CoolFace
Modelpublic

Eimhin03/irish-whisper-shunya-data-no-aug

sourceHugging Faceapache-2.0updated 7mo 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. -->

irish-whisper-shunya-data-no-aug

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

  • —Loss: 3.2185
  • —Wer: 65.0573

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: 0.0001
  • —trainbatchsize: 4
  • —evalbatchsize: 8
  • —seed: 42
  • —gradientaccumulationsteps: 4
  • —totaltrainbatch_size: 16
  • —optimizer: Use OptimizerNames.ADAMWTORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizerargs=No additional optimizer arguments
  • —lrschedulertype: linear
  • —lrschedulerwarmup_steps: 1000
  • —training_steps: 60000
  • —mixedprecisiontraining: Native AMP

Training results

Training LossEpochStepValidation LossWer
0.075228.089950002.523367.0559
0.034056.1798100002.910469.4476
0.035684.2697150003.094871.5024
0.0082112.3596200003.187969.9304
0.0067140.4494250003.217869.5262
0.0063168.5393300003.310272.7150
0.0000196.6292350003.357668.8075
0.0000224.7191400003.430967.1570
0.0000252.8090450003.509665.9219
0.0000280.8989500003.542665.8882
0.0000308.9888550003.314965.1471
0.0000337.0787600003.218565.0573

Framework versions

  • —Transformers 5.3.0.dev0
  • —Pytorch 2.9.0+cu126
  • —Datasets 4.0.0
  • —Tokenizers 0.22.2