wav2vec2-base-arabic-finetuned-continued-v2

This model is a fine-tuned version of Mohammadawad1/wav2vec2-base-arabic-finetuned-continued-v1 on the common_voice_17_0 dataset. It achieves the following results on the evaluation set:

• Loss: 0.6287
• Wer: 0.5388

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: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 64
• 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: constant_with_warmup
• lr_scheduler_warmup_steps: 50
• num_epochs: 3
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.4562	1.0	443	0.6548	0.5561
0.452	2.0	886	0.6418	0.5487
0.4359	2.9944	1326	0.6287	0.5388

Framework versions

• Transformers 4.51.3
• Pytorch 2.6.0+cu124
• Datasets 2.18.0
• Tokenizers 0.21.1