lilt-en-funsd

This model is a fine-tuned version of SCUT-DLVCLab/lilt-roberta-en-base on the funsd-layoutlmv3 dataset. It achieves the following results on the evaluation set:

• Loss: 1.6122
• Answer: {'precision': 0.8905882352941177, 'recall': 0.9265605875152999, 'f1': 0.9082183563287344, 'number': 817}
• Header: {'precision': 0.6379310344827587, 'recall': 0.6218487394957983, 'f1': 0.6297872340425532, 'number': 119}
• Question: {'precision': 0.8964252978918423, 'recall': 0.9080779944289693, 'f1': 0.9022140221402215, 'number': 1077}
• Overall Precision: 0.8794
• Overall Recall: 0.8987
• Overall F1: 0.8889
• Overall Accuracy: 0.8020

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: 5e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• training_steps: 2500

Training results

Training Loss	Epoch	Step	Validation Loss	Answer	Header	Question	Overall Precision	Overall Recall	Overall F1	Overall Accuracy
0.4109	10.53	200	0.9316	{'precision': 0.8169491525423729, 'recall': 0.8849449204406364, 'f1': 0.8495887191539365, 'number': 817}	{'precision': 0.5258620689655172, 'recall': 0.5126050420168067, 'f1': 0.5191489361702126, 'number': 119}	{'precision': 0.8695652173913043, 'recall': 0.8913649025069638, 'f1': 0.8803301237964236, 'number': 1077}	0.8285	0.8664	0.8470	0.7998
0.0432	21.05	400	1.6192	{'precision': 0.8022099447513812, 'recall': 0.8886168910648715, 'f1': 0.843205574912892, 'number': 817}	{'precision': 0.5585585585585585, 'recall': 0.5210084033613446, 'f1': 0.5391304347826087, 'number': 119}	{'precision': 0.9002932551319648, 'recall': 0.8551532033426184, 'f1': 0.8771428571428571, 'number': 1077}	0.8382	0.8490	0.8435	0.7705
0.0145	31.58	600	1.4845	{'precision': 0.8556701030927835, 'recall': 0.9143206854345165, 'f1': 0.8840236686390532, 'number': 817}	{'precision': 0.6274509803921569, 'recall': 0.5378151260504201, 'f1': 0.579185520361991, 'number': 119}	{'precision': 0.8772241992882562, 'recall': 0.9155060352831941, 'f1': 0.8959563834620626, 'number': 1077}	0.8561	0.8927	0.8740	0.7985
0.0063	42.11	800	1.4909	{'precision': 0.8667452830188679, 'recall': 0.8996328029375765, 'f1': 0.8828828828828829, 'number': 817}	{'precision': 0.5766423357664233, 'recall': 0.6638655462184874, 'f1': 0.6171874999999999, 'number': 119}	{'precision': 0.891444342226311, 'recall': 0.8997214484679665, 'f1': 0.8955637707948244, 'number': 1077}	0.8605	0.8857	0.8729	0.8117
0.0041	52.63	1000	1.6197	{'precision': 0.834056399132321, 'recall': 0.9412484700122399, 'f1': 0.8844163312248418, 'number': 817}	{'precision': 0.5945945945945946, 'recall': 0.5546218487394958, 'f1': 0.5739130434782609, 'number': 119}	{'precision': 0.8972407231208372, 'recall': 0.8755803156917363, 'f1': 0.8862781954887218, 'number': 1077}	0.8532	0.8833	0.8680	0.7912
0.0043	63.16	1200	1.6797	{'precision': 0.8763005780346821, 'recall': 0.9277845777233782, 'f1': 0.901307966706302, 'number': 817}	{'precision': 0.6134453781512605, 'recall': 0.6134453781512605, 'f1': 0.6134453781512605, 'number': 119}	{'precision': 0.8998144712430427, 'recall': 0.9006499535747446, 'f1': 0.9002320185614848, 'number': 1077}	0.8734	0.8947	0.8839	0.8046
0.0015	73.68	1400	1.5586	{'precision': 0.8683314415437003, 'recall': 0.9363525091799265, 'f1': 0.9010600706713782, 'number': 817}	{'precision': 0.6486486486486487, 'recall': 0.6050420168067226, 'f1': 0.6260869565217391, 'number': 119}	{'precision': 0.9055555555555556, 'recall': 0.9080779944289693, 'f1': 0.9068150208623087, 'number': 1077}	0.8760	0.9016	0.8886	0.8093
0.001	84.21	1600	1.5060	{'precision': 0.8894230769230769, 'recall': 0.9057527539779682, 'f1': 0.8975136446331109, 'number': 817}	{'precision': 0.711340206185567, 'recall': 0.5798319327731093, 'f1': 0.638888888888889, 'number': 119}	{'precision': 0.8927927927927928, 'recall': 0.9201485608170845, 'f1': 0.9062642889803384, 'number': 1077}	0.8828	0.8942	0.8885	0.8202
0.0004	94.74	1800	1.6122	{'precision': 0.8905882352941177, 'recall': 0.9265605875152999, 'f1': 0.9082183563287344, 'number': 817}	{'precision': 0.6379310344827587, 'recall': 0.6218487394957983, 'f1': 0.6297872340425532, 'number': 119}	{'precision': 0.8964252978918423, 'recall': 0.9080779944289693, 'f1': 0.9022140221402215, 'number': 1077}	0.8794	0.8987	0.8889	0.8020
0.0004	105.26	2000	1.6057	{'precision': 0.8791079812206573, 'recall': 0.9167686658506732, 'f1': 0.8975434391851408, 'number': 817}	{'precision': 0.6260869565217392, 'recall': 0.6050420168067226, 'f1': 0.6153846153846154, 'number': 119}	{'precision': 0.8921124206708976, 'recall': 0.9136490250696379, 'f1': 0.9027522935779816, 'number': 1077}	0.8720	0.8967	0.8842	0.8057
0.0004	115.79	2200	1.6179	{'precision': 0.8848413631022327, 'recall': 0.9216646266829865, 'f1': 0.9028776978417267, 'number': 817}	{'precision': 0.6206896551724138, 'recall': 0.6050420168067226, 'f1': 0.6127659574468085, 'number': 119}	{'precision': 0.8921124206708976, 'recall': 0.9136490250696379, 'f1': 0.9027522935779816, 'number': 1077}	0.8739	0.8987	0.8861	0.8070
0.0002	126.32	2400	1.6142	{'precision': 0.8826291079812206, 'recall': 0.9204406364749081, 'f1': 0.9011384062312762, 'number': 817}	{'precision': 0.6068376068376068, 'recall': 0.5966386554621849, 'f1': 0.6016949152542374, 'number': 119}	{'precision': 0.8965201465201466, 'recall': 0.9090064995357474, 'f1': 0.9027201475334256, 'number': 1077}	0.8743	0.8952	0.8846	0.8074

Framework versions

• Transformers 4.30.2
• Pytorch 2.0.1+cu118
• Datasets 2.13.1
• Tokenizers 0.13.3