deepinfinityai/v3_Large

Whisper Large Medical(11)

This model is a fine-tuned version of openai/whisper-medium on the 11 Sentences dataset. It achieves the following results on the evaluation set:

• Loss: 0.8179
• Wer: 35.7143

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: 4
• eval_batch_size: 8
• seed: 42
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 50
• training_steps: 500
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
1.5756	25.0	50	0.0563	7.1429
0.0009	50.0	100	0.0505	0.0
0.0001	75.0	150	0.0694	0.0
0.0001	100.0	200	0.1348	0.0
0.0001	125.0	250	0.2494	0.0
0.0	150.0	300	0.3976	0.0
0.0	175.0	350	0.5560	7.1429
0.0	200.0	400	0.6852	35.7143
0.0	225.0	450	0.7813	35.7143
0.0	250.0	500	0.8179	35.7143

Framework versions

• Transformers 4.47.1
• Pytorch 2.5.1+cu124
• Datasets 3.2.0
• Tokenizers 0.21.0