903,
|
| 253 |
+
"step": 750
|
| 254 |
+
},
|
| 255 |
+
{
|
| 256 |
+
"epoch": 3.935196950444727,
|
| 257 |
+
"grad_norm": 23.86235237121582,
|
| 258 |
+
"learning_rate": 1.7856739988719684e-05,
|
| 259 |
+
"loss": 1.1646,
|
| 260 |
+
"step": 775
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 4.0,
|
| 264 |
+
"eval_accuracy": 0.8624813153961136,
|
| 265 |
+
"eval_loss": 0.41413968801498413,
|
| 266 |
+
"eval_macro_f1": 0.8617025526281172,
|
| 267 |
+
"eval_runtime": 0.4524,
|
| 268 |
+
"eval_samples_per_second": 1478.779,
|
| 269 |
+
"eval_steps_per_second": 92.838,
|
| 270 |
+
"eval_weighted_f1": 0.8619621402174493,
|
| 271 |
+
"step": 788
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.060991105463787,
|
| 275 |
+
"grad_norm": 48.239681243896484,
|
| 276 |
+
"learning_rate": 1.771573604060914e-05,
|
| 277 |
+
"loss": 0.9527,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.188055908513342,
|
| 282 |
+
"grad_norm": 78.02395629882812,
|
| 283 |
+
"learning_rate": 1.7574732092498594e-05,
|
| 284 |
+
"loss": 0.5694,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 4.315120711562897,
|
| 289 |
+
"grad_norm": 27.761552810668945,
|
| 290 |
+
"learning_rate": 1.7433728144388044e-05,
|
| 291 |
+
"loss": 0.8991,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 4.4421855146124525,
|
| 296 |
+
"grad_norm": 70.77861785888672,
|
| 297 |
+
"learning_rate": 1.7292724196277497e-05,
|
| 298 |
+
"loss": 0.765,
|
| 299 |
+
"step": 875
|
| 300 |
+
},
|
| 301 |
+
{
|
| 302 |
+
"epoch": 4.569250317662007,
|
| 303 |
+
"grad_norm": 104.53265380859375,
|
| 304 |
+
"learning_rate": 1.715172024816695e-05,
|
| 305 |
+
"loss": 0.7464,
|
| 306 |
+
"step": 900
|
| 307 |
+
},
|
| 308 |
+
{
|
| 309 |
+
"epoch": 4.696315120711563,
|
| 310 |
+
"grad_norm": 20.85953712463379,
|
| 311 |
+
"learning_rate": 1.7010716300056404e-05,
|
| 312 |
+
"loss": 0.7537,
|
| 313 |
+
"step": 925
|
| 314 |
+
},
|
| 315 |
+
{
|
| 316 |
+
"epoch": 4.823379923761118,
|
| 317 |
+
"grad_norm": 81.1286849975586,
|
| 318 |
+
"learning_rate": 1.6869712351945854e-05,
|
| 319 |
+
"loss": 0.7003,
|
| 320 |
+
"step": 950
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 4.950444726810673,
|
| 324 |
+
"grad_norm": 10.733736991882324,
|
| 325 |
+
"learning_rate": 1.672870840383531e-05,
|
| 326 |
+
"loss": 0.8773,
|
| 327 |
+
"step": 975
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 5.0,
|
| 331 |
+
"eval_accuracy": 0.8609865470852018,
|
| 332 |
+
"eval_loss": 0.4355868399143219,
|
| 333 |
+
"eval_macro_f1": 0.8609511191736402,
|
| 334 |
+
"eval_runtime": 0.4419,
|
| 335 |
+
"eval_samples_per_second": 1513.824,
|
| 336 |
+
"eval_steps_per_second": 95.038,
|
| 337 |
+
"eval_weighted_f1": 0.8611504956834228,
|
| 338 |
+
"step": 985
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.076238881829733,
|
| 342 |
+
"grad_norm": 115.65019226074219,
|
| 343 |
+
"learning_rate": 1.6587704455724764e-05,
|
| 344 |
+
"loss": 0.4643,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 5.203303684879288,
|
| 349 |
+
"grad_norm": 27.35481071472168,
|
| 350 |
+
"learning_rate": 1.6446700507614214e-05,
|
| 351 |
+
"loss": 0.4502,
|
| 352 |
+
"step": 1025
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 5.330368487928844,
|
| 356 |
+
"grad_norm": 16.8605899810791,
|
| 357 |
+
"learning_rate": 1.6305696559503667e-05,
|
| 358 |
+
"loss": 0.4464,
|
| 359 |
+
"step": 1050
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 5.457433290978399,
|
| 363 |
+
"grad_norm": 78.07289123535156,
|
| 364 |
+
"learning_rate": 1.616469261139312e-05,
|
| 365 |
+
"loss": 0.499,
|
| 366 |
+
"step": 1075
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 5.584498094027954,
|
| 370 |
+
"grad_norm": 91.50618743896484,
|
| 371 |
+
"learning_rate": 1.6023688663282574e-05,
|
| 372 |
+
"loss": 0.7001,
|
| 373 |
+
"step": 1100
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 5.71156289707751,
|
| 377 |
+
"grad_norm": 79.87772369384766,
|
| 378 |
+
"learning_rate": 1.5882684715172024e-05,
|
| 379 |
+
"loss": 0.5603,
|
| 380 |
+
"step": 1125
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 5.838627700127065,
|
| 384 |
+
"grad_norm": 9.959702491760254,
|
| 385 |
+
"learning_rate": 1.5741680767061477e-05,
|
| 386 |
+
"loss": 0.746,
|
| 387 |
+
"step": 1150
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 5.96569250317662,
|
| 391 |
+
"grad_norm": 45.134552001953125,
|
| 392 |
+
"learning_rate": 1.5600676818950934e-05,
|
| 393 |
+
"loss": 0.6466,
|
| 394 |
+
"step": 1175
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 6.0,
|
| 398 |
+
"eval_accuracy": 0.8565022421524664,
|
| 399 |
+
"eval_loss": 0.5075973868370056,
|
| 400 |
+
"eval_macro_f1": 0.8566957729395082,
|
| 401 |
+
"eval_runtime": 0.471,
|
| 402 |
+
"eval_samples_per_second": 1420.253,
|
| 403 |
+
"eval_steps_per_second": 89.164,
|
| 404 |
+
"eval_weighted_f1": 0.8567884789253076,
|
| 405 |
+
"step": 1182
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 6.09148665819568,
|
| 409 |
+
"grad_norm": 32.23667526245117,
|
| 410 |
+
"learning_rate": 1.5459672870840384e-05,
|
| 411 |
+
"loss": 0.3441,
|
| 412 |
+
"step": 1200
|
| 413 |
+
},
|
| 414 |
+
{
|
| 415 |
+
"epoch": 6.2185514612452355,
|
| 416 |
+
"grad_norm": 26.145593643188477,
|
| 417 |
+
"learning_rate": 1.5318668922729838e-05,
|
| 418 |
+
"loss": 0.5518,
|
| 419 |
+
"step": 1225
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 6.34561626429479,
|
| 423 |
+
"grad_norm": 66.14224243164062,
|
| 424 |
+
"learning_rate": 1.5177664974619291e-05,
|
| 425 |
+
"loss": 0.5419,
|
| 426 |
+
"step": 1250
|
| 427 |
+
},
|
| 428 |
+
{
|
| 429 |
+
"epoch": 6.472681067344346,
|
| 430 |
+
"grad_norm": 66.82027435302734,
|
| 431 |
+
"learning_rate": 1.5036661026508743e-05,
|
| 432 |
+
"loss": 0.5322,
|
| 433 |
+
"step": 1275
|
| 434 |
+
},
|
| 435 |
+
{
|
| 436 |
+
"epoch": 6.599745870393901,
|
| 437 |
+
"grad_norm": 101.67127990722656,
|
| 438 |
+
"learning_rate": 1.4895657078398196e-05,
|
| 439 |
+
"loss": 0.5729,
|
| 440 |
+
"step": 1300
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 6.726810673443456,
|
| 444 |
+
"grad_norm": 16.341772079467773,
|
| 445 |
+
"learning_rate": 1.4754653130287648e-05,
|
| 446 |
+
"loss": 0.1972,
|
| 447 |
+
"step": 1325
|
| 448 |
+
},
|
| 449 |
+
{
|
| 450 |
+
"epoch": 6.853875476493012,
|
| 451 |
+
"grad_norm": 88.38990783691406,
|
| 452 |
+
"learning_rate": 1.4613649182177103e-05,
|
| 453 |
+
"loss": 0.332,
|
| 454 |
+
"step": 1350
|
| 455 |
+
},
|
| 456 |
+
{
|
| 457 |
+
"epoch": 6.980940279542566,
|
| 458 |
+
"grad_norm": 158.80921936035156,
|
| 459 |
+
"learning_rate": 1.4472645234066556e-05,
|
| 460 |
+
"loss": 0.5271,
|
| 461 |
+
"step": 1375
|
| 462 |
+
},
|
| 463 |
+
{
|
| 464 |
+
"epoch": 7.0,
|
| 465 |
+
"eval_accuracy": 0.8609865470852018,
|
| 466 |
+
"eval_loss": 0.6492565274238586,
|
| 467 |
+
"eval_macro_f1": 0.8601128173482385,
|
| 468 |
+
"eval_runtime": 0.549,
|
| 469 |
+
"eval_samples_per_second": 1218.477,
|
| 470 |
+
"eval_steps_per_second": 76.496,
|
| 471 |
+
"eval_weighted_f1": 0.8603272426127853,
|
| 472 |
+
"step": 1379
|
| 473 |
+
}
|
| 474 |
+
],
|
| 475 |
+
"logging_steps": 25,
|
| 476 |
+
"max_steps": 3940,
|
| 477 |
+
"num_input_tokens_seen": 0,
|
| 478 |
+
"num_train_epochs": 20,
|
| 479 |
+
"save_steps": 500,
|
| 480 |
+
"stateful_callbacks": {
|
| 481 |
+
"EarlyStoppingCallback": {
|
| 482 |
+
"args": {
|
| 483 |
+
"early_stopping_patience": 3,
|
| 484 |
+
"early_stopping_threshold": 0.0
|
| 485 |
+
},
|
| 486 |
+
"attributes": {
|
| 487 |
+
"early_stopping_patience_counter": 3
|
| 488 |
+
}
|
| 489 |
+
},
|
| 490 |
+
"TrainerControl": {
|
| 491 |
+
"args": {
|
| 492 |
+
"should_epoch_stop": false,
|
| 493 |
+
"should_evaluate": false,
|
| 494 |
+
"should_log": false,
|
| 495 |
+
"should_save": true,
|
| 496 |
+
"should_training_stop": true
|
| 497 |
+
},
|
| 498 |
+
"attributes": {}
|
| 499 |
+
}
|
| 500 |
+
},
|
| 501 |
+
"total_flos": 0.0,
|
| 502 |
+
"train_batch_size": 8,
|
| 503 |
+
"trial_name": null,
|
| 504 |
+
"trial_params": null
|
| 505 |
+
}
|
ood-vifactcheck-phobert-topic-domain-07-s44-20260826T112526Z-ab4112e3/trainer/checkpoint-1379/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-07-s44-20260826T112526Z-ab4112e3/trainer/checkpoint-788/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-topic-domain-07-s44-20260826T112526Z-ab4112e3/trainer/checkpoint-788/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-07-s44-20260826T112526Z-ab4112e3/trainer/checkpoint-788/special_tokens_map.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": "<mask>",
|
| 6 |
+
"pad_token": "<pad>",
|
| 7 |
+
"sep_token": "</s>",
|
| 8 |
+
"unk_token": "<unk>"
|
| 9 |
+
}
|
ood-vifactcheck-phobert-topic-domain-07-s44-20260826T112526Z-ab4112e3/trainer/checkpoint-788/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"64000": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 256,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "PhobertTokenizer",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-phobert-topic-domain-07-s44-20260826T112526Z-ab4112e3/trainer/checkpoint-788/trainer_state.json
ADDED
|
@@ -0,0 +1,304 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 788,
|
| 3 |
+
"best_metric": 0.8617025526281172,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-topic-domain-07-s44-20260826T112526Z-ab4112e3/trainer/checkpoint-788",
|
| 5 |
+
"epoch": 4.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 788,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.12706480304955528,
|
| 14 |
+
"grad_norm": 7.947079658508301,
|
| 15 |
+
"learning_rate": 1.218274111675127e-06,
|
| 16 |
+
"loss": 4.4434,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.25412960609911056,
|
| 21 |
+
"grad_norm": 9.109493255615234,
|
| 22 |
+
"learning_rate": 2.487309644670051e-06,
|
| 23 |
+
"loss": 4.4237,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.3811944091486658,
|
| 28 |
+
"grad_norm": 8.6504545211792,
|
| 29 |
+
"learning_rate": 3.756345177664975e-06,
|
| 30 |
+
"loss": 4.3967,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5082592121982211,
|
| 35 |
+
"grad_norm": 7.3916473388671875,
|
| 36 |
+
"learning_rate": 5.025380710659899e-06,
|
| 37 |
+
"loss": 4.3643,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.6353240152477764,
|
| 42 |
+
"grad_norm": 8.624367713928223,
|
| 43 |
+
"learning_rate": 6.2944162436548225e-06,
|
| 44 |
+
"loss": 4.2858,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.7623888182973316,
|
| 49 |
+
"grad_norm": 11.439811706542969,
|
| 50 |
+
"learning_rate": 7.563451776649747e-06,
|
| 51 |
+
"loss": 4.1555,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.8894536213468869,
|
| 56 |
+
"grad_norm": 33.69758605957031,
|
| 57 |
+
"learning_rate": 8.832487309644671e-06,
|
| 58 |
+
"loss": 4.0057,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 1.0,
|
| 63 |
+
"eval_accuracy": 0.6068759342301944,
|
| 64 |
+
"eval_loss": 0.8771998286247253,
|
| 65 |
+
"eval_macro_f1": 0.5903004195538256,
|
| 66 |
+
"eval_runtime": 0.4748,
|
| 67 |
+
"eval_samples_per_second": 1408.918,
|
| 68 |
+
"eval_steps_per_second": 88.452,
|
| 69 |
+
"eval_weighted_f1": 0.5916977740373158,
|
| 70 |
+
"step": 197
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.0152477763659467,
|
| 74 |
+
"grad_norm": 31.929121017456055,
|
| 75 |
+
"learning_rate": 1.0101522842639596e-05,
|
| 76 |
+
"loss": 3.6638,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.1423125794155018,
|
| 81 |
+
"grad_norm": 43.645084381103516,
|
| 82 |
+
"learning_rate": 1.1370558375634519e-05,
|
| 83 |
+
"loss": 3.4842,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.2693773824650572,
|
| 88 |
+
"grad_norm": 48.55276107788086,
|
| 89 |
+
"learning_rate": 1.2639593908629442e-05,
|
| 90 |
+
"loss": 3.2856,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.3964421855146125,
|
| 95 |
+
"grad_norm": 50.12979507446289,
|
| 96 |
+
"learning_rate": 1.3908629441624366e-05,
|
| 97 |
+
"loss": 3.0123,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.5235069885641677,
|
| 102 |
+
"grad_norm": 37.894901275634766,
|
| 103 |
+
"learning_rate": 1.5177664974619291e-05,
|
| 104 |
+
"loss": 2.6343,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.650571791613723,
|
| 109 |
+
"grad_norm": 33.85085678100586,
|
| 110 |
+
"learning_rate": 1.6446700507614214e-05,
|
| 111 |
+
"loss": 2.4439,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 1.7776365946632784,
|
| 116 |
+
"grad_norm": 40.96009063720703,
|
| 117 |
+
"learning_rate": 1.771573604060914e-05,
|
| 118 |
+
"loss": 2.2235,
|
| 119 |
+
"step": 350
|
| 120 |
+
},
|
| 121 |
+
{
|
| 122 |
+
"epoch": 1.9047013977128335,
|
| 123 |
+
"grad_norm": 49.43701171875,
|
| 124 |
+
"learning_rate": 1.8984771573604063e-05,
|
| 125 |
+
"loss": 2.2721,
|
| 126 |
+
"step": 375
|
| 127 |
+
},
|
| 128 |
+
{
|
| 129 |
+
"epoch": 2.0,
|
| 130 |
+
"eval_accuracy": 0.8370702541106129,
|
| 131 |
+
"eval_loss": 0.4577271640300751,
|
| 132 |
+
"eval_macro_f1": 0.8367230629781583,
|
| 133 |
+
"eval_runtime": 0.4445,
|
| 134 |
+
"eval_samples_per_second": 1505.054,
|
| 135 |
+
"eval_steps_per_second": 94.488,
|
| 136 |
+
"eval_weighted_f1": 0.8369150576357447,
|
| 137 |
+
"step": 394
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.0304955527318933,
|
| 141 |
+
"grad_norm": 52.79537582397461,
|
| 142 |
+
"learning_rate": 1.997179921037789e-05,
|
| 143 |
+
"loss": 2.0769,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.1575603557814484,
|
| 148 |
+
"grad_norm": 61.200138092041016,
|
| 149 |
+
"learning_rate": 1.9830795262267346e-05,
|
| 150 |
+
"loss": 1.7633,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.2846251588310036,
|
| 155 |
+
"grad_norm": 36.30815887451172,
|
| 156 |
+
"learning_rate": 1.96897913141568e-05,
|
| 157 |
+
"loss": 1.8138,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.411689961880559,
|
| 162 |
+
"grad_norm": 62.584781646728516,
|
| 163 |
+
"learning_rate": 1.954878736604625e-05,
|
| 164 |
+
"loss": 1.615,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.5387547649301143,
|
| 169 |
+
"grad_norm": 15.579201698303223,
|
| 170 |
+
"learning_rate": 1.9407783417935703e-05,
|
| 171 |
+
"loss": 1.6278,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 2.6658195679796695,
|
| 176 |
+
"grad_norm": 32.65045928955078,
|
| 177 |
+
"learning_rate": 1.9266779469825156e-05,
|
| 178 |
+
"loss": 1.4847,
|
| 179 |
+
"step": 525
|
| 180 |
+
},
|
| 181 |
+
{
|
| 182 |
+
"epoch": 2.792884371029225,
|
| 183 |
+
"grad_norm": 34.87415313720703,
|
| 184 |
+
"learning_rate": 1.912577552171461e-05,
|
| 185 |
+
"loss": 1.7246,
|
| 186 |
+
"step": 550
|
| 187 |
+
},
|
| 188 |
+
{
|
| 189 |
+
"epoch": 2.91994917407878,
|
| 190 |
+
"grad_norm": 42.98828125,
|
| 191 |
+
"learning_rate": 1.8984771573604063e-05,
|
| 192 |
+
"loss": 1.6823,
|
| 193 |
+
"step": 575
|
| 194 |
+
},
|
| 195 |
+
{
|
| 196 |
+
"epoch": 3.0,
|
| 197 |
+
"eval_accuracy": 0.8146487294469357,
|
| 198 |
+
"eval_loss": 0.534042477607727,
|
| 199 |
+
"eval_macro_f1": 0.8102402234971793,
|
| 200 |
+
"eval_runtime": 0.4423,
|
| 201 |
+
"eval_samples_per_second": 1512.702,
|
| 202 |
+
"eval_steps_per_second": 94.968,
|
| 203 |
+
"eval_weighted_f1": 0.8107100454608398,
|
| 204 |
+
"step": 591
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.04574332909784,
|
| 208 |
+
"grad_norm": 41.373844146728516,
|
| 209 |
+
"learning_rate": 1.8843767625493517e-05,
|
| 210 |
+
"loss": 1.585,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.172808132147395,
|
| 215 |
+
"grad_norm": 49.36249923706055,
|
| 216 |
+
"learning_rate": 1.870276367738297e-05,
|
| 217 |
+
"loss": 1.138,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.2998729351969507,
|
| 222 |
+
"grad_norm": 18.925535202026367,
|
| 223 |
+
"learning_rate": 1.856175972927242e-05,
|
| 224 |
+
"loss": 1.009,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.426937738246506,
|
| 229 |
+
"grad_norm": 38.49395751953125,
|
| 230 |
+
"learning_rate": 1.8420755781161873e-05,
|
| 231 |
+
"loss": 1.047,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 3.554002541296061,
|
| 236 |
+
"grad_norm": 30.579927444458008,
|
| 237 |
+
"learning_rate": 1.8279751833051327e-05,
|
| 238 |
+
"loss": 1.1788,
|
| 239 |
+
"step": 700
|
| 240 |
+
},
|
| 241 |
+
{
|
| 242 |
+
"epoch": 3.681067344345616,
|
| 243 |
+
"grad_norm": 47.26395797729492,
|
| 244 |
+
"learning_rate": 1.813874788494078e-05,
|
| 245 |
+
"loss": 0.9307,
|
| 246 |
+
"step": 725
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 3.8081321473951717,
|
| 250 |
+
"grad_norm": 54.2235107421875,
|
| 251 |
+
"learning_rate": 1.7997743936830234e-05,
|
| 252 |
+
"loss": 1.0903,
|
| 253 |
+
"step": 750
|
| 254 |
+
},
|
| 255 |
+
{
|
| 256 |
+
"epoch": 3.935196950444727,
|
| 257 |
+
"grad_norm": 23.86235237121582,
|
| 258 |
+
"learning_rate": 1.7856739988719684e-05,
|
| 259 |
+
"loss": 1.1646,
|
| 260 |
+
"step": 775
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 4.0,
|
| 264 |
+
"eval_accuracy": 0.8624813153961136,
|
| 265 |
+
"eval_loss": 0.41413968801498413,
|
| 266 |
+
"eval_macro_f1": 0.8617025526281172,
|
| 267 |
+
"eval_runtime": 0.4524,
|
| 268 |
+
"eval_samples_per_second": 1478.779,
|
| 269 |
+
"eval_steps_per_second": 92.838,
|
| 270 |
+
"eval_weighted_f1": 0.8619621402174493,
|
| 271 |
+
"step": 788
|
| 272 |
+
}
|
| 273 |
+
],
|
| 274 |
+
"logging_steps": 25,
|
| 275 |
+
"max_steps": 3940,
|
| 276 |
+
"num_input_tokens_seen": 0,
|
| 277 |
+
"num_train_epochs": 20,
|
| 278 |
+
"save_steps": 500,
|
| 279 |
+
"stateful_callbacks": {
|
| 280 |
+
"EarlyStoppingCallback": {
|
| 281 |
+
"args": {
|
| 282 |
+
"early_stopping_patience": 3,
|
| 283 |
+
"early_stopping_threshold": 0.0
|
| 284 |
+
},
|
| 285 |
+
"attributes": {
|
| 286 |
+
"early_stopping_patience_counter": 0
|
| 287 |
+
}
|
| 288 |
+
},
|
| 289 |
+
"TrainerControl": {
|
| 290 |
+
"args": {
|
| 291 |
+
"should_epoch_stop": false,
|
| 292 |
+
"should_evaluate": false,
|
| 293 |
+
"should_log": false,
|
| 294 |
+
"should_save": true,
|
| 295 |
+
"should_training_stop": false
|
| 296 |
+
},
|
| 297 |
+
"attributes": {}
|
| 298 |
+
}
|
| 299 |
+
},
|
| 300 |
+
"total_flos": 0.0,
|
| 301 |
+
"train_batch_size": 8,
|
| 302 |
+
"trial_name": null,
|
| 303 |
+
"trial_params": null
|
| 304 |
+
}
|
ood-vifactcheck-phobert-topic-domain-07-s44-20260826T112526Z-ab4112e3/trainer/checkpoint-788/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-08-s44-20260826T112804Z-ab4112e3/trainer/checkpoint-1360/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-vifactcheck-phobert-topic-domain-08-s44-20260826T112804Z-ab4112e3/trainer/checkpoint-1360/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-08-s44-20260826T112804Z-ab4112e3/trainer/checkpoint-1360/special_tokens_map.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": "<mask>",
|
| 6 |
+
"pad_token": "<pad>",
|
| 7 |
+
"sep_token": "</s>",
|
| 8 |
+
"unk_token": "<unk>"
|
| 9 |
+
}
|
ood-vifactcheck-phobert-topic-domain-08-s44-20260826T112804Z-ab4112e3/trainer/checkpoint-1360/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"64000": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 256,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "PhobertTokenizer",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-vifactcheck-phobert-topic-domain-08-s44-20260826T112804Z-ab4112e3/trainer/checkpoint-1360/trainer_state.json
ADDED
|
@@ -0,0 +1,509 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 850,
|
| 3 |
+
"best_metric": 0.8367700458730928,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-vifactcheck-phobert-topic-domain-08-s44-20260826T112804Z-ab4112e3/trainer/checkpoint-850",
|
| 5 |
+
"epoch": 8.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1360,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.14727540500736377,
|
| 14 |
+
"grad_norm": 8.028051376342773,
|
| 15 |
+
"learning_rate": 1.4117647058823531e-06,
|
| 16 |
+
"loss": 4.3895,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.29455081001472755,
|
| 21 |
+
"grad_norm": 7.6496195793151855,
|
| 22 |
+
"learning_rate": 2.8823529411764704e-06,
|
| 23 |
+
"loss": 4.4136,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.4418262150220913,
|
| 28 |
+
"grad_norm": 6.894656658172607,
|
| 29 |
+
"learning_rate": 4.352941176470588e-06,
|
| 30 |
+
"loss": 4.3501,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.5891016200294551,
|
| 35 |
+
"grad_norm": 7.361630916595459,
|
| 36 |
+
"learning_rate": 5.823529411764707e-06,
|
| 37 |
+
"loss": 4.2759,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.7363770250368189,
|
| 42 |
+
"grad_norm": 13.567724227905273,
|
| 43 |
+
"learning_rate": 7.294117647058823e-06,
|
| 44 |
+
"loss": 4.1729,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.8836524300441826,
|
| 49 |
+
"grad_norm": 17.20767593383789,
|
| 50 |
+
"learning_rate": 8.764705882352942e-06,
|
| 51 |
+
"loss": 4.0336,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 1.0,
|
| 56 |
+
"eval_accuracy": 0.5632582322357019,
|
| 57 |
+
"eval_loss": 0.9292525053024292,
|
| 58 |
+
"eval_macro_f1": 0.5550093641094379,
|
| 59 |
+
"eval_runtime": 0.4603,
|
| 60 |
+
"eval_samples_per_second": 1253.598,
|
| 61 |
+
"eval_steps_per_second": 80.387,
|
| 62 |
+
"eval_weighted_f1": 0.5565895011649498,
|
| 63 |
+
"step": 170
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 1.0294550810014726,
|
| 67 |
+
"grad_norm": 21.848913192749023,
|
| 68 |
+
"learning_rate": 1.023529411764706e-05,
|
| 69 |
+
"loss": 3.8586,
|
| 70 |
+
"step": 175
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.1767304860088366,
|
| 74 |
+
"grad_norm": 57.24699401855469,
|
| 75 |
+
"learning_rate": 1.1705882352941178e-05,
|
| 76 |
+
"loss": 3.5053,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.3240058910162003,
|
| 81 |
+
"grad_norm": 46.81692123413086,
|
| 82 |
+
"learning_rate": 1.3176470588235294e-05,
|
| 83 |
+
"loss": 3.0883,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.4712812960235642,
|
| 88 |
+
"grad_norm": 31.254060745239258,
|
| 89 |
+
"learning_rate": 1.4647058823529412e-05,
|
| 90 |
+
"loss": 3.0342,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 1.6185567010309279,
|
| 95 |
+
"grad_norm": 39.008460998535156,
|
| 96 |
+
"learning_rate": 1.6117647058823533e-05,
|
| 97 |
+
"loss": 2.713,
|
| 98 |
+
"step": 275
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 1.7658321060382915,
|
| 102 |
+
"grad_norm": 40.096927642822266,
|
| 103 |
+
"learning_rate": 1.7588235294117647e-05,
|
| 104 |
+
"loss": 2.5328,
|
| 105 |
+
"step": 300
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 1.9131075110456552,
|
| 109 |
+
"grad_norm": 45.193626403808594,
|
| 110 |
+
"learning_rate": 1.9058823529411764e-05,
|
| 111 |
+
"loss": 2.5201,
|
| 112 |
+
"step": 325
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"epoch": 2.0,
|
| 116 |
+
"eval_accuracy": 0.7400346620450606,
|
| 117 |
+
"eval_loss": 0.6434793472290039,
|
| 118 |
+
"eval_macro_f1": 0.7375917643230546,
|
| 119 |
+
"eval_runtime": 0.4045,
|
| 120 |
+
"eval_samples_per_second": 1426.524,
|
| 121 |
+
"eval_steps_per_second": 91.476,
|
| 122 |
+
"eval_weighted_f1": 0.7379945451367538,
|
| 123 |
+
"step": 340
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 2.0589101620029453,
|
| 127 |
+
"grad_norm": 42.04426193237305,
|
| 128 |
+
"learning_rate": 1.9941176470588238e-05,
|
| 129 |
+
"loss": 2.3653,
|
| 130 |
+
"step": 350
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.2061855670103094,
|
| 134 |
+
"grad_norm": 28.513450622558594,
|
| 135 |
+
"learning_rate": 1.977777777777778e-05,
|
| 136 |
+
"loss": 1.9617,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 2.353460972017673,
|
| 141 |
+
"grad_norm": 46.21199417114258,
|
| 142 |
+
"learning_rate": 1.961437908496732e-05,
|
| 143 |
+
"loss": 1.6551,
|
| 144 |
+
"step": 400
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 2.500736377025037,
|
| 148 |
+
"grad_norm": 50.72047805786133,
|
| 149 |
+
"learning_rate": 1.9450980392156864e-05,
|
| 150 |
+
"loss": 1.63,
|
| 151 |
+
"step": 425
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 2.6480117820324005,
|
| 155 |
+
"grad_norm": 33.356842041015625,
|
| 156 |
+
"learning_rate": 1.9287581699346406e-05,
|
| 157 |
+
"loss": 1.8081,
|
| 158 |
+
"step": 450
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 2.795287187039764,
|
| 162 |
+
"grad_norm": 60.19363784790039,
|
| 163 |
+
"learning_rate": 1.912418300653595e-05,
|
| 164 |
+
"loss": 1.7406,
|
| 165 |
+
"step": 475
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 2.9425625920471283,
|
| 169 |
+
"grad_norm": 33.84236145019531,
|
| 170 |
+
"learning_rate": 1.8960784313725492e-05,
|
| 171 |
+
"loss": 1.7378,
|
| 172 |
+
"step": 500
|
| 173 |
+
},
|
| 174 |
+
{
|
| 175 |
+
"epoch": 3.0,
|
| 176 |
+
"eval_accuracy": 0.7920277296360485,
|
| 177 |
+
"eval_loss": 0.5614328980445862,
|
| 178 |
+
"eval_macro_f1": 0.78894659112811,
|
| 179 |
+
"eval_runtime": 0.4314,
|
| 180 |
+
"eval_samples_per_second": 1337.651,
|
| 181 |
+
"eval_steps_per_second": 85.777,
|
| 182 |
+
"eval_weighted_f1": 0.7895594206741525,
|
| 183 |
+
"step": 510
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 3.0883652430044184,
|
| 187 |
+
"grad_norm": 107.41642761230469,
|
| 188 |
+
"learning_rate": 1.8797385620915035e-05,
|
| 189 |
+
"loss": 1.3088,
|
| 190 |
+
"step": 525
|
| 191 |
+
},
|
| 192 |
+
{
|
| 193 |
+
"epoch": 3.235640648011782,
|
| 194 |
+
"grad_norm": 76.99580383300781,
|
| 195 |
+
"learning_rate": 1.8633986928104578e-05,
|
| 196 |
+
"loss": 1.2168,
|
| 197 |
+
"step": 550
|
| 198 |
+
},
|
| 199 |
+
{
|
| 200 |
+
"epoch": 3.3829160530191458,
|
| 201 |
+
"grad_norm": 37.805030822753906,
|
| 202 |
+
"learning_rate": 1.847058823529412e-05,
|
| 203 |
+
"loss": 1.0313,
|
| 204 |
+
"step": 575
|
| 205 |
+
},
|
| 206 |
+
{
|
| 207 |
+
"epoch": 3.5301914580265095,
|
| 208 |
+
"grad_norm": 53.588069915771484,
|
| 209 |
+
"learning_rate": 1.830718954248366e-05,
|
| 210 |
+
"loss": 0.8804,
|
| 211 |
+
"step": 600
|
| 212 |
+
},
|
| 213 |
+
{
|
| 214 |
+
"epoch": 3.6774668630338736,
|
| 215 |
+
"grad_norm": 50.77442169189453,
|
| 216 |
+
"learning_rate": 1.8143790849673203e-05,
|
| 217 |
+
"loss": 1.0672,
|
| 218 |
+
"step": 625
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 3.824742268041237,
|
| 222 |
+
"grad_norm": 17.985292434692383,
|
| 223 |
+
"learning_rate": 1.7980392156862746e-05,
|
| 224 |
+
"loss": 1.0904,
|
| 225 |
+
"step": 650
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 3.972017673048601,
|
| 229 |
+
"grad_norm": 31.843887329101562,
|
| 230 |
+
"learning_rate": 1.781699346405229e-05,
|
| 231 |
+
"loss": 1.1251,
|
| 232 |
+
"step": 675
|
| 233 |
+
},
|
| 234 |
+
{
|
| 235 |
+
"epoch": 4.0,
|
| 236 |
+
"eval_accuracy": 0.8076256499133448,
|
| 237 |
+
"eval_loss": 0.5661143064498901,
|
| 238 |
+
"eval_macro_f1": 0.8075518614445029,
|
| 239 |
+
"eval_runtime": 0.4653,
|
| 240 |
+
"eval_samples_per_second": 1240.131,
|
| 241 |
+
"eval_steps_per_second": 79.523,
|
| 242 |
+
"eval_weighted_f1": 0.8073352182026874,
|
| 243 |
+
"step": 680
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 4.117820324005891,
|
| 247 |
+
"grad_norm": 40.300193786621094,
|
| 248 |
+
"learning_rate": 1.7653594771241832e-05,
|
| 249 |
+
"loss": 0.7503,
|
| 250 |
+
"step": 700
|
| 251 |
+
},
|
| 252 |
+
{
|
| 253 |
+
"epoch": 4.265095729013255,
|
| 254 |
+
"grad_norm": 19.95840072631836,
|
| 255 |
+
"learning_rate": 1.7490196078431375e-05,
|
| 256 |
+
"loss": 0.8237,
|
| 257 |
+
"step": 725
|
| 258 |
+
},
|
| 259 |
+
{
|
| 260 |
+
"epoch": 4.412371134020619,
|
| 261 |
+
"grad_norm": 46.056732177734375,
|
| 262 |
+
"learning_rate": 1.7326797385620918e-05,
|
| 263 |
+
"loss": 0.811,
|
| 264 |
+
"step": 750
|
| 265 |
+
},
|
| 266 |
+
{
|
| 267 |
+
"epoch": 4.559646539027982,
|
| 268 |
+
"grad_norm": 30.100013732910156,
|
| 269 |
+
"learning_rate": 1.7163398692810457e-05,
|
| 270 |
+
"loss": 0.8699,
|
| 271 |
+
"step": 775
|
| 272 |
+
},
|
| 273 |
+
{
|
| 274 |
+
"epoch": 4.706921944035346,
|
| 275 |
+
"grad_norm": 64.97836303710938,
|
| 276 |
+
"learning_rate": 1.7e-05,
|
| 277 |
+
"loss": 0.9331,
|
| 278 |
+
"step": 800
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 4.8541973490427095,
|
| 282 |
+
"grad_norm": 30.842016220092773,
|
| 283 |
+
"learning_rate": 1.6836601307189543e-05,
|
| 284 |
+
"loss": 0.7782,
|
| 285 |
+
"step": 825
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 5.0,
|
| 289 |
+
"grad_norm": 86.25627136230469,
|
| 290 |
+
"learning_rate": 1.6673202614379086e-05,
|
| 291 |
+
"loss": 0.6946,
|
| 292 |
+
"step": 850
|
| 293 |
+
},
|
| 294 |
+
{
|
| 295 |
+
"epoch": 5.0,
|
| 296 |
+
"eval_accuracy": 0.8370883882149047,
|
| 297 |
+
"eval_loss": 0.5425134897232056,
|
| 298 |
+
"eval_macro_f1": 0.8367700458730928,
|
| 299 |
+
"eval_runtime": 0.4262,
|
| 300 |
+
"eval_samples_per_second": 1353.983,
|
| 301 |
+
"eval_steps_per_second": 86.824,
|
| 302 |
+
"eval_weighted_f1": 0.8367608736443399,
|
| 303 |
+
"step": 850
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 5.147275405007364,
|
| 307 |
+
"grad_norm": 47.28628921508789,
|
| 308 |
+
"learning_rate": 1.650980392156863e-05,
|
| 309 |
+
"loss": 0.5324,
|
| 310 |
+
"step": 875
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 5.294550810014727,
|
| 314 |
+
"grad_norm": 17.315105438232422,
|
| 315 |
+
"learning_rate": 1.634640522875817e-05,
|
| 316 |
+
"loss": 0.4167,
|
| 317 |
+
"step": 900
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 5.4418262150220915,
|
| 321 |
+
"grad_norm": 117.94629669189453,
|
| 322 |
+
"learning_rate": 1.6183006535947714e-05,
|
| 323 |
+
"loss": 0.6147,
|
| 324 |
+
"step": 925
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 5.589101620029455,
|
| 328 |
+
"grad_norm": 25.299196243286133,
|
| 329 |
+
"learning_rate": 1.6019607843137257e-05,
|
| 330 |
+
"loss": 0.4019,
|
| 331 |
+
"step": 950
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 5.736377025036819,
|
| 335 |
+
"grad_norm": 61.51991653442383,
|
| 336 |
+
"learning_rate": 1.5856209150326797e-05,
|
| 337 |
+
"loss": 0.6984,
|
| 338 |
+
"step": 975
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 5.883652430044183,
|
| 342 |
+
"grad_norm": 24.30650520324707,
|
| 343 |
+
"learning_rate": 1.569281045751634e-05,
|
| 344 |
+
"loss": 0.7553,
|
| 345 |
+
"step": 1000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 6.0,
|
| 349 |
+
"eval_accuracy": 0.8336221837088388,
|
| 350 |
+
"eval_loss": 0.6628820896148682,
|
| 351 |
+
"eval_macro_f1": 0.8331067516648032,
|
| 352 |
+
"eval_runtime": 0.426,
|
| 353 |
+
"eval_samples_per_second": 1354.545,
|
| 354 |
+
"eval_steps_per_second": 86.86,
|
| 355 |
+
"eval_weighted_f1": 0.8331629948030571,
|
| 356 |
+
"step": 1020
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 6.029455081001473,
|
| 360 |
+
"grad_norm": 15.124712944030762,
|
| 361 |
+
"learning_rate": 1.5529411764705882e-05,
|
| 362 |
+
"loss": 0.4527,
|
| 363 |
+
"step": 1025
|
| 364 |
+
},
|
| 365 |
+
{
|
| 366 |
+
"epoch": 6.176730486008837,
|
| 367 |
+
"grad_norm": 48.87225341796875,
|
| 368 |
+
"learning_rate": 1.536601307189543e-05,
|
| 369 |
+
"loss": 0.328,
|
| 370 |
+
"step": 1050
|
| 371 |
+
},
|
| 372 |
+
{
|
| 373 |
+
"epoch": 6.3240058910162,
|
| 374 |
+
"grad_norm": 7.919480323791504,
|
| 375 |
+
"learning_rate": 1.5202614379084968e-05,
|
| 376 |
+
"loss": 0.368,
|
| 377 |
+
"step": 1075
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 6.471281296023564,
|
| 381 |
+
"grad_norm": 5.432925224304199,
|
| 382 |
+
"learning_rate": 1.5039215686274511e-05,
|
| 383 |
+
"loss": 0.5217,
|
| 384 |
+
"step": 1100
|
| 385 |
+
},
|
| 386 |
+
{
|
| 387 |
+
"epoch": 6.618556701030927,
|
| 388 |
+
"grad_norm": 0.8563740849494934,
|
| 389 |
+
"learning_rate": 1.4875816993464054e-05,
|
| 390 |
+
"loss": 0.3912,
|
| 391 |
+
"step": 1125
|
| 392 |
+
},
|
| 393 |
+
{
|
| 394 |
+
"epoch": 6.7658321060382915,
|
| 395 |
+
"grad_norm": 107.15255737304688,
|
| 396 |
+
"learning_rate": 1.4712418300653597e-05,
|
| 397 |
+
"loss": 0.479,
|
| 398 |
+
"step": 1150
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 6.913107511045656,
|
| 402 |
+
"grad_norm": 16.084745407104492,
|
| 403 |
+
"learning_rate": 1.4549019607843138e-05,
|
| 404 |
+
"loss": 0.4546,
|
| 405 |
+
"step": 1175
|
| 406 |
+
},
|
| 407 |
+
{
|
| 408 |
+
"epoch": 7.0,
|
| 409 |
+
"eval_accuracy": 0.8214904679376083,
|
| 410 |
+
"eval_loss": 0.8287202715873718,
|
| 411 |
+
"eval_macro_f1": 0.8198040190719221,
|
| 412 |
+
"eval_runtime": 0.5864,
|
| 413 |
+
"eval_samples_per_second": 984.034,
|
| 414 |
+
"eval_steps_per_second": 63.101,
|
| 415 |
+
"eval_weighted_f1": 0.8197491229376538,
|
| 416 |
+
"step": 1190
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 7.058910162002945,
|
| 420 |
+
"grad_norm": 0.4035715162754059,
|
| 421 |
+
"learning_rate": 1.438562091503268e-05,
|
| 422 |
+
"loss": 0.3647,
|
| 423 |
+
"step": 1200
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 7.206185567010309,
|
| 427 |
+
"grad_norm": 129.20272827148438,
|
| 428 |
+
"learning_rate": 1.4222222222222224e-05,
|
| 429 |
+
"loss": 0.2728,
|
| 430 |
+
"step": 1225
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 7.353460972017673,
|
| 434 |
+
"grad_norm": 0.29202210903167725,
|
| 435 |
+
"learning_rate": 1.4058823529411765e-05,
|
| 436 |
+
"loss": 0.2438,
|
| 437 |
+
"step": 1250
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 7.500736377025037,
|
| 441 |
+
"grad_norm": 27.353788375854492,
|
| 442 |
+
"learning_rate": 1.3895424836601308e-05,
|
| 443 |
+
"loss": 0.2814,
|
| 444 |
+
"step": 1275
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 7.648011782032401,
|
| 448 |
+
"grad_norm": 199.70655822753906,
|
| 449 |
+
"learning_rate": 1.373202614379085e-05,
|
| 450 |
+
"loss": 0.5892,
|
| 451 |
+
"step": 1300
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 7.795287187039764,
|
| 455 |
+
"grad_norm": 98.0313720703125,
|
| 456 |
+
"learning_rate": 1.3568627450980393e-05,
|
| 457 |
+
"loss": 0.3041,
|
| 458 |
+
"step": 1325
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 7.942562592047128,
|
| 462 |
+
"grad_norm": 7.768338203430176,
|
| 463 |
+
"learning_rate": 1.3405228758169935e-05,
|
| 464 |
+
"loss": 0.2861,
|
| 465 |
+
"step": 1350
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 8.0,
|
| 469 |
+
"eval_accuracy": 0.8145580589254766,
|
| 470 |
+
"eval_loss": 0.945428729057312,
|
| 471 |
+
"eval_macro_f1": 0.8153678382686017,
|
| 472 |
+
"eval_runtime": 0.4264,
|
| 473 |
+
"eval_samples_per_second": 1353.065,
|
| 474 |
+
"eval_steps_per_second": 86.765,
|
| 475 |
+
"eval_weighted_f1": 0.8155204896708066,
|
| 476 |
+
"step": 1360
|
| 477 |
+
}
|
| 478 |
+
],
|
| 479 |
+
"logging_steps": 25,
|
| 480 |
+
"max_steps": 3400,
|
| 481 |
+
"num_input_tokens_seen": 0,
|
| 482 |
+
"num_train_epochs": 20,
|
| 483 |
+
"save_steps": 500,
|
| 484 |
+
"stateful_callbacks": {
|
| 485 |
+
"EarlyStoppingCallback": {
|
| 486 |
+
"args": {
|
| 487 |
+
"early_stopping_patience": 3,
|
| 488 |
+
"early_stopping_threshold": 0.0
|
| 489 |
+
},
|
| 490 |
+
"attributes": {
|
| 491 |
+
"early_stopping_patience_counter": 3
|
| 492 |
+
}
|
| 493 |
+
},
|
| 494 |
+
"TrainerControl": {
|
| 495 |
+
"args": {
|
| 496 |
+
"should_epoch_stop": false,
|
| 497 |
+
"should_evaluate": false,
|
| 498 |
+
"should_log": false,
|
| 499 |
+
"should_save": true,
|
| 500 |
+
"should_training_stop": true
|
| 501 |
+
},
|
| 502 |
+
"attributes": {}
|
| 503 |
+
}
|
| 504 |
+
},
|
| 505 |
+
"total_flos": 0.0,
|
| 506 |
+
"train_batch_size": 8,
|
| 507 |
+
"trial_name": null,
|
| 508 |
+
"trial_params": null
|
| 509 |
+
}
|
ood-vifactcheck-phobert-topic-domain-08-s44-20260826T112804Z-ab4112e3/trainer/checkpoint-1360/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-vifactcheck-phobert-topic-domain-08-s44-20260826T112804Z-ab4112e3/trainer/checkpoint-850/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|