# harvardnlp/im2markup

Neural model for converting Image-to-Markup (by Yuntian Deng yuntiandeng.com)

Repository: https://github.com/harvardnlp/im2markup
Canonical: https://ross.abutalabs.com/products/im2markup
Homepage: https://im2markup.yuntiandeng.com
Language: Lua
License: MIT
License Family: permissive
Last push: 2023-10-27T13:42:36+00:00

## Health v2 (maintenance only)
Score: 32/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 3638, "days_push": 1041, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1259, forks 215 (observed 2026-08-28T04:04:09.851324+00:00)

## What it is
A deep learning system (built on Torch) that converts images of rendered text into presentational markup such as LaTeX or HTML, using a CNN encoder with attention-based neural machine translation. It is the reference implementation of the 'What You Get Is What You See' image-to-markup decompiler paper.

## Use cases
- convert images of math formulas to latex
- image to markup generation
- ocr for mathematical expressions
- extract latex from rendered formula screenshots
- train an image-to-text sequence model
- research on attention-based image captioning of markup

## When to choose
- you need to decompile rendered LaTeX or HTML images back into source markup
- you are reproducing or extending the im2markup research paper
- you need attention alignment visualizations between output markup and input image

## When to avoid
- you need a maintained production OCR tool - the project depends on the deprecated Torch framework
- you have no GPU, since only GPU execution is supported
- you want a plug-and-play API rather than a research codebase

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, deep-learning, ocr, image-processing, nlp
- domain: deep-learning, computer-vision, artificial-intelligence
- platform: python, lua
- tags: image-to-latex, encoder-decoder, attention, torch, math-ocr, markup-generation, natural-language-processing, gpu, linux

## Member repositories
- harvardnlp/im2markup (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:09.851324+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T05:05:30.119310+00:00, confidence not recorded.
  - readme: https://github.com/harvardnlp/im2markup (fetched 2026-08-28T04:04:09.851324+00:00, sha 862fa8543356)
  - homepage: https://im2markup.yuntiandeng.com (fetched 2026-08-29T12:17:00.301910+00:00, sha b28f073da72c)
- Data as of 2026-08-30T08:39:29.467469+00:00.