RAM kot virtualni VRAM?

Glej "cuda unified memory" feature.

se vidva spoznata na Stable Diffusion app?
to je treba skonfigurirat nekaj, ne? - na roke.
sam jaz niti ne vem kje (.BAT al .CMD al kje drugje) niti kako.
se to dela v notepadu al v pythonu?
vidim da je python priložen.
sam ne vem kva naredit.

okey. dobil sem odgovor v smislu napotka kje naj najdem koristne informacije.
bom c/p:

""
Hi @Ijus666 Apologies if we missed your message! To be honest, I don't know enough about CUDA to comment, but as far as I understand the way SD is written, I think it either works with GPU memory, or CPU+RAM, and cannot combine the two.

If you'd like to dig into this, you can try reading https://developer.nvidia.com/blog/unifi... and see if it applies to the stable diffusion code: https://github.com/basujindal/stable-di...
""

a mi vi lahko kaj pomagate okol zadnjega odstavka?
jaz nimam čisto nič pojma o programiranju.

dobro. okey bom probal. hvala za opazko.
jaz tega nisem videl.

sem pa še nekaj izvedel za Stable Diffusion AI app:
kaj je treba disejblat pa se veliko VRAM-a sprosti.
okey.

hvala za ta info.

poročam.

Pise, da te optimizirane skopiras v originalno mapo.

Installation

All the modified files are in the optimizedSD folder, so if you have already cloned the original repository you can just download and copy this folder into the original instead of cloning the entire repo. You can also clone this repo and follow the same installation steps as the original (mainly creating the conda environment and placing the weights at the specified location).

Alternatively, if you prefer to use Docker, you can do the following:

Install Docker, Docker Compose plugin, and NVIDIA Container Toolkit
Clone this repo to, e.g., ~/stable-diffusion
Put your downloaded model.ckpt file into ~/sd-data (it's a relative path, you can change it in docker-compose.yml)
cd into ~/stable-diffusion and execute docker compose up --build

This will launch gradio on port 7860 with txt2img. You can also use docker compose run to execute other Python scripts.