Jerry02/whisper-tiny_to_canadian_accent_2

Whisper tiny Canadian

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

• Loss: 0.5672
• Wer: 23.7608

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: 6
• 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: 6000

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.0755	2.9940	1000	0.4602	128.9274
0.0053	5.9880	2000	0.5045	21.1907
0.0007	8.9820	3000	0.5257	22.6855
0.0004	11.9760	4000	0.5463	23.0003
0.0003	14.9701	5000	0.5598	22.8691
0.0003	17.9641	6000	0.5672	23.7608

Framework versions

• Transformers 4.50.0.dev0
• Pytorch 2.6.0+cu124
• Datasets 3.4.1
• Tokenizers 0.21.1