Raghad-DD/whisper-tiny-informal-arabic

Whisper Tiny Informal Arabic

This model is a fine-tuned version of openai/whisper-tiny on the Informal Arabic dataset. It achieves the following results on the evaluation set:

• Loss: 1.1448
• Wer: 63.7075

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 OptimizerNames.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
0.5883	4.0650	1000	0.8899	66.7433
0.2235	8.1301	2000	0.9200	63.0635
0.0966	12.1951	3000	1.0207	63.3211
0.0389	16.2602	4000	1.1064	63.1555
0.0275	20.3252	5000	1.1448	63.7075

Framework versions

• Transformers 4.48.3
• Pytorch 2.5.1+cu124
• Datasets 3.2.0
• Tokenizers 0.21.0