custom-environments

Creating Custom Conda Environments

This directory contains a number of example Conda environment files that you can use as a starting point to create new Conda environments for different purposes. For more information about Conda and Conda environments, please see the Conda Documentation.

This directory contains the following example Conda environments:

• Geospatial:: An environment that contains basic packages for geospatial data analysis.
• SciPy: An environment with foundational packages for data science and scientific computing.
• R: An environment with an R kernel for Jupyter and basic R packages for data science.
• Fastai: An custom environment for fast.ai with its dependencies and sample datasets.

All of these environments include a working Jupyter kernel for Python or R, which enables the packages to work in a Jupyter notebook.

Using these examples

Once you have cloned this repository, you can create custom Conda environments in SageMaker Studio Lab in one of two ways:

First, you can open a Terminal, cd to the environment directory of your choice under this directory and use conda create:

$ cd studio-lab-examples/SciPy
$ conda env create -f scipy.yml

Alternatively, you can right-click on the .yml file from the specific environment directory in the JupyterLab file browser within SageMaker Studio Lab select the "Build Conda Environment" menu item.

Once you have created a new Conda environment in SageMaker Studio Lab, you will be able to create a new notebook with that environment in the JupyterLab launcher. It takes about 1 minute for the application to detect the new environment and kernel.