For dependency management we use conda. Go ahead and install it if you don't have it already.
- Clone the repository
git clone https://github.com/The-Jacob-Lopez/GazeMouse.git
- Create a new conda environment and install the required dependencies
Note that if you intended on using GPU, you will also have to install the
conda env create -f environment.yml
pytorch-cuda
package. This can be done by create the environment using the appropriate environment fileconda env create -f environment-cuda.yml
- Python may struggle to find the appropriate modules created in this repo. Fix this by entering the repo through a shell and running the following command
conda develop ./GazeMouse
You can run the app by entering the repo through a shell and running the following commands
conda activate GazeMouse
python ./GazeMouse/src/app/dev_app.py
Sometimes creating a large environment from an environment.yml
file can take ludicoursly long. It is recommended to use the libmamba solver to speed up this process. Instructions on how to do so can be found (here)[https://www.anaconda.com/blog/conda-is-fast-now].
There is some interesting and unexpected behavior with how pip interacts with MacOS versions which is documented (here)[https://stackoverflow.com/questions/65290242/pythons-platform-mac-ver-reports-incorrect-macos-version/65402241#65402241]. If you run into issues downloading the correct version of opencv-contrib-python
then prepend SYSTEM_VERSION_COMPAT=0
to the previously mentioned conda commans as such
SYSTEM_VERSION_COMPAT=0 conda env create -f environment.yml