layoutlm-funsd

This model is a fine-tuned version of microsoft/layoutlm-base-uncased on the funsd dataset. It achieves the following results on the evaluation set:

• Loss: 0.6178
• Answer: {'precision': 0.6652719665271967, 'recall': 0.7861557478368356, 'f1': 0.7206798866855525, 'number': 809}
• Header: {'precision': 0.29133858267716534, 'recall': 0.31092436974789917, 'f1': 0.3008130081300813, 'number': 119}
• Question: {'precision': 0.7537248028045574, 'recall': 0.8075117370892019, 'f1': 0.7796917497733454, 'number': 1065}
• Overall Precision: 0.6893
• Overall Recall: 0.7692
• Overall F1: 0.7271
• Overall Accuracy: 0.8014

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: 3e-05
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Answer	Header	Question	Overall Precision	Overall Recall	Overall F1	Overall Accuracy
1.5284	1.0	38	1.0167	{'precision': 0.3938144329896907, 'recall': 0.4721878862793572, 'f1': 0.4294547498594716, 'number': 809}	{'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119}	{'precision': 0.5845528455284553, 'recall': 0.6751173708920187, 'f1': 0.6265795206971678, 'number': 1065}	0.4959	0.5524	0.5227	0.6689
0.8661	2.0	76	0.7179	{'precision': 0.630346232179226, 'recall': 0.765142150803461, 'f1': 0.6912339475153545, 'number': 809}	{'precision': 0.2087912087912088, 'recall': 0.15966386554621848, 'f1': 0.18095238095238092, 'number': 119}	{'precision': 0.7058823529411765, 'recall': 0.7436619718309859, 'f1': 0.7242798353909465, 'number': 1065}	0.6515	0.7175	0.6829	0.7596
0.6265	3.0	114	0.6470	{'precision': 0.6458546571136131, 'recall': 0.7799752781211372, 'f1': 0.7066069428891377, 'number': 809}	{'precision': 0.2972972972972973, 'recall': 0.2773109243697479, 'f1': 0.28695652173913044, 'number': 119}	{'precision': 0.7359649122807017, 'recall': 0.787793427230047, 'f1': 0.7609977324263038, 'number': 1065}	0.6746	0.7541	0.7122	0.7879
0.5076	4.0	152	0.6207	{'precision': 0.6680851063829787, 'recall': 0.7762669962917181, 'f1': 0.7181246426529445, 'number': 809}	{'precision': 0.28, 'recall': 0.29411764705882354, 'f1': 0.28688524590163933, 'number': 119}	{'precision': 0.7368421052631579, 'recall': 0.828169014084507, 'f1': 0.7798408488063661, 'number': 1065}	0.6830	0.7752	0.7262	0.8003
0.4471	5.0	190	0.6178	{'precision': 0.6652719665271967, 'recall': 0.7861557478368356, 'f1': 0.7206798866855525, 'number': 809}	{'precision': 0.29133858267716534, 'recall': 0.31092436974789917, 'f1': 0.3008130081300813, 'number': 119}	{'precision': 0.7537248028045574, 'recall': 0.8075117370892019, 'f1': 0.7796917497733454, 'number': 1065}	0.6893	0.7692	0.7271	0.8014

Framework versions

• Transformers 4.26.1
• Pytorch 1.13.1+cu116
• Datasets 2.10.1
• Tokenizers 0.13.2