StephaneBah/whisper-small-rad-FR4_Combo

Whisper Small FR - Radiologie

This model is a fine-tuned version of openai/whisper-small on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.7673
• Wer: 32.0536

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: 3e-05
• train_batch_size: 8
• eval_batch_size: 6
• seed: 3407
• optimizer: Use OptimizerNames.ADAMW_8BIT with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• training_steps: 1000

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
No log	6.25	100	0.7878	100.0
No log	12.5	200	0.7400	39.0006
No log	18.75	300	0.7597	33.4552
No log	25.0	400	0.7574	32.6630
0.0558	31.25	500	0.7604	32.6630
0.0558	37.5	600	0.7635	32.7239
0.0558	43.75	700	0.7639	32.5411
0.0558	50.0	800	0.7660	32.6021
0.0558	56.25	900	0.7668	32.7239
0.0003	62.5	1000	0.7673	32.0536

Framework versions

• Transformers 4.51.3
• Pytorch 2.6.0+cu124
• Datasets 3.6.0
• Tokenizers 0.21.2