Mohamedshaaban2001/MSDC-whisper-base

Whisper Small ar1 - Mohamed Shaaban

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

• Loss: 0.4585
• Wer: 65.2720

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: Adam with betas=(0.9,0.999) and epsilon=1e-08
• 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
0.444	0.42	1000	0.5684	73.7587
0.4161	0.83	2000	0.4995	68.0147
0.3282	1.25	3000	0.4841	68.92
0.2915	1.66	4000	0.4663	67.6120
0.2639	2.08	5000	0.4585	65.2720

Framework versions

• Transformers 4.39.3
• Pytorch 2.1.2
• Datasets 2.18.0
• Tokenizers 0.15.2