octava/whisper-small-id

Whisper Small Id - Inspirasi

This model is a fine-tuned version of openai/whisper-small on the Extracted id video v2 dataset. It achieves the following results on the evaluation set:

• Loss: 0.4480
• Wer: 28.1734

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: 500
• training_steps: 5000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.2601	0.5615	1000	0.3923	29.8060
0.1176	1.1230	2000	0.3954	30.3875
0.0848	1.6844	3000	0.4068	29.2758
0.0317	2.2459	4000	0.4088	26.8850
0.0261	2.8074	5000	0.4480	28.1734

Framework versions

• Transformers 4.49.0
• Pytorch 2.2.0a0+81ea7a4
• Datasets 3.3.2
• Tokenizers 0.21.0