whisper-large-v3-ft-cy-en

This model is a fine-tuned version of openai/whisper-large-v3 on Welsh and English bilingual data originally from Mozilla's Common Voice dataset (see: techiaith/commonvoice_16_1_en_cy).

It achieves the following results on the evaluation set:

• Loss: 0.1480
• Wer: 25.1341

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: 4
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• training_steps: 4000

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.2078	0.25	1000	0.2198	28.7556
0.1623	0.5	2000	0.1800	31.3698
0.1417	0.75	3000	0.1585	18.7051
0.1188	1.01	4000	0.1480	25.1341

Framework versions

• Transformers 4.37.1
• Pytorch 2.1.2+cu121
• Datasets 2.16.1
• Tokenizers 0.15.1