StephaneBah/whisper-small-rad-fr1.1

Whisper Small Fr - Radiologie1.1

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: 0.8172
• Wer: 34.6740

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: 5000

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.0594	31.25	500	0.7913	99.4516
0.0002	62.5	1000	0.7987	99.4516
0.0002	93.75	1500	0.8020	41.0116
0.0001	125.0	2000	0.8071	35.1005
0.0001	156.25	2500	0.8105	35.3443
0.0001	187.5	3000	0.8122	34.9787
0.0001	218.75	3500	0.8153	35.1615
0.0001	250.0	4000	0.8154	34.6130
0.0001	281.25	4500	0.8162	34.9787
0.0001	312.5	5000	0.8172	34.6740

Framework versions

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