Clo123/whisper-tiny_to_american_accent_finetune

Whisper tiny American fine tuned 2

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

• Loss: 0.1867
• Wer: 9.0329

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: 2
• eval_batch_size: 1
• seed: 42
• 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: 2000

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.2838	0.2363	250	0.2934	16.2731
0.2524	0.4726	500	0.2283	11.3946
0.2339	0.7089	750	0.2165	10.7740
0.16	0.9452	1000	0.2035	10.8602
0.0787	1.1815	1250	0.1957	10.2741
0.0521	1.4178	1500	0.1949	10.2913
0.1923	1.6541	1750	0.1888	9.1881
0.0606	1.8904	2000	0.1867	9.0329

Framework versions

• Transformers 4.49.0
• Pytorch 2.6.0+cu124
• Datasets 3.4.1
• Tokenizers 0.21.1