sagemaker-bert-base-arabic-ar-SAS

This model is a fine-tuned version of asafaya/bert-base-arabic on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.2732
• Accuracy: 0.9029

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: 3e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 2
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	291	0.2924	0.8823
0.3433	2.0	582	0.2732	0.9029

Framework versions

• Transformers 4.12.3
• Pytorch 1.9.1
• Datasets 1.15.1
• Tokenizers 0.10.3