datdo2717/11_2

Whisper Small Ori vi

This model is a fine-tuned version of openai/whisper-small on the Common Voice 11.0 dataset. It achieves the following results on the evaluation set:

• Loss: 0.4021
• Wer: 15.2519

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: 16
• eval_batch_size: 8
• seed: 42
• 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: 200
• training_steps: 1000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.458	0.2222	100	0.4649	16.8154
0.4314	0.4444	200	0.4266	16.4319
0.4275	0.6667	300	0.4166	15.5542
0.3946	0.8889	400	0.4107	15.5764
0.2151	1.1111	500	0.4051	15.5616
0.2383	1.3333	600	0.4014	15.3551
0.2176	1.5556	700	0.3979	15.5395
0.2271	1.7778	800	0.3996	15.2371
0.222	2.0	900	0.3966	15.4141
0.1469	2.2222	1000	0.4021	15.2519

Framework versions

• Transformers 4.46.3
• Pytorch 2.4.0
• Datasets 3.1.0
• Tokenizers 0.20.0