The Sagemaker JupyterLab documentation shows in the section Clone conda from an existing environment that you should be able to clone the base environment into a new conda environment:
# create a fresh env from a base environment
conda create --name py310-base-ext --clone base # replace 'base' with another env
But when I try that I get the errors:
The following packages cannot be cloned out of the root environment:
- conda-forge/linux-64::conda-23.11.0-py310hff52083_1
- conda-forge/noarch::conda-libmamba-solver-24.1.0-pyhd8ed1ab_0
Packages: 653
Files: 99
Downloading and Extracting Packages:
PathNotFoundError: /tmp/local_conda_forge_artifacts/linux-64/sagemaker-code-editor-1.83.1-ha1e3dc1_3.conda
PathNotFoundError: /tmp/local_conda_forge_artifacts/linux-64/sagemaker-code-editor-1.83.1-ha1e3dc1_3.conda
The default base environment has 99% of what I need in terms of dependencies, and much of the documentation implies you should not use the base environment directly. But I have not yet found a way to utilize all the packages already installed and still have a new conda / python virtual environment.
Should I just do everything in the base environment and forget about creating a new conda environment? Or is there another way to do this?