whisper-tiny-ft-btb-cv-cvad-ca-cy-2506

This model is a fine-tuned version of openai/whisper-tiny on the techiaith/banc-trawsgrifiadau-bangor train 25.06, techiaith/commonvoice_18_0_cy train+dev+other_with_excluded main, cymen-arfor/lleisiau-arfor train+dev main, techiaith/commonvoice_vad_cy train main dataset. It achieves the following results on the evaluation set:

• Loss: 0.7215
• Wer: 0.4487

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: 16
• 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: 500
• training_steps: 5000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.6084	2.4450	1000	0.7096	0.5239
0.4206	4.8900	2000	0.6443	0.4438
0.2773	7.3350	3000	0.6679	0.4481
0.2302	9.7800	4000	0.6983	0.4484
0.1771	12.2249	5000	0.7215	0.4487

Framework versions

• Transformers 4.52.4
• Pytorch 2.7.0+cu126
• Datasets 3.6.0
• Tokenizers 0.21.1