gcasey2's picture
End of training
d28be1d verified
metadata
license: apache-2.0
base_model: openai/whisper-large-v3
tags:
  - generated_from_trainer
metrics:
  - wer
model-index:
  - name: Whisper Large Korean/English
    results: []

Whisper Large Korean/English

This model is a fine-tuned version of openai/whisper-large-v3 on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 0.8019
  • Wer: 198.2263

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: 773
  • training_steps: 7728
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Wer
0.5546 1.0 773 0.5308 240.1463
0.3284 2.0 1546 0.5160 133.6395
0.176 3.0 2319 0.5582 264.5033
0.0977 4.0 3092 0.6110 155.6417
0.065 5.0 3865 0.6577 194.4118
0.0298 6.0 4638 0.7021 235.0691
0.0109 7.0 5411 0.7408 158.8282
0.0069 8.0 6184 0.7550 201.9574
0.0057 9.0 6957 0.8019 198.2263

Framework versions

  • Transformers 4.37.2
  • Pytorch 2.0.1
  • Datasets 2.17.0
  • Tokenizers 0.15.2