whisper-small-pashto

This model is a fine-tuned version of openai/whisper-small on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.4625
• Wer: 30.8805

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
0.008	0.9901	100	0.4411	28.5655
0.0058	1.9802	200	0.4440	28.3175
0.0073	2.9703	300	0.4664	28.9376
0.0095	3.9604	400	0.4832	30.6738
0.0148	4.9505	500	0.4625	30.8805

Framework versions

• Transformers 4.47.0
• Pytorch 2.2.0
• Datasets 3.2.0
• Tokenizers 0.21.0