jeff7522553/whisper-small-gy

Whisper small TW - from gy

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

• Loss: 0.0166
• Wer: 16.8640

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: Use OptimizerNames.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: 5000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.0046	3.4965	1000	0.0193	19.7097
0.0007	6.9930	2000	0.0161	17.3797
0.0001	10.4895	3000	0.0163	16.7876
0.0001	13.9860	4000	0.0165	16.8640
0.0001	17.4825	5000	0.0166	16.8640

Framework versions

• Transformers 4.50.3
• Pytorch 2.9.1+cu128
• Datasets 4.4.1
• Tokenizers 0.21.4