Autolume-Live is a tool for real-time visual performances using StyleGAN2.
For detailed documentation and demos look at the Metacreation Lab Website.
To successfully run this program, your system must have a 64-bit Linux operating system, and you must download and install the following prerequisites:
git clone git@github.com:Metacreation-Lab/autolume.git
cd
into the cloned repositorysh installer.sh
sudo apt update && sudo apt upgrade
sudo apt install ffmpeg
ffmpeg -version
To successfully run this program, your system must have a 64-bit Windows (Windows 10 & 11 compatible) operating system, and you must download and install the following prerequisites:
While installing Python, make sure to select "Add Python to PATH" as shown below:
Open Command Prompt (windows+R
> cmd
)
Run the following command: where python
C:\Users\Metacreation Lab>where python
C:\Users\Metacreation Lab\AppData\Local\Programs\Python\Python310\python.exe
C:\Users\Metacreation Lab\AppData\Local\Microsoft\WindowsApps\python.exe
Click on the sinstaller.ps1
file and open it (located in your Downloads folder).
Paste the outcome of step 2 to the systemPython
variable under #Define python path
in the script:
# Define python path
$systemPython = "C:\Users\Metacreation Lab\AppData\Local\Programs\Python\Python310\python.exe"
If this is the first time running a script in Windows, you have to enable that functionality first. Detailed instructions to do so, can be found here
Go to Downloads and type PowerShell in the address bar and hit Enter.
Run the following command inside PowerShell: .\sinstaller.ps1
Next time you want to run Autolume, simply click the AutolumeLiveColab shortcut on your Desktop.
Place your StyleGAN2 checkpoints (https://github.com/NVlabs/stylegan2-ada-pytorch) in the models
folder.
Click the AutolumeLiveColab desktop shortcut and get going!
Showcase of current state: https://www.dropbox.com/s/e8yfdtdw3x3s7h0/autolume.mp4?dl=0 \ Trial runs:\ https://www.instagram.com/p/ChsrxkCjJNu/ \ https://www.instagram.com/p/Cg5aMrqjgIM/ \ Artworks: \ https://metacreation.net/autolume-mzton/ \ https://www.metacreation.net/projects/dreamscape/ \ https://metacreation.net/autolume-acedia/
Philippe Pasquier: Director of Metacreation Lab for Creative AI, Principal Investigator \ Arshia Sobhan: Project management \ Jonas Kraasch: Project first-phase development \ Arthur Deleu: Project development \ Mahshid Jabari: DevOps \ Peter Chen: Contributions to the super-resolution module \ Manuel Bolduc: User testing \ Ahmed Abuzuraiq: Research associate
PyTorch StyleGAN2-ada: https://github.com/NVlabs/stylegan2-ada-pytorch \ Underlying GUI code: https://github.com/NVlabs/stylegan3 \ GAN compression: https://github.com/lychenyoko/content-aware-gan-compression \ GANSpace: https://github.com/harskish/ganspace
Some of the dependencies fall under the Nvidia Source Code License.
This code base is still open to changes and bugs might still appear. Please create issues and let us know so we can polish it for the final release! \ As a dedicated research team behind Autolume, we emphasize that our system is a tool for exploration and innovation without direct support.