ssitu/ComfyUI_UltimateSDUpscale

ComfyUI nodes for the Ultimate Stable Diffusion Upscale script by Coyote-A.

What it solves

This project provides a way to upscale images using diffusion models without running into hardware limitations or losing detail. Standard upscaling often struggles with large images because they exceed the memory capacity of the GPU or deviate from the original training size of the diffusion model.

How it works

It implements an image-to-image diffusion process that operates on large images in tiles. By breaking the image into smaller, manageable sections, it can enhance details and increase resolution while keeping the image size within the range the diffusion model was trained on.

Who it’s for

Users of ComfyUI who want to generate high-resolution images with better detail than standard upscaling methods.

Highlights

  • Custom nodes for ComfyUI integrated into the image/upscaling menu.
  • Tiled diffusion process to reduce hardware requirements.
  • Support for high-detail upscaling while maintaining model training constraints.

Related

  • Project
    robertvoy/ComfyUI-Distributed

    A ComfyUI extension that enables parallel image and video generation and distributed upscaling across multiple local or cloud GPUs.

  • Project
    adobe-research/custom-diffusion

    A fast and efficient method for fine-tuning text-to-image diffusion models to learn new concepts from a few images with minimal storage overhead.

  • Project
    Acly/comfyui-tooling-nodes

    A set of ComfyUI custom nodes and API extensions that enable the use of ComfyUI as a backend for external tools, featuring optimized image transfer, regional prompting, and model inspection.

  • Project
    wildminder/ComfyUI-DyPE

    ComfyUI‑DyPE is a set of custom ComfyUI nodes that let you generate ultra‑high‑resolution images (2 K‑4 K+) with modern diffusion‑transformer models (Flux, Qwen‑Image, Z‑Image, etc.) without retraining. It provides training‑free patches (DyPE, SEGA, SPA, HAP) and cascade up‑scalers (PixelRush, FreeScale, HiFlow, I‑Max) that integrate as a single node after the model loader, offering quality, speed, or content‑preserving options.

  • Project
    Jasonzzt/ComfyUI-CacheDiT

    A ComfyUI extension that accelerates Diffusion Transformer (DiT) models through intelligent caching, providing 1.4x to 2.0x speedups for image and video generation.