Quick Start
This guide serves as the fastest way to install IGM. It assumes that have already
- Downloaded the nvidia drivers
- Have a working virtual environment
If this is the case, you can skip to the next section.
Installing on Windows
Tensorflow does not allow us to run IGM on GPU directly on Windows, and the module oggm_shop
does not work on windows. Therefore, we recommend windows-user to install WSL2-ubuntu, which provides a linux/ubuntu terminal. WSL2 terminal can be nicely linked with VS code (with an extension). First, install WSL2-ubuntu
Installing Methods
Once this is done, the options are the following
- pip
- github
Pip
To install the latest version of IGM, simply run
For reproducibility purposes, one might want to install a specific version of IGM. In order to do this, one can specify the version (note, this version must exist on the PyPI servers).
Github
If one wants to have the latest versions, or even, work on a specific hash for reproducibility, one can download IGM through the github repository. This is useful for developers, and researchers alike, who want to have the latest features as well as contribute to IGM's model personally.
One can download the latest version of IGM with git clone
A Note about IGM's Install Location
Please note that where you decide to clone IGM is purely the location where IGMs source code will be installed. After you install IGM with pip install -e
(more on this below), you can run IGM from any location on your computer. This is because installing IGM will create a symbolic link to wherever this folder is installed.
Installing on Mac
IGM's core package, Tensorflow, is unfortunately not natively supported on Mac OS for GPUs. Instead, a "Tensorflow for Mac", called tensorflow-metal, was developed as a workaround. To install IGM on Mac, you can still clone the repository with the above line, but you must additionally change tensorflow
to tensorflow-macos
in setup.py
before running pip install -e
. Here is a working procedure (tested on MacBook Pro M2); we still recommend using a virtual environment such as conda or venv when installing.
setup.py
file, you will need to edit the "install_requires=[...]" line depending on your requirements:
- To use only the CPU:
tensorflow-macos==2.14.0
- To use the GPU:
tensorflow-macos==2.14.0, tensorflow-metal,
Now, once the setup.py
file is ready for your machine and operating system, one can install IGM inside his or her virtual environment. To do this, run the following command in the same level as the setup.py
file:
Note that while the user installs
As IGM is being updated often, make sure you have the latest version by running the following command inside the igm
folder