This model is a fine-tuned version of openai/whisper-small on the common_voice_21_0_eu dataset. It achieves the following results on the evaluation set:

Loss: 0.3205
Wer: 11.0571

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: 3.75e-05
train_batch_size: 64
eval_batch_size: 32
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: 100000
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.0076	11.1111	5000	0.2608	11.6163
0.0049	22.2222	10000	0.2970	11.5374
0.0042	33.3333	15000	0.3066	11.0996
0.0022	44.4444	20000	0.3205	11.0571
0.0021	55.5556	25000	0.3351	11.2886
0.0017	66.6667	30000	0.3326	11.1897
0.0016	77.7778	35000	0.3439	11.1412
0.0019	88.8889	40000	0.3533	11.2487
0.001	100.0	45000	0.3681	11.4507

Framework versions

Transformers 4.52.3
Pytorch 2.6.0+cu124
Datasets 3.6.0
Tokenizers 0.21.1

zuazo/whisper-small-eu-cv21.0

Whisper Small Basque

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

zuazo/
whisper-small-eu-cv21.0