This ComfyUI custom node provides an implementation of OmniGen, a versatile text-to-image and image editing model. It is designed for users of ComfyUI, offering advanced image generation capabilities with support for multiple input images and memory optimization.

How It Works

OmniGen integrates into the ComfyUI workflow, enabling text-to-image generation, image editing, and multi-image composition. The node handles the automatic downloading of necessary OmniGen code and model weights from Hugging Face, simplifying setup. It offers parameters for controlling inference steps, guidance scales (text and image), input image size, and output dimensions, along with memory optimization features like separate CFG inference and model offloading.

Quick Start & Requirements

• Installation: Install via ComfyUI Manager or by cloning the repository into ComfyUI/custom_nodes/ and running pip install -r requirements.txt.
• Prerequisites: ComfyUI, Python.
• Auto-Download: The node automatically downloads approximately 15.5 GB of model weights on first use. Manual download is also supported.
• Documentation: Example workflows are provided in the README.

Highlighted Details

• Supports text-to-image generation and image editing.
• Enables the use of multiple input images for complex prompts and image combinations.
• Includes memory optimization options such as separate_cfg_infer and offload_model.
• Allows flexible control over image size and output dimensions.

Maintenance & Community

The project has seen updates in late 2024 and early 2025. Specific community channels or contributor details are not provided in the README.

Licensing & Compatibility

The repository does not explicitly state a license. The model weights are hosted on Hugging Face under the Shitao/OmniGen-v1 repository, whose license should be consulted for commercial use or closed-source linking.

Limitations & Caveats

The initial setup requires downloading a substantial 15.5 GB model, which may be time-consuming. The README does not specify the license for the custom node code itself, which could impact commercial use.