Skip to content

plain_install.md

Latest commit

 

History

History
23 lines (18 loc) · 1.66 KB

plain_install.md

File metadata and controls

23 lines (18 loc) · 1.66 KB

Running the code without Docker

While the code itself should be platform independent, we recommend running it on GNU/Linux. The code has been tested on Ubuntu 22.04 LTS only.

1. Install System Dependencies

  1. If you want to use the GPU, run nvidia-smi to check if you have a GPU driver installed. If not, install the latest version.
  2. Install Python and pip if necessary. On Ubuntu/Debian run sudo apt install python3 python3-pip.
  3. Make sure that you have pkg-config and FreeType installed. Otherwise, the installation of matplotlib might fail. On Ubuntu/Debian run sudo apt install pkg-config libfreetype6-dev.
  4. Install the g++ compiler for improved performance. On Ubuntu run sudo apt install g++.
  5. Install the oneAPI Math Kernel Library. On Ubuntu run sudo apt install intel-mkl.
  6. Install R. On Ubuntu run sudo apt install r-base.

2. Install Python Dependencies

We recommend using a virtual environment for package management to avoid version conflicts.

  1. Install PyTorch. If you want to use the GPU, choose the latest (highest) CUDA version. Otherwise, select CPU.
  2. Run pip install -r requirements.txt to install the remaining dependencies.
  3. Verify your install.

3 Install R Dependencies

  1. Run Rscript -e 'install.packages(c("lubridate", "dplyr", "tidyr", "ggplot2"))'