lilt-en-funsd

This model is a fine-tuned version of SCUT-DLVCLab/lilt-roberta-en-base on the funsd-layoutlmv3 dataset. It achieves the following results on the evaluation set:

• Loss: 1.1164
• Answer: {'precision': 0.8957575757575758, 'recall': 0.9045287637698899, 'f1': 0.900121802679659, 'number': 817}
• Header: {'precision': 0.5645161290322581, 'recall': 0.5882352941176471, 'f1': 0.5761316872427984, 'number': 119}
• Question: {'precision': 0.904147465437788, 'recall': 0.9108635097493036, 'f1': 0.9074930619796485, 'number': 1077}
• Overall Precision: 0.8800
• Overall Recall: 0.8892
• Overall F1: 0.8846
• Overall Accuracy: 0.8211

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: 5e-05
• train_batch_size: 1
• eval_batch_size: 1
• 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	Answer	Header	Question	Overall Precision	Overall Recall	Overall F1	Overall Accuracy
0.0418	1.34	200	1.1164	{'precision': 0.8957575757575758, 'recall': 0.9045287637698899, 'f1': 0.900121802679659, 'number': 817}	{'precision': 0.5645161290322581, 'recall': 0.5882352941176471, 'f1': 0.5761316872427984, 'number': 119}	{'precision': 0.904147465437788, 'recall': 0.9108635097493036, 'f1': 0.9074930619796485, 'number': 1077}	0.8800	0.8892	0.8846	0.8211
0.0473	2.68	400	1.1164	{'precision': 0.8957575757575758, 'recall': 0.9045287637698899, 'f1': 0.900121802679659, 'number': 817}	{'precision': 0.5645161290322581, 'recall': 0.5882352941176471, 'f1': 0.5761316872427984, 'number': 119}	{'precision': 0.904147465437788, 'recall': 0.9108635097493036, 'f1': 0.9074930619796485, 'number': 1077}	0.8800	0.8892	0.8846	0.8211
0.0444	4.03	600	1.1164	{'precision': 0.8957575757575758, 'recall': 0.9045287637698899, 'f1': 0.900121802679659, 'number': 817}	{'precision': 0.5645161290322581, 'recall': 0.5882352941176471, 'f1': 0.5761316872427984, 'number': 119}	{'precision': 0.904147465437788, 'recall': 0.9108635097493036, 'f1': 0.9074930619796485, 'number': 1077}	0.8800	0.8892	0.8846	0.8211
0.0532	5.37	800	1.1164	{'precision': 0.8957575757575758, 'recall': 0.9045287637698899, 'f1': 0.900121802679659, 'number': 817}	{'precision': 0.5645161290322581, 'recall': 0.5882352941176471, 'f1': 0.5761316872427984, 'number': 119}	{'precision': 0.904147465437788, 'recall': 0.9108635097493036, 'f1': 0.9074930619796485, 'number': 1077}	0.8800	0.8892	0.8846	0.8211
0.0405	6.71	1000	1.1164	{'precision': 0.8957575757575758, 'recall': 0.9045287637698899, 'f1': 0.900121802679659, 'number': 817}	{'precision': 0.5645161290322581, 'recall': 0.5882352941176471, 'f1': 0.5761316872427984, 'number': 119}	{'precision': 0.904147465437788, 'recall': 0.9108635097493036, 'f1': 0.9074930619796485, 'number': 1077}	0.8800	0.8892	0.8846	0.8211
0.0383	8.05	1200	1.1164	{'precision': 0.8957575757575758, 'recall': 0.9045287637698899, 'f1': 0.900121802679659, 'number': 817}	{'precision': 0.5645161290322581, 'recall': 0.5882352941176471, 'f1': 0.5761316872427984, 'number': 119}	{'precision': 0.904147465437788, 'recall': 0.9108635097493036, 'f1': 0.9074930619796485, 'number': 1077}	0.8800	0.8892	0.8846	0.8211
0.0494	9.4	1400	1.1164	{'precision': 0.8957575757575758, 'recall': 0.9045287637698899, 'f1': 0.900121802679659, 'number': 817}	{'precision': 0.5645161290322581, 'recall': 0.5882352941176471, 'f1': 0.5761316872427984, 'number': 119}	{'precision': 0.904147465437788, 'recall': 0.9108635097493036, 'f1': 0.9074930619796485, 'number': 1077}	0.8800	0.8892	0.8846	0.8211

Framework versions

• Transformers 4.28.1
• Pytorch 2.0.0+cu118
• Datasets 2.12.0
• Tokenizers 0.13.3