CoolFace
Modelpublic

Finnish-NLP/whisper-tiny-finnish

sourceHugging Faceapache-2.0updated 3y agoView on Hugging Face
0likes26downloads
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. -->

WhisperTinyFinnishV3

This model is a fine-tuned version of openai/whisper-tiny on the commonvoice11_0 dataset. It achieves the following results on the evaluation set:

  • —Loss: 0.5363
  • —Wer: 45.1376

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-06
  • —trainbatchsize: 32
  • —evalbatchsize: 4
  • —seed: 42
  • —optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • —lrschedulertype: linear
  • —lrschedulerwarmup_steps: 1000
  • —training_steps: 10000
  • —mixedprecisiontraining: Native AMP

Training results

Training LossEpochStepValidation LossWer
0.92360.110000.778358.5187
0.7270.220000.663853.1097
0.68670.330000.611350.2639
0.83480.440000.588248.2661
0.51650.550000.567947.1259
0.55090.660000.554046.6359
0.6390.770000.546646.5228
0.47150.880000.540045.9763
0.63060.990000.536345.1376
0.45981.0100000.535245.4768

Framework versions

  • —Transformers 4.36.2
  • —Pytorch 2.1.0+cu121
  • —Datasets 2.16.1
  • —Tokenizers 0.15.0