whisper-small-jv

This model is a fine-tuned version of openai/whisper-small on the "OpenSLR High quality TTS data for Javanese" dataset. It achieves the following results on the evaluation set:

• Loss: 0.3461
• Wer: 26.7960

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-06
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• training_steps: 5000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.4523	3.0488	1000	0.5512	37.2213
0.244	6.0976	2000	0.4139	30.4913
0.1597	9.1463	3000	0.3683	28.4682
0.126	12.1951	4000	0.3506	27.4773
0.1098	15.2439	5000	0.3461	26.7960

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.0+cu121
• Datasets 3.0.0
• Tokenizers 0.19.1