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: 0.6510
• Answer: {'precision': 0.7025527192008879, 'recall': 0.7824474660074165, 'f1': 0.7403508771929823, 'number': 809}
• Header: {'precision': 0.28421052631578947, 'recall': 0.226890756302521, 'f1': 0.25233644859813087, 'number': 119}
• Question: {'precision': 0.7480916030534351, 'recall': 0.828169014084507, 'f1': 0.7860962566844921, 'number': 1065}
• Overall Precision: 0.7090
• Overall Recall: 0.7737
• Overall F1: 0.7399
• Overall Accuracy: 0.8032

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: 16
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Answer	Header	Question	Overall Precision	Overall Recall	Overall F1	Overall Accuracy
1.7428	1.0	10	1.5458	{'precision': 0.030690537084398978, 'recall': 0.04449938195302843, 'f1': 0.036326942482341064, 'number': 809}	{'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119}	{'precision': 0.18740157480314962, 'recall': 0.22347417840375586, 'f1': 0.2038543897216274, 'number': 1065}	0.1122	0.1375	0.1235	0.4326
1.3991	2.0	20	1.2229	{'precision': 0.1326676176890157, 'recall': 0.11495673671199011, 'f1': 0.12317880794701987, 'number': 809}	{'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119}	{'precision': 0.5, 'recall': 0.5352112676056338, 'f1': 0.5170068027210885, 'number': 1065}	0.3597	0.3327	0.3457	0.5731
1.0911	3.0	30	0.9391	{'precision': 0.47231638418079097, 'recall': 0.5166872682323856, 'f1': 0.4935064935064935, 'number': 809}	{'precision': 0.058823529411764705, 'recall': 0.01680672268907563, 'f1': 0.026143790849673203, 'number': 119}	{'precision': 0.6528268551236749, 'recall': 0.6938967136150235, 'f1': 0.6727355484751935, 'number': 1065}	0.5651	0.5815	0.5732	0.7183
0.8461	4.0	40	0.7784	{'precision': 0.6047717842323651, 'recall': 0.7206427688504327, 'f1': 0.6576424139875917, 'number': 809}	{'precision': 0.15384615384615385, 'recall': 0.06722689075630252, 'f1': 0.0935672514619883, 'number': 119}	{'precision': 0.6666666666666666, 'recall': 0.7455399061032864, 'f1': 0.7039007092198581, 'number': 1065}	0.6275	0.6949	0.6595	0.7638
0.6966	5.0	50	0.7307	{'precision': 0.6315228966986155, 'recall': 0.7330037082818294, 'f1': 0.6784897025171623, 'number': 809}	{'precision': 0.21052631578947367, 'recall': 0.13445378151260504, 'f1': 0.1641025641025641, 'number': 119}	{'precision': 0.6925064599483204, 'recall': 0.7549295774647887, 'f1': 0.7223719676549865, 'number': 1065}	0.6494	0.7090	0.6779	0.7703
0.6037	6.0	60	0.6834	{'precision': 0.657922350472193, 'recall': 0.7750309023485785, 'f1': 0.7116912599318955, 'number': 809}	{'precision': 0.3150684931506849, 'recall': 0.19327731092436976, 'f1': 0.23958333333333334, 'number': 119}	{'precision': 0.7021103896103896, 'recall': 0.812206572769953, 'f1': 0.7531562908141053, 'number': 1065}	0.6709	0.7602	0.7128	0.7915
0.5421	7.0	70	0.6692	{'precision': 0.671306209850107, 'recall': 0.7750309023485785, 'f1': 0.7194492254733217, 'number': 809}	{'precision': 0.2823529411764706, 'recall': 0.20168067226890757, 'f1': 0.23529411764705882, 'number': 119}	{'precision': 0.7227467811158799, 'recall': 0.7906103286384977, 'f1': 0.7551569506726458, 'number': 1065}	0.6836	0.7491	0.7149	0.7931
0.5085	8.0	80	0.6549	{'precision': 0.6901874310915105, 'recall': 0.7737948084054388, 'f1': 0.7296037296037297, 'number': 809}	{'precision': 0.3023255813953488, 'recall': 0.2184873949579832, 'f1': 0.25365853658536586, 'number': 119}	{'precision': 0.7408637873754153, 'recall': 0.8375586854460094, 'f1': 0.7862494490965183, 'number': 1065}	0.7028	0.7747	0.7370	0.7982
0.4692	9.0	90	0.6517	{'precision': 0.6973684210526315, 'recall': 0.7861557478368356, 'f1': 0.7391051714119697, 'number': 809}	{'precision': 0.2903225806451613, 'recall': 0.226890756302521, 'f1': 0.25471698113207547, 'number': 119}	{'precision': 0.7534364261168385, 'recall': 0.8234741784037559, 'f1': 0.7868999551368328, 'number': 1065}	0.7100	0.7727	0.7400	0.8025
0.4538	10.0	100	0.6510	{'precision': 0.7025527192008879, 'recall': 0.7824474660074165, 'f1': 0.7403508771929823, 'number': 809}	{'precision': 0.28421052631578947, 'recall': 0.226890756302521, 'f1': 0.25233644859813087, 'number': 119}	{'precision': 0.7480916030534351, 'recall': 0.828169014084507, 'f1': 0.7860962566844921, 'number': 1065}	0.7090	0.7737	0.7399	0.8032

Framework versions

• Transformers 4.31.0
• Pytorch 2.0.1+cu118
• Datasets 2.14.2
• Tokenizers 0.13.3