output

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: 0.5336
• Wer Ortho: 0.3166
• Wer: 0.3192

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: 3e-05
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: constant_with_warmup
• lr_scheduler_warmup_steps: 15
• training_steps: 90
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer Ortho	Wer
2.9987	0.5172	15	1.7184	0.4640	0.4170
0.7514	1.0345	30	0.5257	0.3790	0.3795
0.307	1.5517	45	0.5051	0.3269	0.3253
0.3075	2.0690	60	0.4907	0.3526	0.3518
0.1492	2.5862	75	0.5120	0.3095	0.3106
0.0719	3.1034	90	0.5336	0.3166	0.3192

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.1+cu121
• Datasets 3.0.1
• Tokenizers 0.19.1