santa hat
deerdeer nosedeer glow
Sign In

Flux In Context - visual identity Lora in Comfy

Flux In Context - visual identity Lora in Comfy

So WizardWhitebeard showed me a cool trick - using https://github.com/ali-vilab/In-Context-LoRA to troll bblink787 's logo - I'm sure this can benefit many, either in creating useful mockups or just by trolling your friends!

The prompt you need for this lora to work is :

The pair of images highlights a logo and its real-world use for a hi-tech farming equipment; [IMAGE1] a black background showcases a logo with a stylized, fish in magenta and cyan, titled “BLINK” in an bold font, with bubble details underneath; [IMAGE2] this logo is applied as a black and white tattoo on lower back of an inmate

There's no need to actually use the regular lora loader bcz flux loras dont do clip (afaik)

Load the logo image in the workflow and it will do the rest for you.

The WF will take your image, duplicate it so there will be something to build on, mask what ever it required (the right frame) and img2img it with the the lora.
Then it will also cut that said image so you can have both the comparison and the new one as separate images.


If this brings you any value, please consider following and all that stuff :P

38

Comments