whisper-tiny-en-US

This model is a fine-tuned version of openai/whisper-tiny on the PolyAI/minds14 dataset. It achieves the following results on the evaluation set:

• Loss: 0.6597
• Wer Ortho: 0.3313
• Wer: 0.3194

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
• 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: constant_with_warmup
• lr_scheduler_warmup_steps: 50
• training_steps: 250
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer Ortho	Wer
4.3538	0.8929	25	2.7788	0.5355	0.4073
1.599	1.7857	50	0.7704	0.4176	0.3878
0.5122	2.6786	75	0.5936	0.3763	0.3560
0.3412	3.5714	100	0.5641	0.3516	0.3276
0.2051	4.4643	125	0.5758	0.3745	0.3471
0.1248	5.3571	150	0.5837	0.3529	0.3329
0.0725	6.25	175	0.6005	0.3541	0.3312
0.0413	7.1429	200	0.6264	0.3455	0.3282
0.0195	8.0357	225	0.6500	0.3387	0.3205
0.0085	8.9286	250	0.6597	0.3313	0.3194

Framework versions

• Transformers 4.48.0
• Pytorch 2.8.0+cu126
• Datasets 3.6.0
• Tokenizers 0.21.4