Style Transfer with Face Swap v2
created a year ago
img2img
ipadapter
upscale
controlnet
face
lora
116 nodes
2.3k
436
Credits
https://www.youtube.com/watch?v=czcgJnoDVd4
Outputs









Description
Workflow for style transfer to pose with face swap.
The idea was to create an easy to use workflow for transfering style and face to a fullbody pose. Inputs are pose reference image, style image, images with face and short prompts for overall composition and face.
Should work with any SDXL model.
Additional work is required to fix hands.
Built-in nodes
Custom nodes
EmptyLatentImage
IPAdapterNoise
ImageUpscaleWithModel
UltralyticsDetectorProvider
ImageScaleBy
VAEEncode
VAEDecode
PreviewImage
IPAdapterAdvanced
ControlNetApplyAdvanced
BboxDetectorSEGS
UpscaleModelLoader
ControlNetLoader
PrepImageForClipVision
LoadImage
SetLatentNoiseMask
MaskPreview+
GetImageSize+
SAMModelLoader (segment anything)
GroundingDinoModelLoader (segment anything)
GroundingDinoSAMSegment (segment anything)
GrowMask
MaskComposite
SegsToCombinedMask
OpenposePreprocessor
ImageResize+
EmptyImage
ImageCompositeMasked
CM_NearestSDXLResolution
IPAdapterModelLoader
ApplyInstantID
InstantIDFaceAnalysis
InstantIDModelLoader
IPAdapterStyleComposition
KSampler
CLIPTextEncode
Zoe_DepthAnythingPreprocessor
CLIPVisionLoader
CheckpointLoaderSimple
LoraLoader
SaveImage
DetailerForEachDebug
LoadImagesFromDir //Inspire
Custom files
new
These are separate files that the creator has uploaded for this workflow.

ComfyUI_00166_ copy1.png

ComfyUI_00188_ copy1.png

ComfyUI_00190_ copy1.png

ComfyUI_00194_ copy1.png

ComfyUI_00199_ copy1.png

ComfyUI_00200_ copy1.png

ComfyUI9__00009_ copy1.png

ComfyUI9__00033_ copy1.png
Models