Credits
inspired by a conversation with NeuraLunk
Outputs
Description
This workflow allows you to create larger scale SDXL images by using Cascade as the base renderer then running a refinement pass through an SDXL model. It also improves fine details and clutter found in the original Cascade render.
Built-in nodes
CLIPLoader
VAELoader
StableCascade_EmptyLatentImage
StableCascade_StageB_Conditioning
KSampler
ConditioningZeroOut
VAEDecode
CLIPTextEncode
VAEEncode
FreeU
CheckpointLoaderSimple
UNETLoader
Custom nodes
CR Prompt Text
BboxDetectorSEGS
UltralyticsDetectorProvider
DetailerForEach
CR Image Output
CascadeResolutions
Seed Everywhere
Models