whisper-large-v3-ft-fz-cv-cy

This model is a fine-tuned version of openai/whisper-large-v3 on the DewiBrynJones/commonvoice_18_0_cy default dataset. It achieves the following results on the evaluation set:

• Loss: 0.5067
• Wer: 0.2715

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: Adam with betas=(0.9,0.999) and epsilon=1e-08
• 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
0.1024	4.0161	1000	0.3592	0.2914
0.0052	8.0321	2000	0.4336	0.2667
0.0014	12.0482	3000	0.4721	0.2708
0.0006	16.0643	4000	0.4972	0.2700
0.0005	20.0803	5000	0.5067	0.2715

Framework versions

• Transformers 4.44.0
• Pytorch 2.4.0+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1