CoolFace
Modelpublic

alikanakar/whisper-synthesized-turkish-4-hour-llr

sourceHugging Faceapache-2.0updated 3y agoView on Hugging Face
0likes8downloads
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-synthesized-turkish-4-hour-llr

This model is a fine-tuned version of openai/whisper-small on the None dataset. It achieves the following results on the evaluation set:

  • —Loss: 0.2356
  • —Wer: 14.9179

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-06
  • —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 LossWer
1.82781.041001.535741.2110
1.20482.082000.805986.3165
0.62063.123000.575194.1612
0.49274.174000.502480.1568
0.42775.215000.445148.6607
0.35026.256000.399932.7490
0.29487.297000.365021.5961
0.248.338000.334417.7756
0.20799.389000.301015.2512
0.141610.4210000.240213.8687
0.107611.4611000.235714.4427
0.090512.512000.235014.0415
0.075913.5413000.234814.0230
0.081914.5814000.234514.0600
0.060815.6215000.234614.3563
0.071216.6716000.235014.2575
0.060817.7117000.235514.8562
0.063218.7518000.235414.9241
0.054419.7919000.235614.9303
0.052920.8320000.235614.9179

Framework versions

  • —Transformers 4.28.0
  • —Pytorch 2.0.0+cu118
  • —Datasets 2.11.0
  • —Tokenizers 0.13.3