whisper_ro_MilDB / README.md
AdaCodruta's picture
End of training
8bfffed verified
metadata
language:
  - ro
license: mit
base_model: openai/whisper-large-v3-turbo
tags:
  - generated_from_trainer
datasets:
  - mozilla-foundation/common_voice_17_0
metrics:
  - wer
model-index:
  - name: Whisper Large_v2 RO CV17
    results:
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: Common Voice 17.0
          type: mozilla-foundation/common_voice_17_0
          args: 'config: ro, split: test'
        metrics:
          - name: Wer
            type: wer
            value: 47.51015670342426

Whisper Large_v2 RO CV17

This model is a fine-tuned version of openai/whisper-large-v3-turbo on the Common Voice 17.0 dataset. It achieves the following results on the evaluation set:

  • Loss: 1.6152
  • Wer: 47.5102

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
  • train_batch_size: 16
  • eval_batch_size: 8
  • seed: 42
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_steps: 500
  • training_steps: 5000
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Wer
0.5004 5.4945 1000 1.1554 106.2565
0.0896 10.9890 2000 1.3810 51.0737
0.0121 16.4835 3000 1.5371 49.9013
0.0027 21.9780 4000 1.5901 49.1468
0.0008 27.4725 5000 1.6152 47.5102

Framework versions

  • Transformers 4.40.2
  • Pytorch 2.5.1+cu118
  • Datasets 3.1.0
  • Tokenizers 0.19.1