whisper-small-dv

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

• Loss: 0.2738
• Wer Ortho: 56.8842
• Wer: 11.0721

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: 3000

Training results

Training Loss	Epoch	Step	Validation Loss	Wer Ortho	Wer
0.1237	1.63	500	0.1702	63.0963	13.1429
0.0494	3.26	1000	0.1662	57.8592	11.6285
0.0315	4.89	1500	0.1894	58.3397	11.3937
0.0121	6.51	2000	0.2257	57.7756	11.5502
0.005	8.14	2500	0.2643	56.9747	11.1573
0.0056	9.77	3000	0.2738	56.8842	11.0721

Framework versions

• Transformers 4.31.0
• Pytorch 2.0.1+cu118
• Datasets 2.14.0
• Tokenizers 0.13.3