S

Structtable InternVL2 1B

Developed by U4R
A multimodal table recognition model based on InternVL2-1B, supporting conversion of table images to LaTeX/HTML/Markdown formats
Downloads 1,833
Release Time : 10/18/2024

Model Overview

This model can accurately extract structured data representations from visual table images, supporting multiple table format conversions and table-related reasoning tasks

Model Features

Multi-format output
Supports converting table images to three common formats: LaTeX, HTML, and Markdown
Efficient inference
Significant inference speed improvement achieved through optimization
Large-scale training data
Trained on DocGenome benchmark and synthetic data, containing over 2 million high-quality image-LaTeX pairs
Cross-domain applicability
Covers table data from 156 subject categories, with broad applicability

Model Capabilities

Table image recognition
Table structure extraction
LaTeX generation
HTML generation
Markdown generation
Table question answering

Use Cases

Academic publishing
Paper table conversion
Convert scanned paper tables into editable LaTeX format
Improves academic writing efficiency and reduces manual input errors
Enterprise applications
Financial statement processing
Automatically recognize financial statement images and convert them into structured data
Simplifies financial data digitization process
Web development
Web table reconstruction
Convert tables from design drafts into HTML code
Accelerates front-end development workflow
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
© 2025AIbase