yoona-J/ASR_Whisper_Cerebral_Palsy

ASR_Whisper_Degenerative_Brain_Argumentation

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

• Loss: 0.4547
• Cer: 19.5682
• Wer: 26.4223

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: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• optimizer: Use OptimizerNames.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: 320
• training_steps: 3232
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Cer	Wer
0.9359	0.9913	400	0.9269	30.1532	43.4307
0.4058	1.9814	800	0.6448	30.0489	48.1589
0.1813	2.9715	1200	0.4917	34.0008	37.9538
0.1193	3.9616	1600	0.4691	28.8523	34.5848
0.0461	4.9517	2000	0.4717	19.7738	27.4559
0.0257	5.9418	2400	0.4440	19.5613	27.2727
0.01	6.9318	2800	0.4532	20.5277	27.4314
0.01	7.9219	3200	0.4547	19.5682	26.4223

Framework versions

• Transformers 4.53.0.dev0
• Pytorch 2.6.0+cu124
• Datasets 3.6.0
• Tokenizers 0.21.1