CoolFace
Modelpublic

utakumi/Hubert-kakeiken-W-5_impulses

sourceHugging Faceapache-2.0updated 2y agoView on Hugging Face
0likes5downloads
Model Card

<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->

Hubert-kakeiken-W-5_impulses

This model is a fine-tuned version of rinna/japanese-hubert-base on the ORIGINALKAKEIKENW5IMPULSES - JA dataset. It achieves the following results on the evaluation set:

  • —Loss: 0.0225
  • —Wer: 0.9994
  • —Cer: 1.0155

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
  • —trainbatchsize: 32
  • —evalbatchsize: 8
  • —seed: 42
  • —gradientaccumulationsteps: 2
  • —totaltrainbatch_size: 64
  • —optimizer: Use adamwtorch with betas=(0.9,0.999) and epsilon=1e-08 and optimizerargs=No additional optimizer arguments
  • —lrschedulertype: cosine
  • —lrschedulerwarmup_steps: 12500
  • —num_epochs: 40.0
  • —mixedprecisiontraining: Native AMP

Training results

Training LossEpochStepValidation LossWerCer
2.8191.040972.57801.01.1284
0.30962.081940.08990.99951.0198
0.21593.0122910.04830.99941.0215
0.18984.0163880.03100.99941.0181
0.16555.0204850.03410.99951.0194
0.15916.0245820.03020.99951.0180
0.14657.0286790.03710.99941.0186
0.14118.0327760.02460.99941.0176
0.12899.0368730.03050.99951.0184
0.130910.0409700.01950.99941.0161
0.128211.0450670.02420.99941.0183
0.119512.0491640.02930.99941.0176
0.118313.0532610.01710.99951.0162
0.112114.0573580.02010.99941.0159
0.10815.0614550.01760.99941.0162
0.109416.0655520.01580.99941.0163
0.105517.0696490.01990.99951.0163
0.099818.0737460.02280.99941.0159
0.099119.0778430.01750.99941.0163
0.097420.0819400.02310.99941.0164
0.090421.0860370.01410.99941.0155
0.088322.0901340.01770.99941.0156
0.088823.0942310.02110.99941.0156
0.083924.0983280.02050.99941.0161
0.083325.01024250.01930.99941.0155
0.08326.01065220.02020.99941.0156
0.0827.01106190.01950.99941.0156
0.075428.01147160.02280.99941.0157
0.076129.01188130.02620.99941.0162
0.072930.01229100.02450.99941.0158
0.071831.01270070.02660.99941.0162
0.071532.01311040.02150.99941.0154
0.070733.01352010.02310.99941.0155
0.068234.01392980.02280.99941.0156
0.067335.01433950.02280.99941.0155
0.066536.01474920.02380.99941.0155
0.065637.01515890.02190.99951.0155
0.067838.01556860.02240.99941.0154
0.065639.01597830.02270.99941.0155
0.065440.01638800.02240.99941.0154

Framework versions

  • —Transformers 4.48.0
  • —Pytorch 2.5.1+cu124
  • —Datasets 3.1.0
  • —Tokenizers 0.21.0