whisper-small-Ypause

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

• Loss: 0.2932
• Cer: 7.7185

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: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• 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
• num_epochs: 2
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Cer
0.2477	0.1290	100	0.2899	7.8184
0.1437	0.2581	200	0.2799	7.4424
0.1141	0.3871	300	0.2833	7.3954
0.1071	0.5161	400	0.2836	7.7244
0.081	0.6452	500	0.2787	7.6010
0.0918	0.7742	600	0.2810	7.6774
0.0844	0.9032	700	0.2767	7.4601
0.0317	1.0323	800	0.2934	7.9065
0.0269	1.1613	900	0.2780	7.5247
0.0283	1.2903	1000	0.2831	7.6304
0.0301	1.4194	1100	0.2997	7.7890
0.0207	1.5484	1200	0.3091	7.9241
0.0227	1.6774	1300	0.2911	7.6480
0.0251	1.8065	1400	0.2917	7.7361
0.0231	1.9355	1500	0.2932	7.7185

Framework versions

• Transformers 4.46.0.dev0
• Pytorch 2.4.1+cu121
• Datasets 3.0.1
• Tokenizers 0.20.0