layoutlmv3-finetuned-arabic_docs

This model is a fine-tuned version of microsoft/layoutlmv3-base on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0174
• Precision: 0.9945
• Recall: 0.9890
• F1: 0.9917
• Accuracy: 0.9966

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: 10
• eval_batch_size: 10
• seed: 42
• 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
• training_steps: 2500

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	6.7568	250	0.0050	0.9972	0.9945	0.9959	0.9983
0.0675	13.5135	500	0.0136	0.9972	0.9945	0.9959	0.9983
0.0675	20.2703	750	0.0143	0.9972	0.9945	0.9959	0.9983
0.0079	27.0270	1000	0.0152	0.9972	0.9945	0.9959	0.9983
0.0079	33.7838	1250	0.0137	0.9972	0.9945	0.9959	0.9983
0.006	40.5405	1500	0.0151	0.9972	0.9945	0.9959	0.9983
0.006	47.2973	1750	0.0161	0.9945	0.9890	0.9917	0.9966
0.005	54.0541	2000	0.0170	0.9945	0.9890	0.9917	0.9966
0.005	60.8108	2250	0.0158	0.9945	0.9890	0.9917	0.9966
0.0044	67.5676	2500	0.0174	0.9945	0.9890	0.9917	0.9966

Framework versions

• Transformers 4.47.0
• Pytorch 2.5.1+cu121
• Datasets 3.3.1
• Tokenizers 0.21.0