In-Context LoRA
Added 2024-11-16 16:39:17 +0000 UTCVideo - https://youtu.be/n2QZEtBU89o
In-Context loras help keep things... in context!
If you get Flux to generate a single, large image and prompt for various parts of that image (such as top-left, bottom-right, etc), then you can crop the final image into a number of smaller (but hopefully consistent) images.
IC-LoRAs can help in such situations, and Ali Vilab has provided 10 examples along with information about how to train your own :)
Links
Download files from - https://huggingface.co/ali-vilab/In-Context-LoRA
GitHub page - https://github.com/ali-vilab/In-Context-LoRA
Paper - https://arxiv.org/abs/2410.23775
Workflow Information
Workflows use FLUX (file links and example - https://comfyanonymous.github.io/ComfyUI_examples/flux/ ) and the In-Context LoRA files.
The "IMGOP" version includes the image outpainting example workflow, with a simple left-right split, best used with the "visual identity" or other "left/right" LoRAs.
Comments
All of these work really well. Thanks for sharing.
Nibmeister
2024-11-16 21:53:19 +0000 UTC
