ComfyUI custom node for NVIDIA PiD pixel diffusion decoding and upscale workflows
-
Updated
Jul 17, 2026 - Python
ComfyUI custom node for NVIDIA PiD pixel diffusion decoding and upscale workflows
PiD-Image-Upscaler is an experimental, advanced super-resolution and image-to-image refinement application based on the state-of-the-art PiD (Pixel Diffusion Decoder) framework by NVIDIA. This application couples local generative text-to-image and image-to-image pipelines.
Tile the latent, not the pixels: native ComfyUI nodes for latent-native tiled NVIDIA PiD decoding. Seam-free 33-107MP masters; tested on Krea 2 Turbo.
Latent-native tiled decoding for NVIDIA PiD: 33MP masters from 2MP latents on stock ComfyUI, no custom nodes. Measures and gates the past-envelope color collapse.
Add a description, image, and links to the nvidia-pid topic page so that developers can more easily learn about it.
To associate your repository with the nvidia-pid topic, visit your repo's landing page and select "manage topics."