numan98/whisper-tiny-nextayah-v6

<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->

Whisper Nextayah Tiny

This model is a fine-tuned version of tarteel-ai/whisper-tiny-ar-quran on the Synth verses dataset. It achieves the following results on the evaluation set:

• —eval_loss: 0.1000
• —eval_wer: 20.6497
• —eval_runtime: 125.1049
• —evalsamplesper_second: 2.638
• —evalstepsper_second: 0.336
• —epoch: 12.3656
• —step: 2300

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: 0.0001
• —trainbatchsize: 16
• —evalbatchsize: 8
• —seed: 42
• —optimizer: Use OptimizerNames.ADAMWTORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizerargs=No additional optimizer arguments
• —lrschedulertype: linear
• —lrschedulerwarmup_steps: 100
• —num_epochs: 15
• —mixedprecisiontraining: Native AMP

Framework versions

• —Transformers 4.48.1
• —Pytorch 2.5.1+cu121
• —Datasets 3.2.0
• —Tokenizers 0.21.0