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

Loss: 0.6323
Wer: 19.3934

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: 256
eval_batch_size: 128
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
0.0014	44.2478	5000	0.4639	21.2055
0.0	88.4956	10000	0.5180	20.2500
0.0	132.7434	15000	0.6086	20.6645
0.0	176.9912	20000	0.7126	20.9749
0.0363	221.2389	25000	0.5293	25.0394
0.0	265.4867	30000	0.5856	19.5755
0.0	309.7345	35000	0.6323	19.3934
0.0	353.9823	40000	0.6857	19.4316
0.0	398.2301	45000	0.7216	19.6509
0.0	442.4779	50000	0.7511	20.0081
0.0	486.7257	55000	0.7904	20.4599
0.0	530.9735	60000	0.8208	20.8414

Framework versions

Transformers 4.52.3
Pytorch 2.6.0+cu124
Datasets 3.6.0
Tokenizers 0.21.1

zuazo/whisper-tiny-eu-cv21.0

Whisper Tiny Basque

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

zuazo/
whisper-tiny-eu-cv21.0