Bateesa/whisper-small-jap

Whisper

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

• Loss: 0.7821
• Wer: 21.2078

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: 8
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED 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: 4000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.0025	24.3902	1000	0.7089	27.9655
0.0001	48.7805	2000	0.7556	21.1359
0.0000	73.1707	3000	0.7753	21.2078
0.0000	97.5610	4000	0.7821	21.2078

Framework versions

• Transformers 5.0.0
• Pytorch 2.10.0+cu128
• Datasets 4.0.0
• Tokenizers 0.22.2