So SDXL is a pretty cool diffusion model recently released by stability AI (https://huggingface.co/stabilityai/stable-diffusion-xl-base-...). It's common to train Low-Rank Adaptations (LoRAs) of this base model to generate art that has lower variance and elements that come from outside of the base model's training set (ex: your face, logos, etc.).
The problem is that the "easiest" ways of making these LoRAs is through the CLI, github script webUIs, or docker images on compute providers like Runpod. If your machine doesn't have a GPU with enough vRAM/power, then you pretty much have to pay to run the training script on some cloud compute.
It's a pretty terrible experience for non-technical people (even programmers like me lol).
So I built this app that lets you train regular + SDXL LoRAs in the browser, save them, share them, generate art with them, and configure them.
The problem is that the "easiest" ways of making these LoRAs is through the CLI, github script webUIs, or docker images on compute providers like Runpod. If your machine doesn't have a GPU with enough vRAM/power, then you pretty much have to pay to run the training script on some cloud compute.
It's a pretty terrible experience for non-technical people (even programmers like me lol).
So I built this app that lets you train regular + SDXL LoRAs in the browser, save them, share them, generate art with them, and configure them.
Let me know what you think :)