bustamiyusoef/Nougat_base_Jawi_Augment_v1

_base_nougat_Mix_Arabic_hwt_Jawi_augment

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

• Loss: 2.0938

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: 10

Training results

Training Loss	Epoch	Step	Validation Loss
No log	0.9474	9	2.6711
3.269	2.0	19	2.4232
2.5448	2.9474	28	2.2949
2.3721	4.0	38	2.2040
2.2662	4.9474	47	2.1548
2.2067	6.0	57	2.1254
2.1674	6.9474	66	2.1111
2.1416	8.0	76	2.0978
2.1368	8.9474	85	2.0947
2.1268	9.4737	90	2.0938

Framework versions

• Transformers 4.46.3
• Pytorch 2.5.1+cu121
• Datasets 3.2.0
• Tokenizers 0.20.3