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: 0.6896
• Answer: {'precision': 0.7152245345016429, 'recall': 0.8071693448702101, 'f1': 0.7584204413472706, 'number': 809}
• Header: {'precision': 0.3697478991596639, 'recall': 0.3697478991596639, 'f1': 0.3697478991596639, 'number': 119}
• Question: {'precision': 0.7833333333333333, 'recall': 0.8384976525821596, 'f1': 0.8099773242630386, 'number': 1065}
• Overall Precision: 0.7320
• Overall Recall: 0.7978
• Overall F1: 0.7635
• Overall Accuracy: 0.8098

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 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
1.7932	1.0	10	1.6151	{'precision': 0.013916500994035786, 'recall': 0.00865265760197775, 'f1': 0.010670731707317074, 'number': 809}	{'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119}	{'precision': 0.24816176470588236, 'recall': 0.1267605633802817, 'f1': 0.1678060907395898, 'number': 1065}	0.1356	0.0712	0.0934	0.3264
1.4724	2.0	20	1.2863	{'precision': 0.11666666666666667, 'recall': 0.1211372064276885, 'f1': 0.11885991510006065, 'number': 809}	{'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119}	{'precision': 0.3751733703190014, 'recall': 0.507981220657277, 'f1': 0.43159154367770247, 'number': 1065}	0.2800	0.3206	0.2989	0.5749
1.1346	3.0	30	0.9582	{'precision': 0.44279176201373, 'recall': 0.4783683559950556, 'f1': 0.45989304812834225, 'number': 809}	{'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119}	{'precision': 0.5849647611589663, 'recall': 0.7014084507042253, 'f1': 0.6379163108454312, 'number': 1065}	0.5202	0.5690	0.5435	0.7037
0.8689	4.0	40	0.7730	{'precision': 0.6069182389937107, 'recall': 0.715698393077874, 'f1': 0.6568349404424276, 'number': 809}	{'precision': 0.15384615384615385, 'recall': 0.06722689075630252, 'f1': 0.0935672514619883, 'number': 119}	{'precision': 0.6672519754170325, 'recall': 0.7136150234741784, 'f1': 0.6896551724137931, 'number': 1065}	0.6280	0.6759	0.6510	0.7590
0.6834	5.0	50	0.6983	{'precision': 0.6349206349206349, 'recall': 0.7416563658838071, 'f1': 0.6841505131128848, 'number': 809}	{'precision': 0.273972602739726, 'recall': 0.16806722689075632, 'f1': 0.20833333333333331, 'number': 119}	{'precision': 0.6741214057507987, 'recall': 0.7924882629107981, 'f1': 0.7285282693137678, 'number': 1065}	0.6449	0.7346	0.6868	0.7813
0.5771	6.0	60	0.6775	{'precision': 0.6438631790744467, 'recall': 0.7911001236093943, 'f1': 0.7099278979478648, 'number': 809}	{'precision': 0.3424657534246575, 'recall': 0.21008403361344538, 'f1': 0.2604166666666667, 'number': 119}	{'precision': 0.7335640138408305, 'recall': 0.7962441314553991, 'f1': 0.7636199909950472, 'number': 1065}	0.6806	0.7592	0.7177	0.7871
0.5055	7.0	70	0.6602	{'precision': 0.6920529801324503, 'recall': 0.7750309023485785, 'f1': 0.7311953352769679, 'number': 809}	{'precision': 0.3069306930693069, 'recall': 0.2605042016806723, 'f1': 0.28181818181818186, 'number': 119}	{'precision': 0.7590788308237378, 'recall': 0.8046948356807512, 'f1': 0.781221513217867, 'number': 1065}	0.7093	0.7602	0.7338	0.7950
0.4549	8.0	80	0.6456	{'precision': 0.6804670912951167, 'recall': 0.792336217552534, 'f1': 0.7321530553969159, 'number': 809}	{'precision': 0.2831858407079646, 'recall': 0.2689075630252101, 'f1': 0.27586206896551724, 'number': 119}	{'precision': 0.7497865072587532, 'recall': 0.8244131455399061, 'f1': 0.7853309481216457, 'number': 1065}	0.6968	0.7782	0.7352	0.8069
0.3945	9.0	90	0.6484	{'precision': 0.6906552094522019, 'recall': 0.7948084054388134, 'f1': 0.7390804597701149, 'number': 809}	{'precision': 0.30833333333333335, 'recall': 0.31092436974789917, 'f1': 0.3096234309623431, 'number': 119}	{'precision': 0.7660869565217391, 'recall': 0.8272300469483568, 'f1': 0.7954853273137698, 'number': 1065}	0.7092	0.7832	0.7444	0.8067
0.3887	10.0	100	0.6674	{'precision': 0.6968085106382979, 'recall': 0.8096415327564895, 'f1': 0.7489994282447112, 'number': 809}	{'precision': 0.31, 'recall': 0.2605042016806723, 'f1': 0.2831050228310502, 'number': 119}	{'precision': 0.790990990990991, 'recall': 0.8244131455399061, 'f1': 0.8073563218390805, 'number': 1065}	0.7274	0.7847	0.7550	0.8115
0.3299	11.0	110	0.6748	{'precision': 0.7125550660792952, 'recall': 0.799752781211372, 'f1': 0.7536400698893418, 'number': 809}	{'precision': 0.3305785123966942, 'recall': 0.33613445378151263, 'f1': 0.33333333333333337, 'number': 119}	{'precision': 0.7663230240549829, 'recall': 0.8375586854460094, 'f1': 0.800358905338717, 'number': 1065}	0.7200	0.7923	0.7544	0.8053
0.3088	12.0	120	0.6757	{'precision': 0.7155361050328227, 'recall': 0.8084054388133498, 'f1': 0.759141033081834, 'number': 809}	{'precision': 0.3904761904761905, 'recall': 0.3445378151260504, 'f1': 0.36607142857142855, 'number': 119}	{'precision': 0.7783641160949868, 'recall': 0.8309859154929577, 'f1': 0.8038147138964576, 'number': 1065}	0.7328	0.7928	0.7616	0.8076
0.2922	13.0	130	0.6892	{'precision': 0.7142857142857143, 'recall': 0.8096415327564895, 'f1': 0.7589803012746235, 'number': 809}	{'precision': 0.38461538461538464, 'recall': 0.37815126050420167, 'f1': 0.38135593220338987, 'number': 119}	{'precision': 0.7850133809099019, 'recall': 0.8262910798122066, 'f1': 0.8051235132662397, 'number': 1065}	0.7332	0.7928	0.7618	0.8076
0.2692	14.0	140	0.6906	{'precision': 0.7212389380530974, 'recall': 0.8059332509270705, 'f1': 0.7612375948628138, 'number': 809}	{'precision': 0.375, 'recall': 0.37815126050420167, 'f1': 0.37656903765690375, 'number': 119}	{'precision': 0.7841409691629956, 'recall': 0.8356807511737089, 'f1': 0.8090909090909091, 'number': 1065}	0.7351	0.7963	0.7645	0.8087
0.2735	15.0	150	0.6896	{'precision': 0.7152245345016429, 'recall': 0.8071693448702101, 'f1': 0.7584204413472706, 'number': 809}	{'precision': 0.3697478991596639, 'recall': 0.3697478991596639, 'f1': 0.3697478991596639, 'number': 119}	{'precision': 0.7833333333333333, 'recall': 0.8384976525821596, 'f1': 0.8099773242630386, 'number': 1065}	0.7320	0.7978	0.7635	0.8098

Framework versions

• Transformers 4.47.1
• Pytorch 2.5.1+cu121
• Datasets 3.2.0
• Tokenizers 0.21.0