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: 0.6735
• Answer: {'precision': 0.7215601300108342, 'recall': 0.823238566131026, 'f1': 0.76905311778291, 'number': 809}
• Header: {'precision': 0.3046875, 'recall': 0.3277310924369748, 'f1': 0.31578947368421056, 'number': 119}
• Question: {'precision': 0.7800175284837861, 'recall': 0.8356807511737089, 'f1': 0.8068902991840435, 'number': 1065}
• Overall Precision: 0.7276
• Overall Recall: 0.8003
• Overall F1: 0.7622
• Overall Accuracy: 0.8080

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

Training results

Training Loss	Epoch	Step	Validation Loss	Answer	Header	Question	Overall Precision	Overall Recall	Overall F1	Overall Accuracy
1.7753	1.0	10	1.5651	{'precision': 0.01791713325867861, 'recall': 0.019777503090234856, 'f1': 0.018801410105757928, 'number': 809}	{'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119}	{'precision': 0.253315649867374, 'recall': 0.17934272300469484, 'f1': 0.21000549752611322, 'number': 1065}	0.1257	0.1039	0.1137	0.3966
1.4505	2.0	20	1.2385	{'precision': 0.2100456621004566, 'recall': 0.22744128553770088, 'f1': 0.21839762611275965, 'number': 809}	{'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119}	{'precision': 0.46676197283774123, 'recall': 0.6131455399061033, 'f1': 0.5300324675324676, 'number': 1065}	0.3657	0.4200	0.3909	0.6293
1.0869	3.0	30	0.8993	{'precision': 0.5091496232508074, 'recall': 0.584672435105068, 'f1': 0.5443037974683544, 'number': 809}	{'precision': 0.046511627906976744, 'recall': 0.01680672268907563, 'f1': 0.02469135802469136, 'number': 119}	{'precision': 0.5931254996003198, 'recall': 0.6967136150234742, 'f1': 0.6407599309153713, 'number': 1065}	0.5475	0.6106	0.5773	0.7210
0.8144	4.0	40	0.7685	{'precision': 0.5755755755755756, 'recall': 0.7107540173053152, 'f1': 0.6360619469026548, 'number': 809}	{'precision': 0.15625, 'recall': 0.08403361344537816, 'f1': 0.10928961748633881, 'number': 119}	{'precision': 0.6641350210970464, 'recall': 0.7389671361502348, 'f1': 0.6995555555555556, 'number': 1065}	0.6103	0.6884	0.6470	0.7562
0.6642	5.0	50	0.6960	{'precision': 0.6472424557752341, 'recall': 0.7688504326328801, 'f1': 0.7028248587570621, 'number': 809}	{'precision': 0.19607843137254902, 'recall': 0.16806722689075632, 'f1': 0.18099547511312217, 'number': 119}	{'precision': 0.6795201371036846, 'recall': 0.7446009389671362, 'f1': 0.7105734767025091, 'number': 1065}	0.6435	0.7200	0.6796	0.7773
0.5578	6.0	60	0.6555	{'precision': 0.6557377049180327, 'recall': 0.7911001236093943, 'f1': 0.7170868347338936, 'number': 809}	{'precision': 0.19327731092436976, 'recall': 0.19327731092436976, 'f1': 0.19327731092436978, 'number': 119}	{'precision': 0.7009038619556286, 'recall': 0.8009389671361502, 'f1': 0.7475898334794041, 'number': 1065}	0.6557	0.7607	0.7043	0.7920
0.484	7.0	70	0.6448	{'precision': 0.6560574948665298, 'recall': 0.7898640296662547, 'f1': 0.7167694896242288, 'number': 809}	{'precision': 0.24509803921568626, 'recall': 0.21008403361344538, 'f1': 0.22624434389140272, 'number': 119}	{'precision': 0.7357859531772575, 'recall': 0.8262910798122066, 'f1': 0.7784166298098186, 'number': 1065}	0.6796	0.7747	0.7240	0.8003
0.4248	8.0	80	0.6501	{'precision': 0.6865828092243187, 'recall': 0.8096415327564895, 'f1': 0.7430516165626773, 'number': 809}	{'precision': 0.23972602739726026, 'recall': 0.29411764705882354, 'f1': 0.2641509433962264, 'number': 119}	{'precision': 0.7493403693931399, 'recall': 0.8, 'f1': 0.7738419618528609, 'number': 1065}	0.6893	0.7737	0.7291	0.7993
0.3833	9.0	90	0.6427	{'precision': 0.7062706270627063, 'recall': 0.7935723114956736, 'f1': 0.7473806752037252, 'number': 809}	{'precision': 0.2777777777777778, 'recall': 0.29411764705882354, 'f1': 0.28571428571428575, 'number': 119}	{'precision': 0.7600685518423308, 'recall': 0.8328638497652582, 'f1': 0.7948028673835125, 'number': 1065}	0.7103	0.7847	0.7456	0.8069
0.3435	10.0	100	0.6499	{'precision': 0.7076271186440678, 'recall': 0.8257107540173053, 'f1': 0.7621220764403879, 'number': 809}	{'precision': 0.3217391304347826, 'recall': 0.31092436974789917, 'f1': 0.3162393162393162, 'number': 119}	{'precision': 0.7789566755083996, 'recall': 0.8272300469483568, 'f1': 0.802367941712204, 'number': 1065}	0.7242	0.7958	0.7583	0.8088
0.3157	11.0	110	0.6661	{'precision': 0.7183406113537117, 'recall': 0.8133498145859085, 'f1': 0.7628985507246376, 'number': 809}	{'precision': 0.32231404958677684, 'recall': 0.3277310924369748, 'f1': 0.32499999999999996, 'number': 119}	{'precision': 0.7774846086191732, 'recall': 0.8300469483568075, 'f1': 0.8029064486830154, 'number': 1065}	0.7272	0.7933	0.7588	0.8052
0.2921	12.0	120	0.6645	{'precision': 0.7142857142857143, 'recall': 0.8281829419035847, 'f1': 0.767029192902118, 'number': 809}	{'precision': 0.30158730158730157, 'recall': 0.31932773109243695, 'f1': 0.310204081632653, 'number': 119}	{'precision': 0.7777777777777778, 'recall': 0.8347417840375587, 'f1': 0.8052536231884059, 'number': 1065}	0.7236	0.8013	0.7605	0.8075
0.2805	13.0	130	0.6742	{'precision': 0.7270742358078602, 'recall': 0.823238566131026, 'f1': 0.7721739130434783, 'number': 809}	{'precision': 0.29850746268656714, 'recall': 0.33613445378151263, 'f1': 0.31620553359683795, 'number': 119}	{'precision': 0.7802101576182137, 'recall': 0.8366197183098592, 'f1': 0.8074309016764839, 'number': 1065}	0.7286	0.8013	0.7632	0.8074
0.2676	14.0	140	0.6739	{'precision': 0.720173535791757, 'recall': 0.8207663782447466, 'f1': 0.7671865973425764, 'number': 809}	{'precision': 0.2932330827067669, 'recall': 0.3277310924369748, 'f1': 0.30952380952380953, 'number': 119}	{'precision': 0.7730434782608696, 'recall': 0.8347417840375587, 'f1': 0.8027088036117382, 'number': 1065}	0.7220	0.7988	0.7585	0.8066
0.2731	15.0	150	0.6735	{'precision': 0.7215601300108342, 'recall': 0.823238566131026, 'f1': 0.76905311778291, 'number': 809}	{'precision': 0.3046875, 'recall': 0.3277310924369748, 'f1': 0.31578947368421056, 'number': 119}	{'precision': 0.7800175284837861, 'recall': 0.8356807511737089, 'f1': 0.8068902991840435, 'number': 1065}	0.7276	0.8003	0.7622	0.8080

Framework versions

• Transformers 4.33.0.dev0
• Pytorch 2.0.1+cpu
• Datasets 2.14.4
• Tokenizers 0.13.3