These workflow are intended to use SD1.5 models and Lora's to generate images at 8k - 16k quickly. please pay attention to the default values and if you build on top of them, feel free to share your work :) TwinAction-SuperUpscale This creates two images by using a feedback loop using the first image using tiled VAE, both go through super upscale 8X using your upscaler.(also available for SD2.1-768)

SD1.5 Comfy Workflows - with Super Upscale

Single action, Lora Support, Tiled VAE, 8X upscaler. uses no custom nodes, so you can easily build in your facedetailer if you want.

Files with SHA256: 20953718174561604b1f01d4b5b6ba1516bb6b4c8c1b083e7033fb36b6facd52