Edit model card

whisper_finetune

This model is a fine-tuned version of openai/whisper-base on the aihub_3 dataset. It achieves the following results on the evaluation set:

  • Loss: 0.4807
  • Cer: 14.7381
  • Wer: 40.8215

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: 32
  • eval_batch_size: 16
  • 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: 4000
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Cer Wer
0.2038 0.4 500 0.4405 13.6475 39.1975
0.1892 0.8 1000 0.4491 14.5230 40.5892
0.1218 1.2 1500 0.4710 14.4216 40.2519
0.1227 1.6 2000 0.4879 14.3981 40.1969
0.1311 2.0 2500 0.4638 14.6655 40.9614
0.0945 2.4 3000 0.4783 14.6635 40.9190
0.0874 2.8 3500 0.4743 14.3360 40.4492
0.0759 3.2 4000 0.4807 14.7381 40.8215

Framework versions

  • Transformers 4.37.0.dev0
  • Pytorch 1.12.1+cu113
  • Datasets 2.15.0
  • Tokenizers 0.15.0
Downloads last month
2
Safetensors
Model size
72.6M params
Tensor type
F32
·

Finetuned from