outputs

This model is a fine-tuned version of KoichiYasuoka/bert-base-japanese-char-extended on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 2.2432
  • F1: 0.1451
  • Precision: 0.1495
  • Recall: 0.1410
  • Accuracy: 0.1914
  • F1 Art: 0.2010
  • F1 Conj: 0.1410
  • F1 Dj: 0.0
  • F1 Dp: 0.0606
  • F1 Dv: 0.0
  • F1 Erb: 0.0
  • F1 Et: 0.0513
  • F1 Ntj: 0.0273
  • F1 Oun: 0.0
  • F1 Ron: 0.1128
  • F1 Ropn: 0.0102
  • F1 Um: 0.0400
  • F1 Unct: 0.6066
  • F1 Ux: 0.0932
  • F1 Ym: 0.0296
  • F1 : 0.0

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

  • learning_rate: 3e-05
  • train_batch_size: 32
  • eval_batch_size: 64
  • seed: 42
  • gradient_accumulation_steps: 2
  • total_train_batch_size: 64
  • optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_ratio: 0.06
  • num_epochs: 10

Training results

Training Loss Epoch Step Validation Loss F1 Precision Recall Accuracy F1 Art F1 Conj F1 Dj F1 Dp F1 Dv F1 Erb F1 Et F1 Ntj F1 Oun F1 Ron F1 Ropn F1 Um F1 Unct F1 Ux F1 Ym F1
2.3878 1.0 638 2.4024 0.1351 0.1440 0.1272 0.2013 0.1619 0.1305 0.0061 0.0106 0.0 0.0 0.0813 0.0377 0.0 0.0 0.0191 0.0567 0.5886 0.1625 0.0446 0.0
2.3651 2.0 1276 2.3474 0.1368 0.1418 0.1321 0.1769 0.1957 0.1369 0.0 0.0228 0.0 0.0 0.0250 0.0253 0.0139 0.1175 0.0 0.0205 0.5886 0.0 0.0343 0.0
2.3634 3.0 1914 2.3653 0.1380 0.1403 0.1358 0.1947 0.1914 0.1278 0.0 0.0248 0.0 0.0 0.0783 0.0780 0.0 0.0995 0.0118 0.0 0.5886 0.1129 0.0627 0.0
2.3592 4.0 2552 2.3568 0.1356 0.1365 0.1346 0.1774 0.1776 0.1375 0.0061 0.0282 0.0 0.0 0.0729 0.0283 0.0 0.0966 0.0037 0.0229 0.5886 0.1038 0.0470 0.0
2.3378 5.0 3190 2.3542 0.1386 0.1498 0.1289 0.2005 0.1914 0.1416 0.0 0.0168 0.0 0.0 0.0302 0.0335 0.0057 0.1079 0.0118 0.0161 0.5886 0.0158 0.0182 0.0
2.388 6.0 3828 2.3583 0.1413 0.1578 0.1279 0.2129 0.1672 0.1466 0.0 0.0183 0.0 0.0552 0.0586 0.0349 0.0 0.0995 0.0 0.0229 0.5886 0.1217 0.0391 0.0
2.3177 7.0 4466 2.3540 0.1389 0.1523 0.1277 0.2005 0.1672 0.1434 0.0 0.0026 0.0 0.0 0.0 0.0234 0.0 0.1126 0.0291 0.0229 0.5886 0.1179 0.0182 0.0
2.3434 8.0 5104 2.3745 0.1349 0.1418 0.1286 0.1984 0.1776 0.1341 0.0 0.0006 0.0 0.0 0.0855 0.0402 0.0057 0.1079 0.0118 0.0437 0.5886 0.1038 0.0325 0.0
2.3773 9.0 5742 2.3556 0.1386 0.1444 0.1332 0.1920 0.1914 0.1399 0.0 0.0468 0.0 0.0 0.0583 0.0377 0.0 0.0995 0.0118 0.0437 0.5886 0.0363 0.0390 0.0
2.3234 10.0 6380 2.3626 0.1416 0.1468 0.1367 0.1930 0.1776 0.1385 0.0 0.0617 0.0 0.0 0.0549 0.0377 0.0 0.1079 0.0118 0.0437 0.5886 0.0794 0.0390 0.0

Framework versions

  • Transformers 4.53.3
  • Pytorch 2.6.0+cu124
  • Datasets 4.1.1
  • Tokenizers 0.21.2
Downloads last month
2
Safetensors
Model size
90.3M params
Tensor type
F32
ยท
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support

Model tree for HuyHoang1977/outputs