rcastrovexler/whisper-small-es-pr

Whisper Small ES-PR - Roberto Castro-Vexler

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

• Loss: 0.1820
• Wer: 6.1053

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 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
0.0001	37.0370	1000	0.1530	5.8947
0.0001	74.0741	2000	0.1689	6.1053
0.0	111.1111	3000	0.1783	6.0526
0.0	148.1481	4000	0.1820	6.1053

Framework versions

• Transformers 4.48.3
• Pytorch 2.8.0+cu126
• Datasets 3.6.0
• Tokenizers 0.21.4