santhosh643/whisper-medium-english

whisper_fine_tune_Santhosh

This model is a fine-tuned version of openai/whisper-medium on the Medical Speech, Transcription, and Intent dataset. It achieves the following results on the evaluation set:

• Loss: 0.0725
• Wer: 4.3103

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: 2
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 100
• training_steps: 600
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.5679	0.2825	100	0.2478	13.0980
0.1266	0.5650	200	0.1474	8.1003
0.0872	0.8475	300	0.1034	5.9266
0.0399	1.1299	400	0.0865	5.3507
0.0229	1.4124	500	0.0771	4.1709
0.0259	1.6949	600	0.0725	4.3103

Framework versions

• Transformers 4.45.0.dev0
• Pytorch 2.4.0+cu121
• Datasets 2.21.1.dev0
• Tokenizers 0.19.1