CoolFace
Modelpublic

yoona-J/speech_impediment_audio

sourceHugging Faceapache-2.0updated 2y agoView on Hugging Face
0likes7downloads
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. -->

Whisper_impediment

This model is a fine-tuned version of openai/whisper-base on the speechimpedimentaudio dataset. It achieves the following results on the evaluation set:

  • —Loss: 0.3906
  • —Cer: 14.8734

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: 8
  • —seed: 42
  • —optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • —lrschedulertype: linear
  • —lrschedulerwarmup_steps: 500
  • —training_steps: 2000
  • —mixedprecisiontraining: Native AMP

Training results

Training LossEpochStepValidation LossCer
0.042620.02000.289615.5063
0.001640.04000.322514.5570
0.000660.06000.344713.9241
0.000380.08000.358814.5570
0.0002100.010000.368614.5570
0.0002120.012000.376514.8734
0.0002140.014000.382714.8734
0.0001160.016000.386914.8734
0.0001180.018000.389614.8734
0.0001200.020000.390614.8734

Framework versions

  • —Transformers 4.45.0.dev0
  • —Pytorch 2.3.1+cu121
  • —Datasets 2.20.0
  • —Tokenizers 0.19.1