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: 0.6947
• Answer: {'precision': 0.7250554323725056, 'recall': 0.8084054388133498, 'f1': 0.7644652250146114, 'number': 809}
• Header: {'precision': 0.30158730158730157, 'recall': 0.31932773109243695, 'f1': 0.310204081632653, 'number': 119}
• Question: {'precision': 0.767586821015138, 'recall': 0.8093896713615023, 'f1': 0.7879341864716636, 'number': 1065}
• Overall Precision: 0.7225
• Overall Recall: 0.7797
• Overall F1: 0.75
• Overall Accuracy: 0.8070

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: Use OptimizerNames.ADAMW_TORCH 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
1.742	1.0	10	1.5266	{'precision': 0.027950310559006212, 'recall': 0.03337453646477132, 'f1': 0.030422535211267605, 'number': 809}	{'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119}	{'precision': 0.2287292817679558, 'recall': 0.19436619718309858, 'f1': 0.21015228426395938, 'number': 1065}	0.1251	0.1174	0.1211	0.4247
1.412	2.0	20	1.2278	{'precision': 0.19525801952580196, 'recall': 0.173053152039555, 'f1': 0.1834862385321101, 'number': 809}	{'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119}	{'precision': 0.4582560296846011, 'recall': 0.463849765258216, 'f1': 0.4610359309379375, 'number': 1065}	0.3532	0.3181	0.3347	0.5888
1.0962	3.0	30	0.9645	{'precision': 0.4753157290470723, 'recall': 0.511742892459827, 'f1': 0.4928571428571428, 'number': 809}	{'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119}	{'precision': 0.6110183639398998, 'recall': 0.6873239436619718, 'f1': 0.6469288555015466, 'number': 1065}	0.5478	0.5750	0.5611	0.7154
0.838	4.0	40	0.7924	{'precision': 0.6248671625929861, 'recall': 0.7268232385661311, 'f1': 0.672, 'number': 809}	{'precision': 0.12698412698412698, 'recall': 0.06722689075630252, 'f1': 0.08791208791208792, 'number': 119}	{'precision': 0.6594863297431649, 'recall': 0.7474178403755869, 'f1': 0.7007042253521126, 'number': 1065}	0.6296	0.6984	0.6622	0.7647
0.6636	5.0	50	0.7294	{'precision': 0.6722037652270211, 'recall': 0.7503090234857849, 'f1': 0.7091121495327103, 'number': 809}	{'precision': 0.2077922077922078, 'recall': 0.13445378151260504, 'f1': 0.16326530612244897, 'number': 119}	{'precision': 0.6664086687306502, 'recall': 0.8084507042253521, 'f1': 0.7305897327110734, 'number': 1065}	0.6532	0.7446	0.6959	0.7781
0.5632	6.0	60	0.6983	{'precision': 0.660164271047228, 'recall': 0.7948084054388134, 'f1': 0.7212563095905777, 'number': 809}	{'precision': 0.21739130434782608, 'recall': 0.12605042016806722, 'f1': 0.1595744680851064, 'number': 119}	{'precision': 0.7283842794759825, 'recall': 0.7830985915492957, 'f1': 0.7547511312217194, 'number': 1065}	0.6819	0.7486	0.7137	0.7905
0.4868	7.0	70	0.6635	{'precision': 0.7008830022075055, 'recall': 0.7849196538936959, 'f1': 0.7405247813411079, 'number': 809}	{'precision': 0.25742574257425743, 'recall': 0.2184873949579832, 'f1': 0.23636363636363636, 'number': 119}	{'precision': 0.7467248908296943, 'recall': 0.8028169014084507, 'f1': 0.7737556561085973, 'number': 1065}	0.7045	0.7607	0.7315	0.7993
0.4332	8.0	80	0.6626	{'precision': 0.6882168925964547, 'recall': 0.8158220024721878, 'f1': 0.7466063348416289, 'number': 809}	{'precision': 0.2727272727272727, 'recall': 0.226890756302521, 'f1': 0.24770642201834864, 'number': 119}	{'precision': 0.7463456577815993, 'recall': 0.8150234741784037, 'f1': 0.7791741472172352, 'number': 1065}	0.7001	0.7802	0.7380	0.7992
0.3853	9.0	90	0.6623	{'precision': 0.7160220994475138, 'recall': 0.8009888751545118, 'f1': 0.7561260210035006, 'number': 809}	{'precision': 0.30927835051546393, 'recall': 0.25210084033613445, 'f1': 0.2777777777777778, 'number': 119}	{'precision': 0.753448275862069, 'recall': 0.8206572769953052, 'f1': 0.7856179775280899, 'number': 1065}	0.7179	0.7787	0.7471	0.8031
0.3733	10.0	100	0.6695	{'precision': 0.7180327868852459, 'recall': 0.8121137206427689, 'f1': 0.7621809744779582, 'number': 809}	{'precision': 0.28846153846153844, 'recall': 0.25210084033613445, 'f1': 0.26905829596412556, 'number': 119}	{'precision': 0.77068345323741, 'recall': 0.8046948356807512, 'f1': 0.7873220027560864, 'number': 1065}	0.7245	0.7747	0.7488	0.8085
0.3201	11.0	110	0.6826	{'precision': 0.7122381477398015, 'recall': 0.7985166872682324, 'f1': 0.752913752913753, 'number': 809}	{'precision': 0.32142857142857145, 'recall': 0.3025210084033613, 'f1': 0.3116883116883117, 'number': 119}	{'precision': 0.7510620220900595, 'recall': 0.8300469483568075, 'f1': 0.7885816235504014, 'number': 1065}	0.7131	0.7858	0.7477	0.8048
0.3027	12.0	120	0.6841	{'precision': 0.7213656387665198, 'recall': 0.8096415327564895, 'f1': 0.762958648806057, 'number': 809}	{'precision': 0.34210526315789475, 'recall': 0.3277310924369748, 'f1': 0.33476394849785407, 'number': 119}	{'precision': 0.7768744354110207, 'recall': 0.8075117370892019, 'f1': 0.7918968692449355, 'number': 1065}	0.7299	0.7797	0.7540	0.8068
0.2902	13.0	130	0.6871	{'precision': 0.7210065645514223, 'recall': 0.8145859085290482, 'f1': 0.7649448636099826, 'number': 809}	{'precision': 0.32142857142857145, 'recall': 0.3025210084033613, 'f1': 0.3116883116883117, 'number': 119}	{'precision': 0.7732506643046945, 'recall': 0.819718309859155, 'f1': 0.7958067456700091, 'number': 1065}	0.7276	0.7868	0.7560	0.8073
0.2694	14.0	140	0.6911	{'precision': 0.7197802197802198, 'recall': 0.8096415327564895, 'f1': 0.7620709714950552, 'number': 809}	{'precision': 0.32456140350877194, 'recall': 0.31092436974789917, 'f1': 0.31759656652360513, 'number': 119}	{'precision': 0.7796762589928058, 'recall': 0.8140845070422535, 'f1': 0.7965089572806615, 'number': 1065}	0.7299	0.7822	0.7551	0.8083
0.2721	15.0	150	0.6947	{'precision': 0.7250554323725056, 'recall': 0.8084054388133498, 'f1': 0.7644652250146114, 'number': 809}	{'precision': 0.30158730158730157, 'recall': 0.31932773109243695, 'f1': 0.310204081632653, 'number': 119}	{'precision': 0.767586821015138, 'recall': 0.8093896713615023, 'f1': 0.7879341864716636, 'number': 1065}	0.7225	0.7797	0.75	0.8070

Framework versions

• Transformers 4.48.3
• Pytorch 2.5.1+cu124
• Tokenizers 0.21.0