Ahmed107/distill-ar

Hamsa distill alfa

This model is a fine-tuned version of distil-whisper/distil-large-v2 on the nadsoft/Jordan-Audio dataset. It achieves the following results on the evaluation set:

• Loss: 0.9732
• Wer Ortho: 47.5105
• Wer: 45.2234

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: 0.0002
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: constant_with_warmup
• lr_scheduler_warmup_steps: 50
• training_steps: 8000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer Ortho	Wer
0.2094	7.04	2000	0.8198	48.5575	46.3918
0.0883	14.08	4000	0.9112	47.4174	44.6048
0.0662	21.13	6000	0.9644	46.8125	44.6277
0.0496	28.17	8000	0.9732	47.5105	45.2234

Framework versions

• Transformers 4.35.0
• Pytorch 2.1.0+cu118
• Datasets 2.14.6
• Tokenizers 0.14.1