lmv2-g-aadhaar-236doc-06-14 Open Source Model - Efficiently Extract Key Information from Documents Such as Aadhaar Cards

Lmv2 G Aadhaar 236doc 06 14

Developed by Sebabrata

This model is a fine-tuned version based on microsoft/layoutlmv2-base-uncased, specializing in document information extraction tasks, excelling in extracting fields such as Aadhaar card numbers, date of birth, gender, and names.

Sequence Labeling

Transformers

#Document Information Extraction #High-precision Recognition #Identity Verification

Downloads 52

Release Time : 6/14/2022

Model Overview

A document information extraction model based on the LayoutLMv2 architecture, specifically designed to extract key field information from structured documents, such as ID numbers, date of birth, gender, and names.

Model Features

High-precision Information Extraction

Achieves high precision and recall rates on key fields such as Aadhaar numbers, date of birth, gender, and names.

Based on LayoutLMv2 Architecture

Utilizes text and layout information for joint modeling to enhance document understanding capabilities.

Multi-field Joint Recognition

Capable of simultaneously recognizing various types of information fields within documents.

Model Capabilities

Document Information Extraction

Structured Data Recognition

ID Information Parsing

Use Cases

Identity Verification

Aadhaar Card Information Extraction

Extracts key information from Indian Aadhaar ID cards

Aadhaar number extraction F1 score reaches 0.9890

Document Processing

Date of Birth Recognition

Accurately identifies date of birth information from documents

Date of birth extraction F1 score reaches 0.9892

Personal Information Extraction

Extracts personal information such as name and gender from documents

Name extraction F1 score 0.9474, gender extraction F1 score 0.9892

🚀 lmv2-g-aadhaar-236doc-06-14

This model is a fine - tuned version of [microsoft/layoutlmv2 - base - uncased](https://huggingface.co/microsoft/layoutlmv2 - base - uncased) on the None dataset. It offers high - precision results in various aspects such as Aadhaar, Dob, Gender, and Name recognition, providing reliable performance for related tasks.

🚀 Quick Start

This model is a fine - tuned version of [microsoft/layoutlmv2 - base - uncased](https://huggingface.co/microsoft/layoutlmv2 - base - uncased) on the None dataset. It achieves the following results on the evaluation set:

Loss: 0.0427
Aadhaar Precision: 0.9783
Aadhaar Recall: 1.0
Aadhaar F1: 0.9890
Aadhaar Number: 45
Dob Precision: 0.9787
Dob Recall: 1.0
Dob F1: 0.9892
Dob Number: 46
Gender Precision: 1.0
Gender Recall: 0.9787
Gender F1: 0.9892
Gender Number: 47
Name Precision: 0.9574
Name Recall: 0.9375
Name F1: 0.9474
Name Number: 48
Overall Precision: 0.9785
Overall Recall: 0.9785
Overall F1: 0.9785
Overall Accuracy: 0.9939

📚 Documentation

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 4e - 05
train_batch_size: 1
eval_batch_size: 1
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon = 1e - 08
lr_scheduler_type: constant
num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Aadhaar Precision	Aadhaar Recall	Aadhaar F1	Aadhaar Number	Dob Precision	Dob Recall	Dob F1	Dob Number	Gender Precision	Gender Recall	Gender F1	Gender Number	Name Precision	Name Recall	Name F1	Name Number	Overall Precision	Overall Recall	Overall F1	Overall Accuracy
1.0024	1.0	188	0.5819	0.9348	0.9556	0.9451	45	1.0	1.0	1.0	46	1.0	0.9574	0.9783	47	0.5172	0.625	0.5660	48	0.8410	0.8817	0.8609	0.9744
0.4484	2.0	376	0.3263	0.8980	0.9778	0.9362	45	1.0	1.0	1.0	46	1.0	0.9787	0.9892	47	0.6842	0.8125	0.7429	48	0.8838	0.9409	0.9115	0.9733
0.2508	3.0	564	0.2230	0.9318	0.9111	0.9213	45	1.0	1.0	1.0	46	1.0	0.9787	0.9892	47	0.8913	0.8542	0.8723	48	0.9560	0.9355	0.9457	0.9811
0.165	4.0	752	0.1728	0.9362	0.9778	0.9565	45	1.0	1.0	1.0	46	1.0	0.9787	0.9892	47	0.8444	0.7917	0.8172	48	0.9457	0.9355	0.9405	0.9844
0.1081	5.0	940	0.0987	0.8958	0.9556	0.9247	45	1.0	1.0	1.0	46	1.0	0.9787	0.9892	47	1.0	0.9167	0.9565	48	0.9728	0.9624	0.9676	0.9928
0.0834	6.0	1128	0.0984	0.8980	0.9778	0.9362	45	1.0	1.0	1.0	46	1.0	0.9574	0.9783	47	0.8148	0.9167	0.8627	48	0.9227	0.9624	0.9421	0.9833
0.0676	7.0	1316	0.0773	0.9362	0.9778	0.9565	45	1.0	1.0	1.0	46	1.0	0.9787	0.9892	47	0.9111	0.8542	0.8817	48	0.9620	0.9516	0.9568	0.9894
0.0572	8.0	1504	0.0786	0.8235	0.9333	0.8750	45	1.0	1.0	1.0	46	1.0	0.9787	0.9892	47	0.8936	0.875	0.8842	48	0.9263	0.9462	0.9362	0.9872
0.0481	9.0	1692	0.0576	0.9375	1.0	0.9677	45	1.0	1.0	1.0	46	1.0	0.9787	0.9892	47	0.9362	0.9167	0.9263	48	0.9679	0.9731	0.9705	0.99
0.0349	10.0	1880	0.0610	0.9574	1.0	0.9783	45	1.0	1.0	1.0	46	1.0	0.9787	0.9892	47	0.8958	0.8958	0.8958	48	0.9626	0.9677	0.9651	0.9894
0.0287	11.0	2068	0.0978	0.9091	0.8889	0.8989	45	1.0	1.0	1.0	46	1.0	0.9787	0.9892	47	0.9348	0.8958	0.9149	48	0.9615	0.9409	0.9511	0.985
0.0297	12.0	2256	0.0993	0.9375	1.0	0.9677	45	1.0	1.0	1.0	46	1.0	0.9787	0.9892	47	0.7959	0.8125	0.8041	48	0.9312	0.9462	0.9387	0.9833
0.0395	13.0	2444	0.0824	0.9362	0.9778	0.9565	45	1.0	1.0	1.0	46	1.0	0.9787	0.9892	47	0.875	0.875	0.875	48	0.9519	0.9570	0.9544	0.9872
0.0333	14.0	2632	0.0788	0.8913	0.9111	0.9011	45	1.0	1.0	1.0	46	1.0	0.9787	0.9892	47	0.9556	0.8958	0.9247	48	0.9617	0.9462	0.9539	0.9867
0.0356	15.0	2820	0.0808	0.84	0.9333	0.8842	45	1.0	1.0	1.0	46	1.0	0.9787	0.9892	47	0.9565	0.9167	0.9362	48	0.9468	0.9570	0.9519	0.9867
0.0192	16.0	3008	0.0955	0.8462	0.9778	0.9072	45	0.9787	1.0	0.9892	46	0.9583	0.9787	0.9684	47	0.9070	0.8125	0.8571	48	0.9211	0.9409	0.9309	0.9822
0.016	17.0	3196	0.0936	0.9130	0.9333	0.9231	45	1.0	1.0	1.0	46	1.0	0.9787	0.9892	47	0.9318	0.8542	0.8913	48	0.9615	0.9409	0.9511	0.9867
0.0218	18.0	3384	0.1009	0.9545	0.9333	0.9438	45	1.0	1.0	1.0	46	1.0	0.9787	0.9892	47	0.8571	0.875	0.8660	48	0.9514	0.9462	0.9488	0.9844
0.0165	19.0	3572	0.0517	0.9574	1.0	0.9783	45	1.0	1.0	1.0	46	1.0	0.9787	0.9892	47	0.9333	0.875	0.9032	48	0.9728	0.9624	0.9676	0.9906
0.0198	20.0	3760	0.0890	0.9167	0.9778	0.9462	45	1.0	1.0	1.0	46	1.0	0.9787	0.9892	47	0.9149	0.8958	0.9053	48	0.9572	0.9624	0.9598	0.9867
0.0077	21.0	3948	0.0835	0.9574	1.0	0.9783	45	1.0	1.0	1.0	46	1.0	0.9787	0.9892	47	0.88	0.9167	0.8980	48	0.9577	0.9731	0.9653	0.9872
0.0088	22.0	4136	0.0427	0.9783	1.0	0.9890	45	0.9787	1.0	0.9892	46	1.0	0.9787	0.9892	47	0.9574	0.9375	0.9474	48	0.9785	0.9785	0.9785	0.9939
0.0078	23.0	4324	0.0597	0.9574	1.0	0.9783	45	1.0	1.0	1.0	46	1.0	0.9787	0.9892	47	0.8654	0.9375	0.9	48	0.9529	0.9785	0.9655	0.9889
0.0178	24.0	4512	0.0524	0.9574	1.0	0.9783	45	1.0	1.0	1.0	46	1.0	0.9787	0.9892	47	1.0	0.875	0.9333	48	0.9890	0.9624	0.9755	0.9922
0.012	25.0	4700	0.0637	0.9375	1.0	0.9677	45	1.0	1.0	1.0	46	1.0	0.9787	0.9892	47	0.8491	0.9375	0.8911	48	0.9430	0.9785	0.9604	0.9867
0.0135	26.0	4888	0.0668	0.9184	1.0	0.9574	45	1.0	1.0	1.0	46	1.0	0.9787	0.9892	47	0.86	0.8958	0.8776	48	0.9424	0.9677	0.9549	0.9867
0.0123	27.0	5076	0.0713	0.9565	0.9778	0.9670	45	1.0	1.0	1.0	46	1.0	0.9787	0.9892	47	0.9375	0.9375	0.9375	48	0.9731	0.9731	0.9731	0.9911
0.0074	28.0	5264	0.0675	0.9362	0.9778	0.9565	45	1.0	1.0	1.0	46	1.0	0.9787	0.9892	47	0.9	0.9375	0.9184	48	0.9577	0.9731	0.9653	0.99
0.0051	29.0	5452	0.0713	0.9362	0.9778	0.9565	45	1.0	1.0	1.0	46	1.0	0.9787	0.9892	47	0.9167	0.9167	0.9167	48	0.9626	0.9677	0.9651	0.9906
0.0027	30.0	5640	0.0725	0.9362	0.9778	0.9565	45	1.0	1.0	1.0	46	1.0	0.9787	0.9892	47	0.9167	0.9167	0.9167	48	0.9626	0.9677	0.9651	0.9906

Framework versions

Transformers 4.20.0.dev0
Pytorch 1.11.0+cu113
Datasets 2.2.2
Tokenizers 0.12.1

📄 License

This project is licensed under the CC - BY - NC - SA 4.0 license.

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご