mirodil/whisper-tiny-uzbek

Whisper Tiny Uzbek

This model is a fine-tuned version of openai/whisper-tiny on the mozilla-foundation/common_voice_13_0 dataset. It achieves the following results on the evaluation set:

• Loss: 0.2981
• Wer Ortho: 47.7812
• Wer: 36.7906

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: cosine_with_restarts
• lr_scheduler_warmup_steps: 50
• num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Wer Ortho	Wer
0.2929	0.8	3000	0.3281	50.8851	40.4395
0.2194	1.59	6000	0.3110	49.2325	37.9320
0.177	2.39	9000	0.3003	47.8700	36.8366
0.1574	3.18	12000	0.2997	48.2291	37.0491
0.1524	3.98	15000	0.2958	47.2395	36.4400
0.1455	4.77	18000	0.2981	47.7812	36.7906

Framework versions

• Transformers 4.35.1
• Pytorch 2.1.0
• Datasets 2.14.6
• Tokenizers 0.14.1