leliw/whisper-small-pl

Whisper Small PL

This model is a fine-tuned version of whisper-small-pl on the Common Voice 17.0 dataset. It achieves the following results on the evaluation set:

• Loss: 0.5061
• Wer: 120.2622

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: 50
• eval_batch_size: 8
• seed: 42
• optimizer: Use 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: 100
• training_steps: 1000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.0329	0.2	200	0.5055	105.0792
0.0276	0.4	400	0.4741	116.0295
0.0147	1.185	600	0.6035	97.4604
0.0091	1.385	800	0.6017	106.7723
0.0129	2.17	1000	0.5061	120.2622

Framework versions

• Transformers 4.49.0
• Pytorch 2.6.0+cu124
• Datasets 3.3.1
• Tokenizers 0.21.0