momo103197/whisper-small-zh-TW-mix

Whisper Small zh-TW - Fine-tune-mix

This model is a fine-tuned version of openai/whisper-small on the Common Voice 15, 16, 16_1 dataset. It achieves the following results on the evaluation set:

• Loss: 0.2202
• Cer: 9.9010

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: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• 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
0.1194	0.4429	1000	0.2126	10.9873
0.0496	0.8857	2000	0.2120	11.1762
0.0075	1.3286	3000	0.2199	10.1705
0.0058	1.7715	4000	0.2202	9.9010

Framework versions

• Transformers 4.48.3
• Pytorch 2.5.1+cu121
• Datasets 3.2.0
• Tokenizers 0.21.0