CoolFace
Modelpublic

weewtr1121/outs

sourceHugging Faceapache-2.0updated 2y agoView on Hugging Face
0likes4downloads
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. -->

outs

This model is a fine-tuned version of biodatlab/whisper-th-medium-combined on an unknown dataset. It achieves the following results on the evaluation set:

  • —Loss: 0.0878
  • —Cer: 5.4469

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: 1e-05
  • —trainbatchsize: 16
  • —evalbatchsize: 16
  • —seed: 42
  • —gradientaccumulationsteps: 2
  • —totaltrainbatch_size: 32
  • —optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • —lrschedulertype: cosinewithrestarts
  • —lrschedulerwarmup_steps: 5000
  • —num_epochs: 15.0
  • —mixedprecisiontraining: Native AMP

Training results

Training LossEpochStepValidation LossCer
0.02750.9926670.08044.9648
0.03592.01350.08076.0688
0.03172.99262020.08135.8626
0.01964.02700.08194.9681
0.0214.99263370.08234.8683
0.02096.04050.08314.8783
0.02236.99264720.08665.0080
0.01738.05400.08645.2541
0.02318.99266070.08394.6422
0.013110.06750.08605.3405
0.013410.99267420.08995.2674
0.018312.08100.08735.8094
0.017212.99268770.08934.9215
0.017814.09450.08555.7096
0.012714.888910050.08785.4469

Framework versions

  • —Transformers 4.44.2
  • —Pytorch 2.4.1+cu121
  • —Datasets 3.0.1
  • —Tokenizers 0.19.1