whisper-large-v3-ar-resturant-11.2

This model is a fine-tuned version of openai/whisper-large-v3 on the heikal/arabic_call_splitted_8,9 and 10 dataset. It achieves the following results on the evaluation set:

• Loss: 2.0517
• Wer: 70.1149

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: 5e-06
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• 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: 750
• training_steps: 1000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.0134	125.0	500	1.6941	67.8161
0.0004	250.0	1000	2.0517	70.1149

Framework versions

• Transformers 4.47.0
• Pytorch 2.5.1+cu121
• Datasets 3.1.0
• Tokenizers 0.21.0