CoolFace
Modelpublic

samil24/whisper-small-serbian-v3

sourceHugging Faceapache-2.0updated 1y agoView on Hugging Face
0likes68downloads
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-small-serbian-v3

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: 0.4611
  • —Wer: 20.9326

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-05
  • —trainbatchsize: 16
  • —evalbatchsize: 16
  • —seed: 42
  • —gradientaccumulationsteps: 2
  • —totaltrainbatch_size: 32
  • —optimizer: Use OptimizerNames.ADAMWTORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizerargs=No additional optimizer arguments
  • —lrschedulertype: linear
  • —lrschedulerwarmup_steps: 1250
  • —num_epochs: 20
  • —mixedprecisiontraining: Native AMP

Training results

Training LossEpochStepValidation LossWer
0.34230.92345000.322327.8919
0.25881.845810000.268524.6306
0.18332.768215000.266523.3260
0.11543.690720000.271323.0167
0.07874.613125000.288122.8452
0.04565.535530000.316122.7374
0.0256.458035000.331822.6472
0.01517.380440000.365722.9566
0.00848.302945000.378622.1788
0.00639.225350000.387521.9738
0.003710.147755000.393222.0038
0.003111.070260000.410121.7705
0.003211.993565000.415221.8005
0.00312.916070000.417621.7935
0.001513.838475000.421321.5230
0.000814.760880000.437721.3339
0.000515.683385000.438921.2225
0.000216.605790000.444521.0935
0.000317.528295000.453621.0404
0.000118.4506100000.458920.9220
0.000119.3730105000.461120.9326

Framework versions

  • —Transformers 4.51.3
  • —Pytorch 2.6.0+cu124
  • —Datasets 3.6.0
  • —Tokenizers 0.21.4