layoutlm-funsd

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

• Loss: 0.9802
• Answer: {'precision': 0.7281879194630873, 'recall': 0.8046971569839307, 'f1': 0.7645331767469172, 'number': 809}
• Header: {'precision': 0.43884892086330934, 'recall': 0.5126050420168067, 'f1': 0.4728682170542636, 'number': 119}
• Question: {'precision': 0.8128390596745028, 'recall': 0.844131455399061, 'f1': 0.8281897742975588, 'number': 1065}
• Overall Precision: 0.7532
• Overall Recall: 0.8083
• Overall F1: 0.7798
• Overall Accuracy: 0.8137

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: 2
• eval_batch_size: 2
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 15
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Answer	Header	Question	Overall Precision	Overall Recall	Overall F1	Overall Accuracy
1.3673	1.0	75	0.7834	{'precision': 0.6307870370370371, 'recall': 0.6736711990111248, 'f1': 0.6515242080095638, 'number': 809}	{'precision': 0.05660377358490566, 'recall': 0.025210084033613446, 'f1': 0.03488372093023256, 'number': 119}	{'precision': 0.6363636363636364, 'recall': 0.7690140845070422, 'f1': 0.6964285714285714, 'number': 1065}	0.6202	0.6859	0.6514	0.7643
0.7401	2.0	150	0.7013	{'precision': 0.6420704845814978, 'recall': 0.7206427688504327, 'f1': 0.6790914385556204, 'number': 809}	{'precision': 0.2169811320754717, 'recall': 0.19327731092436976, 'f1': 0.20444444444444446, 'number': 119}	{'precision': 0.7446443873179092, 'recall': 0.815962441314554, 'f1': 0.7786738351254481, 'number': 1065}	0.6763	0.7401	0.7068	0.7636
0.5127	3.0	225	0.6228	{'precision': 0.7036637931034483, 'recall': 0.8071693448702101, 'f1': 0.7518710420264824, 'number': 809}	{'precision': 0.2987012987012987, 'recall': 0.3865546218487395, 'f1': 0.336996336996337, 'number': 119}	{'precision': 0.7757255936675461, 'recall': 0.828169014084507, 'f1': 0.8010899182561309, 'number': 1065}	0.7125	0.7933	0.7507	0.8033
0.3405	4.0	300	0.6358	{'precision': 0.7230419977298524, 'recall': 0.7873918417799752, 'f1': 0.7538461538461538, 'number': 809}	{'precision': 0.29577464788732394, 'recall': 0.35294117647058826, 'f1': 0.3218390804597701, 'number': 119}	{'precision': 0.7768888888888889, 'recall': 0.8206572769953052, 'f1': 0.7981735159817351, 'number': 1065}	0.7230	0.7792	0.7501	0.8055
0.2492	5.0	375	0.6565	{'precision': 0.7119914346895075, 'recall': 0.8220024721878862, 'f1': 0.7630522088353414, 'number': 809}	{'precision': 0.40869565217391307, 'recall': 0.3949579831932773, 'f1': 0.4017094017094017, 'number': 119}	{'precision': 0.8032056990204809, 'recall': 0.8469483568075117, 'f1': 0.8244972577696525, 'number': 1065}	0.7431	0.8098	0.7750	0.8148
0.1761	6.0	450	0.7601	{'precision': 0.7114754098360656, 'recall': 0.8046971569839307, 'f1': 0.7552204176334106, 'number': 809}	{'precision': 0.4482758620689655, 'recall': 0.4369747899159664, 'f1': 0.44255319148936173, 'number': 119}	{'precision': 0.8157156220767072, 'recall': 0.8187793427230047, 'f1': 0.8172446110590441, 'number': 1065}	0.75	0.7903	0.7696	0.8165
0.1326	7.0	525	0.8064	{'precision': 0.7289719626168224, 'recall': 0.7713226205191595, 'f1': 0.7495495495495494, 'number': 809}	{'precision': 0.41134751773049644, 'recall': 0.48739495798319327, 'f1': 0.4461538461538461, 'number': 119}	{'precision': 0.7851528384279476, 'recall': 0.844131455399061, 'f1': 0.813574660633484, 'number': 1065}	0.7381	0.7933	0.7647	0.8066
0.104	8.0	600	0.8490	{'precision': 0.7248618784530386, 'recall': 0.8108776266996292, 'f1': 0.765460910151692, 'number': 809}	{'precision': 0.4154929577464789, 'recall': 0.4957983193277311, 'f1': 0.4521072796934866, 'number': 119}	{'precision': 0.8113382899628253, 'recall': 0.819718309859155, 'f1': 0.8155067725361981, 'number': 1065}	0.7480	0.7968	0.7716	0.8095
0.0751	9.0	675	0.8807	{'precision': 0.7271714922048997, 'recall': 0.8071693448702101, 'f1': 0.7650849443468072, 'number': 809}	{'precision': 0.40625, 'recall': 0.4369747899159664, 'f1': 0.4210526315789474, 'number': 119}	{'precision': 0.8076580587711487, 'recall': 0.8516431924882629, 'f1': 0.8290676416819013, 'number': 1065}	0.7501	0.8088	0.7784	0.8105
0.0556	10.0	750	0.9078	{'precision': 0.7152466367713004, 'recall': 0.788627935723115, 'f1': 0.7501469723691946, 'number': 809}	{'precision': 0.4014084507042254, 'recall': 0.4789915966386555, 'f1': 0.4367816091954024, 'number': 119}	{'precision': 0.8066604995374653, 'recall': 0.8187793427230047, 'f1': 0.8126747437092265, 'number': 1065}	0.7409	0.7863	0.7629	0.8071
0.0494	11.0	825	0.9615	{'precision': 0.7342342342342343, 'recall': 0.8059332509270705, 'f1': 0.7684148497348262, 'number': 809}	{'precision': 0.4206896551724138, 'recall': 0.5126050420168067, 'f1': 0.46212121212121215, 'number': 119}	{'precision': 0.8015943312666076, 'recall': 0.8497652582159625, 'f1': 0.8249772105742936, 'number': 1065}	0.7484	0.8118	0.7788	0.8022
0.0383	12.0	900	0.9451	{'precision': 0.7216721672167217, 'recall': 0.8108776266996292, 'f1': 0.7636786961583235, 'number': 809}	{'precision': 0.3935483870967742, 'recall': 0.5126050420168067, 'f1': 0.44525547445255476, 'number': 119}	{'precision': 0.8148487626031164, 'recall': 0.8347417840375587, 'f1': 0.8246753246753246, 'number': 1065}	0.7452	0.8058	0.7743	0.8148
0.0316	13.0	975	0.9593	{'precision': 0.734533183352081, 'recall': 0.8071693448702101, 'f1': 0.769140164899882, 'number': 809}	{'precision': 0.4025974025974026, 'recall': 0.5210084033613446, 'f1': 0.45421245421245426, 'number': 119}	{'precision': 0.8153564899451554, 'recall': 0.8375586854460094, 'f1': 0.826308476146364, 'number': 1065}	0.7520	0.8063	0.7782	0.8120
0.0286	14.0	1050	0.9804	{'precision': 0.7295173961840629, 'recall': 0.8034610630407911, 'f1': 0.7647058823529411, 'number': 809}	{'precision': 0.4246575342465753, 'recall': 0.5210084033613446, 'f1': 0.46792452830188674, 'number': 119}	{'precision': 0.8177697189483227, 'recall': 0.8469483568075117, 'f1': 0.8321033210332104, 'number': 1065}	0.7542	0.8098	0.7810	0.8130
0.0273	15.0	1125	0.9802	{'precision': 0.7281879194630873, 'recall': 0.8046971569839307, 'f1': 0.7645331767469172, 'number': 809}	{'precision': 0.43884892086330934, 'recall': 0.5126050420168067, 'f1': 0.4728682170542636, 'number': 119}	{'precision': 0.8128390596745028, 'recall': 0.844131455399061, 'f1': 0.8281897742975588, 'number': 1065}	0.7532	0.8083	0.7798	0.8137

Framework versions

• Transformers 4.57.0
• Pytorch 2.8.0+cu128
• Datasets 4.2.0
• Tokenizers 0.22.0