whisper-medical-data

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

• Loss: 0.7069
• Wer: 23.3573

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: 100
• training_steps: 1000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
1.0106	4.17	100	1.1117	24.8717
0.043	8.33	200	0.5784	28.0287
0.0049	12.5	300	0.6418	23.7936
0.001	16.67	400	0.6730	23.2290
0.0005	20.83	500	0.6854	23.3573
0.0004	25.0	600	0.6938	23.3316
0.0003	29.17	700	0.6996	23.4343
0.0003	33.33	800	0.7034	23.3830
0.0002	37.5	900	0.7061	23.3573
0.0002	41.67	1000	0.7069	23.3573

Framework versions

• Transformers 4.38.2
• Pytorch 2.1.0+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2