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: 0.7180
• Answer: {'precision': 0.7269700332963374, 'recall': 0.8096415327564895, 'f1': 0.7660818713450291, 'number': 809}
• Header: {'precision': 0.2992125984251969, 'recall': 0.31932773109243695, 'f1': 0.30894308943089427, 'number': 119}
• Question: {'precision': 0.7881205673758865, 'recall': 0.8347417840375587, 'f1': 0.8107615139078888, 'number': 1065}
• Overall Precision: 0.7338
• Overall Recall: 0.7938
• Overall F1: 0.7626
• Overall Accuracy: 0.8036

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: 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
1.7814	1.0	10	1.5919	{'precision': 0.015602836879432624, 'recall': 0.013597033374536464, 'f1': 0.01453104359313078, 'number': 809}	{'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119}	{'precision': 0.14666666666666667, 'recall': 0.09295774647887324, 'f1': 0.11379310344827587, 'number': 1065}	0.0797	0.0552	0.0652	0.3502
1.446	2.0	20	1.2482	{'precision': 0.15749235474006115, 'recall': 0.1273176761433869, 'f1': 0.14080656185919346, 'number': 809}	{'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119}	{'precision': 0.45069393718042366, 'recall': 0.5793427230046948, 'f1': 0.5069843878389482, 'number': 1065}	0.3559	0.3613	0.3586	0.5919
1.0955	3.0	30	0.9653	{'precision': 0.44740024183796856, 'recall': 0.4573547589616811, 'f1': 0.45232273838630804, 'number': 809}	{'precision': 0.030303030303030304, 'recall': 0.008403361344537815, 'f1': 0.013157894736842105, 'number': 119}	{'precision': 0.5621703089675961, 'recall': 0.7004694835680751, 'f1': 0.6237458193979933, 'number': 1065}	0.5107	0.5605	0.5344	0.7004
0.8413	4.0	40	0.8063	{'precision': 0.6070686070686071, 'recall': 0.7218788627935723, 'f1': 0.6595143986448335, 'number': 809}	{'precision': 0.1111111111111111, 'recall': 0.05042016806722689, 'f1': 0.06936416184971099, 'number': 119}	{'precision': 0.6669527896995708, 'recall': 0.7295774647887324, 'f1': 0.6968609865470853, 'number': 1065}	0.6268	0.6859	0.6550	0.7495
0.6708	5.0	50	0.7123	{'precision': 0.6703539823008849, 'recall': 0.7490729295426453, 'f1': 0.7075306479859895, 'number': 809}	{'precision': 0.11458333333333333, 'recall': 0.09243697478991597, 'f1': 0.10232558139534885, 'number': 119}	{'precision': 0.6787401574803149, 'recall': 0.8093896713615023, 'f1': 0.7383297644539615, 'number': 1065}	0.6515	0.7421	0.6939	0.7841
0.566	6.0	60	0.7005	{'precision': 0.6466466466466466, 'recall': 0.7985166872682324, 'f1': 0.7146017699115044, 'number': 809}	{'precision': 0.13725490196078433, 'recall': 0.11764705882352941, 'f1': 0.12669683257918554, 'number': 119}	{'precision': 0.7279151943462897, 'recall': 0.7737089201877935, 'f1': 0.7501137915339099, 'number': 1065}	0.6646	0.7446	0.7023	0.7826
0.4969	7.0	70	0.6764	{'precision': 0.6817691477885652, 'recall': 0.7812113720642769, 'f1': 0.7281105990783411, 'number': 809}	{'precision': 0.21666666666666667, 'recall': 0.2184873949579832, 'f1': 0.21757322175732216, 'number': 119}	{'precision': 0.7367521367521368, 'recall': 0.8093896713615023, 'f1': 0.7713646532438478, 'number': 1065}	0.6856	0.7627	0.7221	0.7950
0.4295	8.0	80	0.6735	{'precision': 0.7071823204419889, 'recall': 0.7911001236093943, 'f1': 0.7467911318553092, 'number': 809}	{'precision': 0.21311475409836064, 'recall': 0.2184873949579832, 'f1': 0.21576763485477177, 'number': 119}	{'precision': 0.7489397794741306, 'recall': 0.8291079812206573, 'f1': 0.78698752228164, 'number': 1065}	0.7022	0.7772	0.7378	0.7992
0.3774	9.0	90	0.6814	{'precision': 0.7023809523809523, 'recall': 0.8022249690976514, 'f1': 0.7489901904212348, 'number': 809}	{'precision': 0.2773109243697479, 'recall': 0.2773109243697479, 'f1': 0.2773109243697479, 'number': 119}	{'precision': 0.7639372822299652, 'recall': 0.8234741784037559, 'f1': 0.7925892453682785, 'number': 1065}	0.7115	0.7822	0.7452	0.8047
0.3678	10.0	100	0.6885	{'precision': 0.7193370165745856, 'recall': 0.8046971569839307, 'f1': 0.7596266044340723, 'number': 809}	{'precision': 0.264, 'recall': 0.2773109243697479, 'f1': 0.27049180327868855, 'number': 119}	{'precision': 0.7690972222222222, 'recall': 0.831924882629108, 'f1': 0.799278304014434, 'number': 1065}	0.7195	0.7878	0.7521	0.8050
0.3137	11.0	110	0.6976	{'precision': 0.7161572052401747, 'recall': 0.8108776266996292, 'f1': 0.7605797101449275, 'number': 809}	{'precision': 0.27007299270072993, 'recall': 0.31092436974789917, 'f1': 0.2890625, 'number': 119}	{'precision': 0.7719756309834639, 'recall': 0.8328638497652582, 'f1': 0.8012646793134598, 'number': 1065}	0.7175	0.7928	0.7533	0.8018
0.2983	12.0	120	0.6996	{'precision': 0.7108433734939759, 'recall': 0.8022249690976514, 'f1': 0.7537746806039489, 'number': 809}	{'precision': 0.304, 'recall': 0.31932773109243695, 'f1': 0.31147540983606553, 'number': 119}	{'precision': 0.7810283687943262, 'recall': 0.8272300469483568, 'f1': 0.8034655722754217, 'number': 1065}	0.7239	0.7868	0.7540	0.8029
0.2786	13.0	130	0.7114	{'precision': 0.7172949002217295, 'recall': 0.799752781211372, 'f1': 0.7562828755113968, 'number': 809}	{'precision': 0.29838709677419356, 'recall': 0.31092436974789917, 'f1': 0.3045267489711935, 'number': 119}	{'precision': 0.7814977973568282, 'recall': 0.8328638497652582, 'f1': 0.8063636363636365, 'number': 1065}	0.7270	0.7883	0.7564	0.8019
0.2632	14.0	140	0.7151	{'precision': 0.7256438969764838, 'recall': 0.8009888751545118, 'f1': 0.7614571092831962, 'number': 809}	{'precision': 0.3064516129032258, 'recall': 0.31932773109243695, 'f1': 0.31275720164609055, 'number': 119}	{'precision': 0.7858407079646018, 'recall': 0.8338028169014085, 'f1': 0.8091116173120729, 'number': 1065}	0.7331	0.7898	0.7604	0.8033
0.2668	15.0	150	0.7180	{'precision': 0.7269700332963374, 'recall': 0.8096415327564895, 'f1': 0.7660818713450291, 'number': 809}	{'precision': 0.2992125984251969, 'recall': 0.31932773109243695, 'f1': 0.30894308943089427, 'number': 119}	{'precision': 0.7881205673758865, 'recall': 0.8347417840375587, 'f1': 0.8107615139078888, 'number': 1065}	0.7338	0.7938	0.7626	0.8036

Framework versions

• Transformers 4.41.2
• Pytorch 2.3.1+cu121
• Tokenizers 0.19.1