3 ms·
Show HN: Consumerizing Stable Diffusion
After generating our first image with Stable Diffusion, it was almost impossible to not drop everything we were doing and bust out a simple interface to consumerize the magic of this AI.
Upload your face, train a model, and make great prompts. The challenge we faced was scaling cloud GPUs for each user who trains a model while still keeping a profit margin. But we'd like to think we did it (to some degree), and would love for Hacker News and those that haven't already created a model to give it a go.
Will be a fun way to spend part of your Thanksgiving :)
- sech8420 4y agoWill give this a go. I saw an app kind of like this but the images did not look like me. Was very unsatisfied. And I know it was just released, but is it using Stable Diffusion 2?