Outpainting With Minimal Manual Input
5.0
0 reviews
Description
This workflow tries to minimise manual input needed by using Florence2 and IPAdapter style transfer to capture the content and style of the original image, and then use mask filling and Fooocus inpainting to extend the existing image beyond its borders.
(Note: over 12GB VRAM might be needed. Skipping the IPAdapter nodes can greatly reduce VRAM use.)
Discussion
(No comments yet)
Loading...
Resources (1)
Reviews
No reviews yet
Versions (2)
- latest (a year ago)
- v20240714-124448
Node Details
Primitive Nodes (6)
DownloadAndLoadFlorence2Model (1)
Florence2Run (1)
Reroute (4)
Custom Nodes (21)
ComfyUI
- VAEEncode (1)
- SetLatentNoiseMask (1)
- PreviewImage (2)
- CLIPTextEncode (2)
- VAEDecode (1)
- SaveImage (1)
- LoadImage (1)
- ImagePadForOutpaint (1)
- CheckpointLoaderSimple (1)
- KSampler (1)
- MaskPreview+ (1)
- INPAINT_ApplyFooocusInpaint (1)
- INPAINT_LoadFooocusInpaint (1)
- INPAINT_MaskedFill (1)
- IPAdapterUnifiedLoader (1)
- PrepImageForClipVision (1)
- IPAdapterAdvanced (1)
- ShowText|pysssss (1)
- SDXLPromptStyler (1)
Model Details
Checkpoints (1)
zavychromaxl_v90.safetensors
LoRAs (0)