LayoutLM_2

This model is a fine-tuned version of BadreddineHug/LayoutLM_1 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.4785
• Precision: 0.6599
• Recall: 0.7638
• F1: 0.7080
• Accuracy: 0.9097

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-06
• train_batch_size: 2
• eval_batch_size: 2
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• training_steps: 1500

Training results

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1	Accuracy
No log	3.7	100	0.4266	0.6597	0.7480	0.7011	0.9110
No log	7.41	200	0.4415	0.6575	0.7559	0.7033	0.9084
No log	11.11	300	0.4478	0.6575	0.7559	0.7033	0.9084
No log	14.81	400	0.4481	0.6690	0.7638	0.7132	0.9123
0.0237	18.52	500	0.4551	0.6644	0.7638	0.7106	0.9097
0.0237	22.22	600	0.4542	0.6736	0.7638	0.7159	0.9097
0.0237	25.93	700	0.4536	0.6783	0.7638	0.7185	0.9123
0.0237	29.63	800	0.4662	0.6644	0.7638	0.7106	0.9097
0.0237	33.33	900	0.4716	0.6486	0.7559	0.6982	0.9071
0.0146	37.04	1000	0.4644	0.6577	0.7717	0.7101	0.9097
0.0146	40.74	1100	0.4732	0.6599	0.7638	0.7080	0.9097
0.0146	44.44	1200	0.4727	0.6667	0.7717	0.7153	0.9110
0.0146	48.15	1300	0.4774	0.6531	0.7559	0.7007	0.9097
0.0146	51.85	1400	0.4780	0.6599	0.7638	0.7080	0.9097
0.0128	55.56	1500	0.4785	0.6599	0.7638	0.7080	0.9097

Framework versions

• Transformers 4.29.2
• Pytorch 2.0.1+cu118
• Datasets 2.14.4
• Tokenizers 0.13.3