sujithvemi/whisper-medium-physician-dictation-gpt-4-turbo

Whisper Large v3 Physician Dictation GPT 4 turbo

This model is a fine-tuned version of openai/whisper-medium on the Physician Dictation GPT 4 Turbo dataset. It achieves the following results on the evaluation set:

• Loss: 0.1587
• Wer: 5.2149

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: 4
• 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: 8500
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.0314	3.9683	500	0.1036	4.6175
0.0088	7.9365	1000	0.1090	5.0712
0.0009	11.9048	1500	0.1229	5.1880
0.0013	15.8730	2000	0.1329	5.6506
0.0031	19.8413	2500	0.1360	5.3227
0.0004	23.8095	3000	0.1288	4.9589
0.0008	27.7778	3500	0.1381	4.9724
0.0001	31.7460	4000	0.1391	5.0667
0.0001	35.7143	4500	0.1435	5.4710
0.0001	39.6825	5000	0.1445	5.2688
0.0002	43.6508	5500	0.1490	5.1880
0.0	47.6190	6000	0.1510	5.3093
0.0	51.5873	6500	0.1540	5.2868
0.0	55.5556	7000	0.1559	5.3452
0.0	59.5238	7500	0.1573	5.2104
0.0	63.4921	8000	0.1583	5.2239
0.0	67.4603	8500	0.1587	5.2149

Framework versions

• Transformers 4.40.2
• Pytorch 2.0.1+cu117
• Datasets 2.19.1
• Tokenizers 0.19.1