whisper-medium-en-cv-4.0

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

• Loss: 0.4337
• Wer: 13.7048

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: 5e-06
• train_batch_size: 64
• eval_batch_size: 64
• seed: 42
• optimizer: Use OptimizerNames.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: 120
• training_steps: 1125
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.2967	0.2	225	0.4531	14.3072
0.2258	0.4	450	0.4317	13.4538
0.1697	1.0996	675	0.4308	13.0020
0.1653	1.2996	900	0.4358	13.7550
0.1394	1.4996	1125	0.4337	13.7048

Framework versions

• Transformers 4.51.3
• Pytorch 2.6.0+cu124
• Datasets 3.5.0
• Tokenizers 0.21.1