CoolFace
Modelpublic

bijaykumarsingh/whisper-turbo-as-LDCIL-sentence-Aligned_Dataset

sourceHugging Facemitupdated 7mo agoView on Hugging Face
0likes6downloads
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-turbo-as-LDCIL-sentence-Aligned_Dataset

This model is a fine-tuned version of openai/whisper-large-v3-turbo on the audiofolder dataset. It achieves the following results on the evaluation set:

  • —Loss: 0.0987
  • —Wer: 33.2956

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
  • —optimizer: Use adamwtorch with betas=(0.9,0.999) and epsilon=1e-08 and optimizerargs=No additional optimizer arguments
  • —lrschedulertype: constantwithwarmup
  • —lrschedulerwarmup_steps: 500
  • —training_steps: 5000

Training results

Training LossEpochStepValidation LossWer
0.19140.46045000.198154.3022
0.12910.920810000.134843.5821
0.10411.381215000.119639.9390
0.09351.841620000.102736.7426
0.06532.302025000.105636.7208
0.06192.762430000.101235.1193
0.04323.222835000.103434.7837
0.0463.683240000.100234.4787
0.03354.143645000.096833.0341
0.03544.604150000.098733.2956

Framework versions

  • —Transformers 4.46.3
  • —Pytorch 2.4.0+cu121
  • —Datasets 3.1.0
  • —Tokenizers 0.20.3