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

Loss: 0.5741
Cer: 16.7548
Wer: 19.9232

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: 4
total_train_batch_size: 32
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: 280
training_steps: 2828
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Cer	Validation Loss	Wer
0.5726	1.9814	400	22.8761	0.7755	28.2161
0.1687	3.9616	800	19.4152	0.6453	23.4518
0.0637	5.9418	1200	18.9512	0.6946	22.7384
0.0215	7.9219	1600	18.6748	0.6044	22.2010
0.0132	9.9021	2000	17.9068	0.6177	21.3079
0.0085	11.8872	2400	18.1934	0.5656	21.7711
0.0071	13.8674	2800	16.7548	0.5741	19.9232

Framework versions

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

yoona-J/ASR_Whisper_Cerebral_Palsy_Aug

ASR_Whisper_Cerebral_Palsy_Arg

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

yoona-J/
ASR_Whisper_Cerebral_Palsy_Aug