Скачать или смотреть Introduction to Anaconda environments

An introduction to Anaconda environments, installing packages from conda, conda-forge and pypi (pip), and saving/loading environments.

Create empty environment (called temp_env here):
conda create -n temp_env

Create environment with given Python version (just use python alone for latest available version):
conda create -n temp_env python=3.7

Activate environment (before launching Spyder or Jupyter-Lab):
conda activate temp_env

Deactivate environment:
conda deactivate

List packages in environment
conda list

List environments:
conda env list

After activation environment install packages with conda, e.g.
conda install jupyterlab spyder numpy pandas matplotlib scikit-learn

Or install from Conda Forge:
conda install -c conda-forge geoplot

Or install from PyPi:
pip install forecast-tools

Remove environment (deactivate first):
conda env remove -n temp_env

Export environment to yml file (activate first). This exports all packages with version numbers (use actual greater sign, rather than [greater than] for export - YouTube description won't allow me to use it!):
conda env export [greater than] environment.yml

Set up environment from yml file:
conda env create -f environment.yml

An example simple yml file (we can put version numbers on some or all packages if we like with =, or with == for pip packages, but without version numbers Anaconda will work out best versions to use, though this may take some time):

name: scipy_env
channels:
defaults
conda-forge
dependencies:
jupyterlab
matplotlib
numpy
pip
pandas
scikit-learn
scipy
spyder
pip:
simpy