Edit model card

Visualize in Weights & Biases Visualize in Weights & Biases Visualize in Weights & Biases Visualize in Weights & Biases

whisper-small-th

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

  • Loss: 0.1596
  • Wer: 64.8535

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
  • gradient_accumulation_steps: 2
  • total_train_batch_size: 32
  • 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 Wer
0.2535 0.7294 1000 0.2177 73.9061
0.1453 1.4588 2000 0.1778 69.6909
0.0923 2.1882 3000 0.1648 65.8303
0.0781 2.9176 4000 0.1596 64.8535

Framework versions

  • Transformers 4.41.0
  • Pytorch 2.3.0+cu121
  • Datasets 2.19.1
  • Tokenizers 0.19.1
Downloads last month
17
Safetensors
Model size
242M params
Tensor type
F32
·

Finetuned from

Dataset used to train Porameht/whisper-small-th

Evaluation results