Training Loras for Flux 1.0 [dev] 

I love open-source - and non-profit projects based on open-source.

Tost.AI is a site where you can try out new models and pipelines like Live Portrait.

You can train your own LoRa and create cool stuff. With just 6 photos, you can consistently generate the same object. Details of the object are captured so accurately that even text remains intact.

E.g. you can take photos in a white box and create a product photo with a model (or on a model), or an image in some unique location, etc.

Worth noting that, even though the weights are under a non-profit license, you fully own the images generated by Flux.

i like pretending i am smart. /lovejoy by day; yoink anytime. we /honk: goosegang.netlify.app

serial /yoink'r building in the /arweave ecosystem with PDS.inc

your dive into coding is v inspiring 300 $DEGEN

😀 I'm not used to registering with mail.
I like to just link my wallet and that's it)