ananth-docai1

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: 0.7024
• Answer: {'precision': 0.7113513513513513, 'recall': 0.8133498145859085, 'f1': 0.7589388696655133, 'number': 809}
• Header: {'precision': 0.30952380952380953, 'recall': 0.3277310924369748, 'f1': 0.31836734693877555, 'number': 119}
• Question: {'precision': 0.7811387900355872, 'recall': 0.8244131455399061, 'f1': 0.8021927820922796, 'number': 1065}
• Overall Precision: 0.7241
• Overall Recall: 0.7903
• Overall F1: 0.7558
• Overall Accuracy: 0.8106

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
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Answer	Header	Question	Overall Precision	Overall Recall	Overall F1	Overall Accuracy
1.7944	1.0	10	1.6233	{'precision': 0.01929260450160772, 'recall': 0.014833127317676144, 'f1': 0.016771488469601678, 'number': 809}	{'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119}	{'precision': 0.27685325264750377, 'recall': 0.17183098591549295, 'f1': 0.2120509849362688, 'number': 1065}	0.1520	0.0978	0.1190	0.3505
1.5001	2.0	20	1.2971	{'precision': 0.11125, 'recall': 0.1100123609394314, 'f1': 0.11062771908017402, 'number': 809}	{'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119}	{'precision': 0.4044059795436664, 'recall': 0.48262910798122066, 'f1': 0.4400684931506849, 'number': 1065}	0.2912	0.3026	0.2968	0.5348
1.136	3.0	30	0.9852	{'precision': 0.4911699779249448, 'recall': 0.5500618046971569, 'f1': 0.5189504373177842, 'number': 809}	{'precision': 0.0, 'recall': 0.0, 'f1': 0.0, 'number': 119}	{'precision': 0.6086587436332768, 'recall': 0.6732394366197183, 'f1': 0.6393223361569326, 'number': 1065}	0.5562	0.5830	0.5693	0.6941
0.8567	4.0	40	0.8143	{'precision': 0.627744510978044, 'recall': 0.7775030902348579, 'f1': 0.6946438431805633, 'number': 809}	{'precision': 0.06666666666666667, 'recall': 0.01680672268907563, 'f1': 0.026845637583892617, 'number': 119}	{'precision': 0.6987179487179487, 'recall': 0.7164319248826291, 'f1': 0.7074640704682429, 'number': 1065}	0.6563	0.6994	0.6772	0.7467
0.6998	5.0	50	0.7133	{'precision': 0.6534859521331946, 'recall': 0.7762669962917181, 'f1': 0.7096045197740113, 'number': 809}	{'precision': 0.2, 'recall': 0.11764705882352941, 'f1': 0.14814814814814817, 'number': 119}	{'precision': 0.7243532560214094, 'recall': 0.7624413145539906, 'f1': 0.7429094236047575, 'number': 1065}	0.6757	0.7296	0.7016	0.7781
0.5886	6.0	60	0.6775	{'precision': 0.648406374501992, 'recall': 0.8046971569839307, 'f1': 0.7181467181467182, 'number': 809}	{'precision': 0.25806451612903225, 'recall': 0.13445378151260504, 'f1': 0.17679558011049723, 'number': 119}	{'precision': 0.712947189097104, 'recall': 0.7859154929577464, 'f1': 0.7476552032157214, 'number': 1065}	0.6714	0.7546	0.7106	0.7890
0.5185	7.0	70	0.6770	{'precision': 0.6755888650963597, 'recall': 0.7799752781211372, 'f1': 0.7240390131956398, 'number': 809}	{'precision': 0.2079207920792079, 'recall': 0.17647058823529413, 'f1': 0.19090909090909092, 'number': 119}	{'precision': 0.7341337907375644, 'recall': 0.8037558685446009, 'f1': 0.7673688928731511, 'number': 1065}	0.6851	0.7566	0.7191	0.7955
0.4672	8.0	80	0.6729	{'precision': 0.683083511777302, 'recall': 0.788627935723115, 'f1': 0.7320711417096959, 'number': 809}	{'precision': 0.23300970873786409, 'recall': 0.20168067226890757, 'f1': 0.21621621621621623, 'number': 119}	{'precision': 0.747431506849315, 'recall': 0.819718309859155, 'f1': 0.7819077474249888, 'number': 1065}	0.6961	0.7702	0.7313	0.8007
0.4188	9.0	90	0.6664	{'precision': 0.6888888888888889, 'recall': 0.8046971569839307, 'f1': 0.74230330672748, 'number': 809}	{'precision': 0.2727272727272727, 'recall': 0.25210084033613445, 'f1': 0.26200873362445415, 'number': 119}	{'precision': 0.7708703374777975, 'recall': 0.8150234741784037, 'f1': 0.792332268370607, 'number': 1065}	0.7102	0.7772	0.7422	0.8045
0.3724	10.0	100	0.6845	{'precision': 0.6928721174004193, 'recall': 0.8170580964153276, 'f1': 0.7498581962563812, 'number': 809}	{'precision': 0.33, 'recall': 0.2773109243697479, 'f1': 0.30136986301369867, 'number': 119}	{'precision': 0.7818343722172751, 'recall': 0.8244131455399061, 'f1': 0.8025594149908593, 'number': 1065}	0.7221	0.7888	0.7540	0.8047
0.3402	11.0	110	0.6830	{'precision': 0.7118093174431203, 'recall': 0.8121137206427689, 'f1': 0.7586605080831409, 'number': 809}	{'precision': 0.3090909090909091, 'recall': 0.2857142857142857, 'f1': 0.296943231441048, 'number': 119}	{'precision': 0.787422497785651, 'recall': 0.8347417840375587, 'f1': 0.8103919781221514, 'number': 1065}	0.7308	0.7928	0.7605	0.8129
0.3219	12.0	120	0.6944	{'precision': 0.7179203539823009, 'recall': 0.8022249690976514, 'f1': 0.7577349678925861, 'number': 809}	{'precision': 0.3220338983050847, 'recall': 0.31932773109243695, 'f1': 0.32067510548523204, 'number': 119}	{'precision': 0.781882145998241, 'recall': 0.8347417840375587, 'f1': 0.8074477747502271, 'number': 1065}	0.7300	0.7908	0.7592	0.8097
0.3004	13.0	130	0.6978	{'precision': 0.7147540983606557, 'recall': 0.8084054388133498, 'f1': 0.7587006960556845, 'number': 809}	{'precision': 0.33043478260869563, 'recall': 0.31932773109243695, 'f1': 0.32478632478632474, 'number': 119}	{'precision': 0.7890974084003575, 'recall': 0.8291079812206573, 'f1': 0.8086080586080587, 'number': 1065}	0.7329	0.7903	0.7605	0.8144
0.2942	14.0	140	0.7001	{'precision': 0.7145945945945946, 'recall': 0.8170580964153276, 'f1': 0.7623990772779701, 'number': 809}	{'precision': 0.30708661417322836, 'recall': 0.3277310924369748, 'f1': 0.3170731707317073, 'number': 119}	{'precision': 0.7820284697508897, 'recall': 0.8253521126760563, 'f1': 0.8031064412973961, 'number': 1065}	0.7256	0.7923	0.7575	0.8108
0.2853	15.0	150	0.7024	{'precision': 0.7113513513513513, 'recall': 0.8133498145859085, 'f1': 0.7589388696655133, 'number': 809}	{'precision': 0.30952380952380953, 'recall': 0.3277310924369748, 'f1': 0.31836734693877555, 'number': 119}	{'precision': 0.7811387900355872, 'recall': 0.8244131455399061, 'f1': 0.8021927820922796, 'number': 1065}	0.7241	0.7903	0.7558	0.8106

Framework versions

• Transformers 4.25.1
• Pytorch 1.13.1+cu117
• Datasets 2.7.1
• Tokenizers 0.13.2