# Ucas-HaoranWei/Vary

[ECCV 2024] Official code implementation of Vary: Scaling Up the Vision Vocabulary of Large Vision Language Models.

Repository: https://github.com/Ucas-HaoranWei/Vary
Canonical: https://ross.abutalabs.com/products/vary
Language: Python
License Family: other
Last push: 2024-12-30T07:43:26+00:00

## Health v2 (maintenance only)
Score: 26/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 0, release rhythm 35, longevity 71
- inputs: {"age_days": 1000, "days_push": 611, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1889, forks 145 (observed 2026-08-28T04:05:49.391381+00:00)

## What it is
Official ECCV 2024 implementation of Vary, a method for scaling up the vision vocabulary of large vision-language models. It provides training and inference code for Vary models that perform dense document OCR, chart parsing, and visual understanding.

## Use cases
- parse text from document images with a vision-language model
- train a vision-language model with an expanded vision vocabulary
- run OCR on dense PDF page images
- extract charts into structured data
- reproduce ECCV 2024 Vary paper results

## When to choose
- you need a research-grade vision-language model for document OCR
- you want to experiment with vision vocabulary scaling for LVLMs
- you need English and Chinese document parsing

## When to avoid
- you need a production-supported library with a license and long-term maintenance
- you want a general-purpose chat assistant rather than document perception
- you lack GPU resources for large model inference

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, computer-vision, ocr, nlp, llm-training
- domain: computer-vision, large-language-models, deep-learning, image-processing, pdf
- platform: python
- tags: vision-language-model, eccv-2024, document-understanding, vision-vocabulary, research-code, gpu, linux

## Member repositories
- Ucas-HaoranWei/Vary (main) score 26

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:49.391381+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 2026-08-30T03:13:05.718119+00:00, confidence not recorded.
  - readme: https://github.com/Ucas-HaoranWei/Vary (fetched 2026-08-28T04:05:49.391381+00:00, sha bae06206a647)
- Data as of 2026-08-30T08:39:29.467469+00:00.