Noursene/whisper-tiny-2000

Whisper Small Eng - noursene

This model is a fine-tuned version of openai/whisper-small-2000 on the medical audio trascription dataset. It achieves the following results on the evaluation set:

• Loss: 0.1612
• Wer: 10.5366

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: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• training_steps: 2000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.191	3.0257	500	0.2308	14.0727
0.0375	6.0514	1000	0.1570	10.9975
0.0045	9.0772	1500	0.1594	10.7598
0.0029	12.1029	2000	0.1612	10.5366

Framework versions

• Transformers 4.40.2
• Pytorch 2.2.1+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1