deepinfinityai/v3_02

Whisper Medium 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: 0.4603
• Wer: 44.0

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
2.2535	25.0	50	0.5853	32.0
0.004	50.0	100	0.4700	12.0
0.0	75.0	150	0.4688	16.0
0.0	100.0	200	0.4657	44.0
0.0	125.0	250	0.4635	44.0
0.0	150.0	300	0.4620	44.0
0.0	175.0	350	0.4607	44.0
0.0	200.0	400	0.4613	44.0
0.0	225.0	450	0.4600	44.0
0.0	250.0	500	0.4603	44.0

Framework versions

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