whisper-large-v3-DI-28.08.25

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

• Loss: 0.5691
• Wer: 22.8586

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: 2e-05
• train_batch_size: 32
• eval_batch_size: 16
• 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: 150
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
No log	1.0	88	0.4634	35.4412
No log	2.0	176	0.5425	35.6356
No log	3.0	264	0.5338	26.4351
No log	4.0	352	0.5240	37.7867
No log	5.0	440	0.5445	33.2901
0.2671	6.0	528	0.5544	30.2449
0.2671	7.0	616	0.5712	31.1131
0.2671	8.0	704	0.5839	24.4266
0.2671	9.0	792	0.5667	25.6836
0.2671	10.0	880	0.5691	22.8586

Framework versions

• Transformers 4.46.2
• Pytorch 2.4.1+cu121
• Datasets 3.1.0
• Tokenizers 0.20.0