Want to have your own Text to Image A.I.? Let's get one up and running on Ubuntu 24.01.1 (I had it running on 22.04 too). Let's install
FOOOCUS (yes that is really three 'O's). The
Fooocus project is built entirely on the Stable Diffusion XL architecture and uses a local web front end. Under the
ADVANCED option on the web front end you can easily just tick a box to change the style and there are a lot of them. You can also change the resolution of the generated images. For more information about
Fooocus or to find instructions to get it running on a Windows or Mac or different ways to install it on Linux check out their
GIT page.You should have a modern graphics card but you don't have to, it will run on CPU only but slowly. According to the
GIT Page you'll run about 17 times slower without a modern graphics card and you'll need a minimum of 32gb of ram. I'm using a Nvidia RTX 3060, 32gb ram, AMD 5900X CPU and generate an image in about 30 to 45 seconds depending on settings.
From their
GIT Page here's a list of GPUs, GPU's minimum ram, minimum system ram needed and what you can expect performance wise.
Nvidia on Linux/Windows
GPU Card - GPU - System - Note
RTX 4XXX - 4GB - 8GB - fastest
RTX 3XXX - 4GB - 8GB - faster than RTX 2XXX
RTX 2XXX - 4GB - 8GB - faster than GTX 1XXX
GTX 1XXX - 8GB - 8GB - only marginally faster than CPU
GTX 9XXX - 8GB - 8GB - faster or slower than CPU
Linux AMD GPU
GPU Card - GPU - System - Note
AMD GPU - 8GB - 8GB - Via ROCm, about 1.5x slower than Nvidia RTX 3XXX
Windows/Linux/Mac
GPU Card - GPU - System - Note
CPU Only - 0GB - 32GB - About 17x slower than Nvidia RTX 3XXX
Windows AMD GPU
GPU Card - GPU - System - Note
AMD GPU - 8GB - 8GB - Via DirectML (* ROCm is on hold), about 3x slower RTX 3XXX
MAC
GPU Card - GPU - System - Note
M1/M2 MPS - Shared - Shared - About 9x slower than Nvidia RTX 3XXX
SAMPLE IMAGE
Enough selling it, lets get it up and running.
We will be using the Linux (Using Python Venv) version. Your Linux needs to have Python 3.10 or greater installed, and your Python can be called with the command python3 with your venv system working;
Open a Terminal Window and enter the following command, waiting for each one to finish before executing the next one (some will take a few minutes to execute).
git clone https://github.com/lllyasviel/Fooocus.git
cd Fooocus
python3 -m venv fooocus_env
source fooocus_env/bin/activate
pip install -r requirements_versions.txt
Once that's done you launch the software with:
source fooocus_env/bin/activate
python entry_with_update.py
If you want to open a remote port;
source fooocus_env/bin/activate
python entry_with_update.py --listen
For additional information you should really read the GIT PAGE for FOOOCUS