Outputs
Description
Generate 5 images using Lightning LoRAs using SDXL models (not SDXL lightning models). Then choose one of those 5 for Hi-Res, Upscale, and Detailing. The 5 (low-quality) images will be saved in a sub-folder for quick viewing after running a batch, then you can refine which ever you like by simply reloading that image workflow and selecting it for processing.
Built-in nodes
Custom nodes
BNK_CLIPTextEncodeAdvanced
VAELoader
KSampler
MathExpression|pysssss
VAEDecode
PreviewImage
LoraLoaderModelOnly
LoraLoader
FaceDetailer
MaskToImage
UltralyticsDetectorProvider
MediaPipe-FaceMeshPreprocessor
MediaPipeFaceMeshToSEGS
SegsToCombinedMask
ImpactSEGSToMaskList
Convert Masks to Images
MaskToSEGS
BboxDetectorSEGS
UpscaleModelLoader
DetailerForEachDebug
SaveImage
UltimateSDUpscale
CLIPSetLastLayer
ImageUpscaleWithModel
ImageScaleBy
VAEEncode
CR SDXL Aspect Ratio
LoRA Stacker
Text _O
CR Apply LoRA Stack
Seed Generator
int _O
Cfg Literal
PrimitiveNode
ImpactSwitch
CheckpointLoaderSimple
LatentSwitch
Models
sdxl_vae.safetensors