Remake for SDXL : Removing Object and Filling with Background
5.0
0 reviews
Description
I have remade the “Remove Object and Fill with Background” , which many of you enjoy, for SDXL. It should now be more flexible and of higher quality.
Improvements:
- Mask generation by SAM before resizing
- Fooocus Inpaint
- Replaced ControlNet Inpaint with this
- Thanks to lllyasviel, who developed Fooocus Inpaint, and Acly, who introduced it to ComfyUI
- IP-Adapter
Future Improvements:
- I am currently using WD14-tagger as image2prompt, but it is not suitable for realistic photos.
- If MLLMs like MoE-LLaVa become more accessible, we can expect further development.
Discussion
(No comments yet)
Loading...
Reviews
No reviews yet
Versions (1)
- latest (a year ago)
Node Details
Primitive Nodes (4)
IPAdapterApply (1)
Reroute (3)
Custom Nodes (25)
ComfyUI
- CheckpointLoaderSimple (1)
- KSampler (1)
- VAEDecode (1)
- GrowMask (1)
- CLIPTextEncode (2)
- CLIPVisionLoader (1)
- JoinImageWithAlpha (1)
- ImageScale (1)
- LoadImage (1)
- SaveImage (2)
- InvertMask (1)
- INPAINT_VAEEncodeInpaintConditioning (1)
- INPAINT_ApplyFooocusInpaint (1)
- INPAINT_LoadFooocusInpaint (1)
- INPAINT_LoadInpaintModel (1)
- INPAINT_InpaintWithModel (1)
- WD14Tagger|pysssss (1)
- IPAdapterModelLoader (1)
- ImageResize (1)
- SAMModelLoader (segment anything) (1)
- GroundingDinoModelLoader (segment anything) (1)
- GroundingDinoSAMSegment (segment anything) (1)
- Image Size to Number (1)
Model Details
Checkpoints (1)
📷-XL\RealismEngineSDXL_V2.0_FP16_VAE.safetensors
LoRAs (0)