whisper-base-ft-btb-ca-ec-cv-cy-en

This model is a fine-tuned version of openai/whisper-base on the DewiBrynJones/banc-trawsgrifiadau-bangor-clean train main, cymen-arfor/15awr train+dev+test main, wanasash/enwaucymraeg train+dev+test main, DewiBrynJones/commonvoice_18_0_cy_en train main dataset. It achieves the following results on the evaluation set:

• Loss: 0.7135
• Wer: 0.5089

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: 16
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• 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
1.7363	0.3260	1000	0.9845	0.6733
1.4155	0.6520	2000	0.8237	0.5717
1.3809	0.9780	3000	0.7566	0.5274
1.0882	1.3040	4000	0.7265	0.5286
1.0903	1.6300	5000	0.7135	0.5089

Framework versions

• Transformers 4.46.1
• Pytorch 2.5.1+cu124
• Datasets 3.1.0
• Tokenizers 0.20.1