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: 0.7293
• Answer: {'precision': 0.7119205298013245, 'recall': 0.7972805933250927, 'f1': 0.7521865889212828, 'number': 809}
• Header: {'precision': 0.30327868852459017, 'recall': 0.31092436974789917, 'f1': 0.3070539419087137, 'number': 119}
• Question: {'precision': 0.7780701754385965, 'recall': 0.8328638497652582, 'f1': 0.8045351473922903, 'number': 1065}
• Overall Precision: 0.7237
• Overall Recall: 0.7873
• Overall F1: 0.7541
• Overall Accuracy: 0.7994

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
1.8121	1.0	10	1.5778	{'precision': 0.02838221381267739, 'recall': 0.037082818294190356, 'f1': 0.03215434083601286, 'number': 809}	{'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119}	{'precision': 0.19402985074626866, 'recall': 0.20751173708920187, 'f1': 0.20054446460980033, 'number': 1065}	0.1143	0.1259	0.1198	0.4198
1.4301	2.0	20	1.2246	{'precision': 0.15695067264573992, 'recall': 0.12978986402966625, 'f1': 0.14208389715832204, 'number': 809}	{'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119}	{'precision': 0.46811702925731435, 'recall': 0.5859154929577465, 'f1': 0.5204336947456214, 'number': 1065}	0.3641	0.3658	0.3650	0.5873
1.0718	3.0	30	0.9296	{'precision': 0.48633879781420764, 'recall': 0.5500618046971569, 'f1': 0.5162412993039442, 'number': 809}	{'precision': 0.029411764705882353, 'recall': 0.008403361344537815, 'f1': 0.013071895424836602, 'number': 119}	{'precision': 0.5914927768860353, 'recall': 0.692018779342723, 'f1': 0.6378191259195154, 'number': 1065}	0.5390	0.5936	0.5649	0.7243
0.814	4.0	40	0.7678	{'precision': 0.5895765472312704, 'recall': 0.6711990111248455, 'f1': 0.6277456647398845, 'number': 809}	{'precision': 0.1724137931034483, 'recall': 0.12605042016806722, 'f1': 0.14563106796116504, 'number': 119}	{'precision': 0.64, 'recall': 0.7661971830985915, 'f1': 0.6974358974358974, 'number': 1065}	0.6018	0.6894	0.6427	0.7677
0.6527	5.0	50	0.7178	{'precision': 0.6438653637350705, 'recall': 0.7330037082818294, 'f1': 0.6855491329479768, 'number': 809}	{'precision': 0.3, 'recall': 0.17647058823529413, 'f1': 0.22222222222222224, 'number': 119}	{'precision': 0.675, 'recall': 0.8112676056338028, 'f1': 0.7368869936034115, 'number': 1065}	0.6508	0.7416	0.6932	0.7851
0.5605	6.0	60	0.6839	{'precision': 0.6655982905982906, 'recall': 0.7700865265760197, 'f1': 0.7140401146131805, 'number': 809}	{'precision': 0.29411764705882354, 'recall': 0.21008403361344538, 'f1': 0.2450980392156863, 'number': 119}	{'precision': 0.7222222222222222, 'recall': 0.8178403755868544, 'f1': 0.7670629678555702, 'number': 1065}	0.6821	0.7622	0.7199	0.7950
0.4793	7.0	70	0.6672	{'precision': 0.6862955032119914, 'recall': 0.792336217552534, 'f1': 0.7355134825014343, 'number': 809}	{'precision': 0.25961538461538464, 'recall': 0.226890756302521, 'f1': 0.242152466367713, 'number': 119}	{'precision': 0.7552083333333334, 'recall': 0.8169014084507042, 'f1': 0.7848443843031124, 'number': 1065}	0.7023	0.7717	0.7354	0.8014
0.4262	8.0	80	0.6747	{'precision': 0.6762886597938145, 'recall': 0.8108776266996292, 'f1': 0.7374929735806633, 'number': 809}	{'precision': 0.24509803921568626, 'recall': 0.21008403361344538, 'f1': 0.22624434389140272, 'number': 119}	{'precision': 0.7618228718830611, 'recall': 0.831924882629108, 'f1': 0.7953321364452423, 'number': 1065}	0.7011	0.7863	0.7412	0.8000
0.3773	9.0	90	0.6885	{'precision': 0.6932314410480349, 'recall': 0.7849196538936959, 'f1': 0.736231884057971, 'number': 809}	{'precision': 0.30701754385964913, 'recall': 0.29411764705882354, 'f1': 0.30042918454935624, 'number': 119}	{'precision': 0.7628865979381443, 'recall': 0.8338028169014085, 'f1': 0.7967698519515478, 'number': 1065}	0.7101	0.7817	0.7442	0.8015
0.3709	10.0	100	0.6915	{'precision': 0.6982758620689655, 'recall': 0.8009888751545118, 'f1': 0.7461139896373058, 'number': 809}	{'precision': 0.3106796116504854, 'recall': 0.2689075630252101, 'f1': 0.28828828828828823, 'number': 119}	{'precision': 0.7681660899653979, 'recall': 0.8338028169014085, 'f1': 0.7996398018910401, 'number': 1065}	0.7170	0.7868	0.7502	0.8041
0.3123	11.0	110	0.7102	{'precision': 0.7045951859956237, 'recall': 0.796044499381953, 'f1': 0.7475333720255369, 'number': 809}	{'precision': 0.3, 'recall': 0.3025210084033613, 'f1': 0.301255230125523, 'number': 119}	{'precision': 0.7717013888888888, 'recall': 0.8347417840375587, 'f1': 0.8019846639603068, 'number': 1065}	0.7177	0.7873	0.7509	0.8003
0.2944	12.0	120	0.7214	{'precision': 0.7073707370737073, 'recall': 0.7948084054388134, 'f1': 0.7485448195576251, 'number': 809}	{'precision': 0.34285714285714286, 'recall': 0.3025210084033613, 'f1': 0.32142857142857145, 'number': 119}	{'precision': 0.774390243902439, 'recall': 0.8347417840375587, 'f1': 0.8034342521464076, 'number': 1065}	0.7253	0.7868	0.7548	0.8031
0.286	13.0	130	0.7283	{'precision': 0.7105263157894737, 'recall': 0.8009888751545118, 'f1': 0.7530505520046484, 'number': 809}	{'precision': 0.336283185840708, 'recall': 0.31932773109243695, 'f1': 0.32758620689655166, 'number': 119}	{'precision': 0.7801418439716312, 'recall': 0.8262910798122066, 'f1': 0.8025535795713634, 'number': 1065}	0.7274	0.7858	0.7554	0.7994
0.2609	14.0	140	0.7260	{'precision': 0.7144444444444444, 'recall': 0.7948084054388134, 'f1': 0.7524868344060853, 'number': 809}	{'precision': 0.3217391304347826, 'recall': 0.31092436974789917, 'f1': 0.3162393162393162, 'number': 119}	{'precision': 0.7796312554872695, 'recall': 0.8338028169014085, 'f1': 0.8058076225045372, 'number': 1065}	0.7279	0.7868	0.7562	0.8000
0.264	15.0	150	0.7293	{'precision': 0.7119205298013245, 'recall': 0.7972805933250927, 'f1': 0.7521865889212828, 'number': 809}	{'precision': 0.30327868852459017, 'recall': 0.31092436974789917, 'f1': 0.3070539419087137, 'number': 119}	{'precision': 0.7780701754385965, 'recall': 0.8328638497652582, 'f1': 0.8045351473922903, 'number': 1065}	0.7237	0.7873	0.7541	0.7994

Framework versions

• Transformers 4.48.3
• Pytorch 2.5.1+cu124
• Datasets 3.3.2
• Tokenizers 0.21.0