heikal/whisper-large-v3-ar-resturant5

whisper large-v3 ar resturant 5

This model is a fine-tuned version of openai/whisper-large-v3 on the Arabic calls splitted dataset. It achieves the following results on the evaluation set:

• Loss: 1.8319
• Wer: 59.4828

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: 1000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.0005	250.0	1000	1.8319	59.4828

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.1+cu121
• Datasets 3.0.0
• Tokenizers 0.19.1