Mohanad03/whisper-medium-arz-1

Whisper Medium Arz - All

This model is a fine-tuned version of openai/whisper-small on an unknown dataset. It achieves the following results on the evaluation set:

• eval_loss: 1.6689
• eval_wer: 61.0832
• eval_runtime: 898.9718
• eval_samples_per_second: 0.941
• eval_steps_per_second: 0.236
• epoch: 2.0
• step: 612

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: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 16
• 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: 10
• num_epochs: 2
• mixed_precision_training: Native AMP

Framework versions

• Transformers 4.49.0
• Pytorch 2.5.1+cu121
• Datasets 3.3.2
• Tokenizers 0.21.0