This is a brief note on how to set up your ML/DL development environment on your own workstation for success. You do not necessarily require a powerful laptop or workstation as ultimately your goal for doing ML/DL on your workstation is to figure out what sort of experiments to run. Not fitting large models but rather to run a set of tests and figure out answers to questions such as:
- What data to use
- How to prepare data
- What models to use
- What configuration to consider and use
Once these preliminary questions have been answered, you can then move on to running larger experiments on servers and clusters. Just not on your workstations. This may mean that you use a smaller data samples or scale down your model runs. Work with what you have.
We support Windows 10 running Windows Subsystem for Linux (WSL), macOS and Linux as the operating systems of choice in the curriculum.
If you are running anything else, please consider running a virtualisation software like VirtualBox or Hyper-V and install a Linux distribution such as Ubuntu as a VM and use that as your development environment.
At the minimum, please install the following:
- Anaconda Distribution. This is a popular distribution of Python with the necessary scientific and ML/DL packages.
- A decent text editor. If you do not have any preference yet, Visual Studio Code is recommended.
- Git for version control. Please install the package provided by your operating system.
- Make for automation. Please install the package provided by your operating system.
In later weeks, we will cover things such as containers and orchestration. For the most part, most of the tasks will be run in our cluster environment but you are also recommended to install Docker for your operating system. Please refer to the install instructions and install it on your workstation.