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: 0.6840
• Answer: {'precision': 0.7045203969128997, 'recall': 0.7898640296662547, 'f1': 0.7447552447552448, 'number': 809}
• Header: {'precision': 0.3465346534653465, 'recall': 0.29411764705882354, 'f1': 0.3181818181818182, 'number': 119}
• Question: {'precision': 0.7655709342560554, 'recall': 0.8309859154929577, 'f1': 0.7969383160738407, 'number': 1065}
• Overall Precision: 0.7204
• Overall Recall: 0.7822
• Overall F1: 0.7501
• 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: 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
1.81	1.0	10	1.6151	{'precision': 0.004709576138147566, 'recall': 0.003708281829419036, 'f1': 0.004149377593360995, 'number': 809}	{'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119}	{'precision': 0.1045016077170418, 'recall': 0.06103286384976526, 'f1': 0.07705986959098991, 'number': 1065}	0.0540	0.0341	0.0418	0.3338
1.4675	2.0	20	1.2857	{'precision': 0.20516129032258065, 'recall': 0.1965389369592089, 'f1': 0.20075757575757577, 'number': 809}	{'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119}	{'precision': 0.4057971014492754, 'recall': 0.5258215962441315, 'f1': 0.45807770961145194, 'number': 1065}	0.3336	0.3608	0.3467	0.5615
1.1095	3.0	30	0.9749	{'precision': 0.4744136460554371, 'recall': 0.5500618046971569, 'f1': 0.5094447624499141, 'number': 809}	{'precision': 0.02857142857142857, 'recall': 0.008403361344537815, 'f1': 0.012987012987012986, 'number': 119}	{'precision': 0.5271260997067448, 'recall': 0.6751173708920187, 'f1': 0.592013174145739, 'number': 1065}	0.4985	0.5845	0.5381	0.6871
0.8396	4.0	40	0.7887	{'precision': 0.6004098360655737, 'recall': 0.7243510506798516, 'f1': 0.6565826330532212, 'number': 809}	{'precision': 0.1016949152542373, 'recall': 0.05042016806722689, 'f1': 0.06741573033707866, 'number': 119}	{'precision': 0.6529942575881871, 'recall': 0.7474178403755869, 'f1': 0.6970227670753065, 'number': 1065}	0.6158	0.6964	0.6536	0.7516
0.6706	5.0	50	0.7238	{'precision': 0.6403887688984882, 'recall': 0.7330037082818294, 'f1': 0.6835734870317004, 'number': 809}	{'precision': 0.1875, 'recall': 0.12605042016806722, 'f1': 0.1507537688442211, 'number': 119}	{'precision': 0.6530303030303031, 'recall': 0.8093896713615023, 'f1': 0.7228511530398324, 'number': 1065}	0.6320	0.7376	0.6807	0.7773
0.5725	6.0	60	0.6919	{'precision': 0.6587473002159827, 'recall': 0.754017305315204, 'f1': 0.7031700288184438, 'number': 809}	{'precision': 0.24050632911392406, 'recall': 0.15966386554621848, 'f1': 0.1919191919191919, 'number': 119}	{'precision': 0.7182978723404255, 'recall': 0.7924882629107981, 'f1': 0.7535714285714286, 'number': 1065}	0.6757	0.7391	0.7060	0.7819
0.4918	7.0	70	0.6613	{'precision': 0.685807150595883, 'recall': 0.7824474660074165, 'f1': 0.7309468822170901, 'number': 809}	{'precision': 0.32941176470588235, 'recall': 0.23529411764705882, 'f1': 0.2745098039215686, 'number': 119}	{'precision': 0.730092204526404, 'recall': 0.8178403755868544, 'f1': 0.7714791851195748, 'number': 1065}	0.6960	0.7687	0.7306	0.7954
0.4364	8.0	80	0.6622	{'precision': 0.6994475138121546, 'recall': 0.7824474660074165, 'f1': 0.7386231038506416, 'number': 809}	{'precision': 0.25961538461538464, 'recall': 0.226890756302521, 'f1': 0.242152466367713, 'number': 119}	{'precision': 0.7337826453243471, 'recall': 0.8178403755868544, 'f1': 0.7735346358792186, 'number': 1065}	0.6972	0.7682	0.7310	0.7961
0.3848	9.0	90	0.6633	{'precision': 0.7046460176991151, 'recall': 0.7873918417799752, 'f1': 0.7437244600116752, 'number': 809}	{'precision': 0.3106796116504854, 'recall': 0.2689075630252101, 'f1': 0.28828828828828823, 'number': 119}	{'precision': 0.7519446845289542, 'recall': 0.8169014084507042, 'f1': 0.7830783078307832, 'number': 1065}	0.7112	0.7722	0.7404	0.7973
0.3806	10.0	100	0.6619	{'precision': 0.6894679695982627, 'recall': 0.7849196538936959, 'f1': 0.7341040462427746, 'number': 809}	{'precision': 0.29896907216494845, 'recall': 0.24369747899159663, 'f1': 0.2685185185185185, 'number': 119}	{'precision': 0.7575236457437661, 'recall': 0.8272300469483568, 'f1': 0.7908438061041293, 'number': 1065}	0.7084	0.7752	0.7403	0.8005
0.3245	11.0	110	0.6781	{'precision': 0.7051569506726457, 'recall': 0.7775030902348579, 'f1': 0.7395649617871839, 'number': 809}	{'precision': 0.32710280373831774, 'recall': 0.29411764705882354, 'f1': 0.3097345132743363, 'number': 119}	{'precision': 0.7420701168614358, 'recall': 0.8347417840375587, 'f1': 0.7856827220503757, 'number': 1065}	0.7069	0.7792	0.7413	0.7994
0.3037	12.0	120	0.6741	{'precision': 0.7049723756906078, 'recall': 0.788627935723115, 'f1': 0.7444574095682615, 'number': 809}	{'precision': 0.32, 'recall': 0.2689075630252101, 'f1': 0.2922374429223744, 'number': 119}	{'precision': 0.7791519434628975, 'recall': 0.828169014084507, 'f1': 0.8029130632680928, 'number': 1065}	0.7263	0.7787	0.7516	0.8001
0.2917	13.0	130	0.6849	{'precision': 0.7, 'recall': 0.7787391841779975, 'f1': 0.7372732592159158, 'number': 809}	{'precision': 0.32673267326732675, 'recall': 0.2773109243697479, 'f1': 0.30000000000000004, 'number': 119}	{'precision': 0.7703056768558952, 'recall': 0.828169014084507, 'f1': 0.7981900452488688, 'number': 1065}	0.7199	0.7752	0.7466	0.8011
0.2692	14.0	140	0.6823	{'precision': 0.7019867549668874, 'recall': 0.7861557478368356, 'f1': 0.7416909620991254, 'number': 809}	{'precision': 0.35051546391752575, 'recall': 0.2857142857142857, 'f1': 0.3148148148148148, 'number': 119}	{'precision': 0.7642980935875217, 'recall': 0.828169014084507, 'f1': 0.7949526813880126, 'number': 1065}	0.7195	0.7787	0.7480	0.8019
0.2721	15.0	150	0.6840	{'precision': 0.7045203969128997, 'recall': 0.7898640296662547, 'f1': 0.7447552447552448, 'number': 809}	{'precision': 0.3465346534653465, 'recall': 0.29411764705882354, 'f1': 0.3181818181818182, 'number': 119}	{'precision': 0.7655709342560554, 'recall': 0.8309859154929577, 'f1': 0.7969383160738407, 'number': 1065}	0.7204	0.7822	0.7501	0.8036

Framework versions

• Transformers 4.53.0
• Pytorch 2.6.0+cu124
• Datasets 3.6.0
• Tokenizers 0.21.2