Wishwa98/ASRForCommonVoice

Whisper Small for Common Accent

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

• Loss: 0.4234
• Wer Ortho: 17.9229
• Wer: 13.0605

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: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: constant_with_warmup
• lr_scheduler_warmup_steps: 50
• training_steps: 2500

Training results

Training Loss	Epoch	Step	Validation Loss	Wer Ortho	Wer
0.1012	1.14	500	0.3215	16.3784	11.5941
0.0345	2.28	1000	0.3483	16.6496	11.8450
0.018	3.42	1500	0.3829	17.1622	12.4707
0.0075	4.57	2000	0.4069	17.8667	13.0116
0.0059	5.71	2500	0.4234	17.9229	13.0605

Framework versions

• Transformers 4.33.0.dev0
• Pytorch 2.0.1+cu118
• Datasets 2.14.4
• Tokenizers 0.13.3