ComfyUI Plugin
2.9k 2026-05-01
nunchaku-ai/ComfyUI-nunchaku
A ComfyUI plugin that integrates Nunchaku, an efficient inference engine for 4-bit quantized neural networks, to accelerate AI model execution.
Core Features
Seamless integration with ComfyUI workflows.
Accelerated inference for 4-bit quantized neural networks.
Support for various advanced AI models like Z-Image-Turbo and Qwen-Image-Edit.
Significant performance boosts and reduced VRAM usage (e.g., 20-30% Z-Image boost, asynchronous offloading).
Compatibility with LoRA and INT4 support for specific GPUs.
Detailed Introduction
ComfyUI-nunchaku is a crucial plugin that brings the power of Nunchaku, an efficient inference engine, directly into the ComfyUI ecosystem. Nunchaku specializes in accelerating 4-bit neural networks quantized with SVDQuant, enabling users to run complex AI models with significantly reduced computational overhead and VRAM consumption. This integration allows ComfyUI users to leverage state-of-the-art quantized models for tasks like image generation and editing, achieving faster results and more efficient resource utilization, even on less powerful hardware.