DALPA_new_AR

This model is a fine-tuned version of MohamedRashad/arabic-small-nougat on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.0557

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: 0.0001
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 6
• total_train_batch_size: 48
• 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
• num_epochs: 15

Training results

Training Loss	Epoch	Step	Validation Loss
9.292	0.99	66	1.5167
7.3274	1.99	132	1.2984
7.1742	2.99	198	1.2041
6.794	3.99	264	1.1550
6.3938	4.99	330	1.1210
6.3269	5.99	396	1.0967
6.0447	6.99	462	1.0809
5.922	7.99	528	1.0692
6.1488	8.99	594	1.0613
5.9237	9.99	660	1.0576
5.9782	10.99	726	1.0558
5.9258	11.99	792	1.0559
5.9427	12.99	858	1.0558
5.9771	13.99	924	1.0556
6.2395	14.99	990	1.0557

Framework versions

• Transformers 4.47.1
• Pytorch 2.5.1+cu121
• Datasets 4.1.1
• Tokenizers 0.21.0