cd of/addons
git clone https://github.com/runwayml/ofxRunway.git
git clone https://github.com/fred-dev/ofxHTTP.git
git clone https://github.com/fred-dev/ofxIO.git
git clone https://github.com/bakercp/ofxMediaType.git
git clone https://github.com/bakercp/ofxNetworkUtils.git
git clone https://github.com/bakercp/ofxSSLManager.git
git clone https://github.com/bakercp/ofxSerial.git
git clone https://github.com/mneunomne/ofxGLFWJoystick.git
git clone https://github.com/danomatika/ofxMidi.git
OF is riddled with bugs nowadays. In particular, this is an important fix re: Poco Link to discussion
The essential steps are to go into ofxPoco/addon_config.mk file and comment out the first ADDON_INCLUDES line, then remove or rename the /libs folder in ofxPoco as well
then make clean
I also had to make changes to ofxIO: https://github.com/fred-dev/ofxIO/commit/f1e96d278ac6940291509e8f1e75506789ad9217 and https://github.com/fred-dev/ofxHTTP these are reflected in the addons git clones above
On Ubuntu, I had to
sudo usermod -a -G dialout <username>
sudo apt remove brltty
Build the docker from Dockerfile
docker build . --tag inference-engine
Run a docker container (the latest image automatically loads the python script)
docker run --expose 9000 --network="host" -v $PWD:/workspace/ --gpus all -it inference-engine
then run the python
python3 runway_model.py
then hit Connect in Runway
then run the OF app
Run a docker container, expose port 9000 and 127.0.0.1, mount current working directory to /workspace/ in the container, enable gpu, and load the tf1 image into a bash terminal
docker run --expose 9000 --network="host" -v $PWD:/workspace/ --gpus all -it my/tf1 /bin/bash
The dnnlib I was using before was incompatible with SG2-ada, so keep an eye out for that. I borrowed a lot from https://github.com/usufyan29/stylegan2_runway
https://help.runwayml.com/hc/en-us/articles/4401808774419-Adding-Models stylegan w runway implementation