Using Miniconda

Reason: Anaconda includes most packages, especially for artificial intelligence, but takes up too much space. Miniconda is more lightweight - install only the packages you need. Following the principle of “use only what you need,” I chose to use Miniconda.

Installation

(Skipped…)

Creating Virtual Environments

conda create -n python_3.9 python=3.9

conda create -n env_name python=version

Managing Environments

# Switch to the newly created python_3.9 environment
conda activate python_3.9

# Delete environment (think twice before doing this)
conda remove -n python_3.9 --all

# Exit current environment
conda deactivate

# Return to default environment
conda activate base

Installing Packages

conda install package_name

Using Requirements Files

pip can use requirements files for quick installation and can also generate them. pip usage specifications:

pip freeze > requirements.txt
pip install -r requirements.txt

conda can install from this file:

conda install --yes --file requirements.txt

Exporting and Importing Environments

conda can completely export environments to .yml files:

conda env export > freeze.yml

Then create conda environments directly from exported files:

conda env create -f freeze.yml