# apple/ml-mgie

Repository: https://github.com/apple/ml-mgie
Canonical: https://ross.abutalabs.com/products/ml-mgie
Language: Python
License: NOASSERTION
License Family: other
Last push: 2024-03-15T21:35:14+00:00

## Health v2 (maintenance only)
Score: 26/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 35, longevity 70
- inputs: {"age_days": 992, "days_push": 901, "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 3874, forks 251 (observed 2026-08-28T04:08:27.873907+00:00)

## What it is
MGIE (MLLM-Guided Image Editing) is Apple's research implementation of instruction-based image editing guided by multimodal large language models, from an ICLR 2024 Spotlight paper. It uses an MLLM (LLaVA-based) to derive expressive editing instructions from brief natural-language commands and performs image manipulation end-to-end with diffusion models.

## Use cases
- edit images with natural language instructions
- research multimodal LLM guided image manipulation
- reproduce MGIE paper results
- build instruction-based photo editing demos
- explore MLLM visual-aware response generation

## When to choose
- you need research-grade instruction-based image editing with MLLM guidance
- you want to build on or study the MGIE ICLR'24 method
- you have GPU resources and want a LLaVA + diffusion editing pipeline

## When to avoid
- you need a production-ready or actively maintained image editor
- you lack a CUDA GPU or want lightweight CPU inference
- you need a permissively licensed model for commercial use

## Facets
- artifact type: library
- maturity: experimental
- function: image-processing, machine-learning, deep-learning, llm-inference
- domain: computer-vision, image-processing, large-language-models, artificial-intelligence, deep-learning
- platform: python
- tags: instruction-based-image-editing, multimodal-llm, diffusion, iclr-2024, research-code, apple, gpu, linux

## Member repositories
- apple/ml-mgie (main) score 26

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:08:27.873907+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 2026-08-29T18:25:46.510035+00:00, confidence not recorded.
  - readme: https://github.com/apple/ml-mgie (fetched 2026-08-28T04:08:27.873907+00:00, sha 53e2f1e26c9f)
- Data as of 2026-08-30T08:39:29.467469+00:00.