Most artists who are not tech savvy to spin up their local Stable diffusion instances are probably going to pay and use one of the online services like DALL-E 2 or something.
Packaging Python is a dumpster fire of local environment complexity. But this is an extraordinarily solvable problem. I won’t call it solved today. But we have Proof of Life of solutions.
And yeah it’s not super duper hard. It’s just not how Python projects are structured by default. Simple Diffusion isn’t setup that way. Some random person on the internet spent some time to package it up nicely and in exchange they hit #1 on HN. Possibly more than once!
- download the Python embedded environment,
- modify `pythonXX.pth` to `import site`,
- install pip in the environment,
- install all dependencies with pip,
- install your code as an executable module (with a `__main__.py` file).
You can ship this as a ZIP file and simply run `./python.exe -m your_module arg1 arg2 arg3...`
Nvidia has primarily been associated with gaming. With Machine Learning that’s starting to change. With tools like Stable Diffusion every single artist is going to want a $1600 GPU and honestly that’s a bargain for them!
I think it’d be sweet if you could spend $100 to $10,000 and get a linear increases in cores, ram, and performance. That’s not how it works. But philosophically I have no problem with it.
Hopefully AMD steps up and brings some real competition.