shg1421/whisper-small-decoder

Whisper

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

• Loss: 0.3409
• Wer: 10.5283

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: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• training_steps: 4000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.0305	4.55	1000	0.2471	11.1359
0.0117	9.09	2000	0.3168	10.3795
0.0024	13.64	3000	0.3312	10.4291
0.0006	18.18	4000	0.3409	10.5283

Framework versions

• Transformers 4.39.0.dev0
• Pytorch 1.12.1+cu102
• Datasets 2.15.0
• Tokenizers 0.15.0