ComfyUI Node: Tiled Diffusion
Run Tiled Diffusion without installing ComfyUI
This node already runs on ComfyICU. Open it in the browser, hit run, and you're billed for GPU seconds — not for setup, and not while you're idle.
Category
_for_testing
Inputs
model MODEL
method
- MultiDiffusion
- Mixture of Diffusers
- SpotDiffusion
tile_width INT
tile_height INT
tile_overlap INT
tile_batch_size INT
Outputs
MODEL
Extension: Tiled Diffusion & VAE for ComfyUI
The extension enables large image drawing & upscaling with limited VRAM via the following techniques: 1.Two SOTA diffusion tiling algorithms: a/Mixture of Diffusers and a/MultiDiffusion 2.pkuliyi2015's Tiled VAE algorithm.
Authored by shiimizu
Looking for a different node?
Also provided by 1 other extension
More nodes in Tiled Diffusion & VAE for ComfyUI
Other _for_testing nodes
Run Tiled Diffusion without installing ComfyUI
This node already runs on ComfyICU. Open it in the browser, hit run, and you're billed for GPU seconds — not for setup, and not while you're idle.