# JIA-Lab-research/DreamOmni2

This project is the official implementation of 'DreamOmni2: Multimodal Instruction-based Editing and Generation  (CVPR2026 Highlight)''

Repository: https://github.com/JIA-Lab-research/DreamOmni2
Canonical: https://ross.abutalabs.com/products/dreamomni2
Language: Python
License: Apache-2.0
License Family: permissive
Topics: image-editing, image-generation, unified-generation-editing-model
Last push: 2026-04-11T14:51:35+00:00

## Health v2 (maintenance only)
Score: 51/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 76, release rhythm 35, longevity 24
- inputs: {"age_days": 339, "days_push": 144, "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 1978, forks 173 (observed 2026-08-28T04:06:01.550817+00:00)

## What it is
DreamOmni2 is the official PyTorch implementation of a CVPR 2026 Highlight model for multimodal instruction-based image editing and generation. It supports subject-driven generation referencing concrete objects or abstract attributes, and editing guided by both text instructions and reference images.

## Use cases
- edit images with text and reference image instructions
- generate images of a subject with consistent identity and pose
- transfer abstract attributes like style, texture, or makeup to images
- run instruction-based image editing locally
- benchmark multimodal editing models with DreamOmni2Bench
- integrate image editing into ComfyUI workflows

## When to choose
- you need open-source instruction-based editing that accepts reference images
- you want subject-driven generation with strong identity and pose consistency
- you need to reproduce or build on a published research model
- you want to reference abstract attributes like artistic style or material

## When to avoid
- you need a polished end-user photo editor with a GUI
- you lack a GPU or can't run large diffusion models
- you need lightweight or real-time image editing
- you need a commercially supported product rather than research code

## Facets
- artifact type: library
- maturity: active
- function: image-processing, machine-learning, deep-learning
- domain: image-processing, artificial-intelligence, computer-vision
- platform: python
- tags: image-editing, image-generation, diffusion-model, instruction-based-editing, subject-driven-generation, multimodal, research-model, comfyui, gpu, linux

## Member repositories
- JIA-Lab-research/DreamOmni2 (main) score 51

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:01.550817+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 2026-08-30T03:04:43.874055+00:00, confidence not recorded.
  - readme: https://github.com/JIA-Lab-research/DreamOmni2 (fetched 2026-08-28T04:06:01.550817+00:00, sha 5a6fbc9bafe6)
- Data as of 2026-08-30T08:39:29.467469+00:00.