wme_30s_speed_20_1.0

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: 1.0985
• Wer: 31.6901

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: 4e-05
• train_batch_size: 48
• eval_batch_size: 32
• 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: 44
• training_steps: 264
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
No log	0	0	1.8678	47.3362
0.7443	0.3333	88	1.1949	34.2315
0.52	1.0038	176	1.1157	31.9045
0.2008	1.3371	264	1.0985	31.6901

Framework versions

• Transformers 4.51.3
• Pytorch 2.5.1+cu124
• Datasets 3.6.0
• Tokenizers 0.21.0