
Powered By RTX 4090
This workflow has only been tested more than 10 times and does not guarantee a better effect, but currently, the test results are slightly better than before. Please refer to the video tutorial for details.
T8star-Aix
2025-05-06 Update
T8star-Aix
2025-05-06 Update
Workflow introduction
This workflow has only been tested more than 10 times and does not guarantee a better effect, but currently, the test results are slightly better than before. Please refer to the video tutorial for details.
Nodes Information
48
CLIPTextEncode
CLIPVisionEncode
CheckpointLoaderSimple
DualCLIPLoader
InpaintModelConditioning
KSampler
LoadImage
LoraLoaderModelOnly
PreviewImage
SaveImage
StyleModelApply
StyleModelLoader
VAEDecode
VAELoader
AdvancedVisionLoader
ApplyFBCacheOnModel
DifferentialDiffusion
DisplayText_Zho
FluxGuidance
GetImageSize+
ImageConcanate
ImageCrop+
ImitationHueNode
ImpactSwitch
InvertMask
JWImageResizeByLongerSide
JWInteger
LayerMask: BiRefNetUltraV2
LayerMask: LoadBiRefNetModelV2
LayerMask: LoadFlorence2Model
LayerUtility: Florence2Image2Prompt
LayerUtility: ImageScaleByAspectRatio V2
LayerUtility: PurgeVRAM V2
LayerUtility: TextJoin
MaskPreview+
MathExpression|pysssss
Note
Note Plus (mtb)
PMRF
PrimitiveNode
Replace Text _O
Reroute
String Literal
TTP_Expand_And_Mask
TeaCache
easy showAnything
image_concat_mask(rh)
preview_mask