How to Use Conda for Deep Learning

In this post, we will go through the steps to set up a deep learning environment on your local machine. This includes installing necessary libraries and tools to get you started with deep learning projects.

Step 1: Install Python

Conda is a popular package manager and environment management system for Python and other programming languages. You can use it to create isolated environments for your projects and manage dependencies. - website: https://www.anaconda.com/docs/getting-started/miniconda/install/linux-install - installation command:

curl -O https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh
bash ~/Miniconda3-latest-Linux-x86_64.sh
source ~/.zshrc

# after installation, you can use initialization command to set up conda for zsh:
conda init
# to disable conda auto-activation, you can run the following command:
conda config --set auto_activate_base false

To get better network performance, you can change the default conda channel to Tsinghua University mirror:

conda config --add channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free/
conda config --add channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/main/
conda config --add channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/pytorch/
conda config --add channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/pytorch/linux-64/
conda config --set show_channel_urls yes

To get better network performance when installing packages with pip, you can change the default pip source to Tsinghua University mirror by running the following command:

pip config set global.index-url https://pypi.tuna.tsinghua.edu.cn/simple

Conda Commands Table:

Command	Description
conda create -n myenv python=3.8	Create a new virtual environment named myenv with Python 3.8
conda activate myenv	Activate the virtual environment myenv
conda install numpy pandas	Install NumPy and Pandas in the active environment
conda list	List all installed packages in the active environment
conda deactivate	Deactivate the current virtual environment
conda env remove -n myenv	Remove the virtual environment named myenv
conda env export > environment.yml	Export the current environment to a YAML file
conda env create -f environment.yml	Create a new environment from a YAML file

Step 2: Create a Virtual Environment

It is recommended to create a new virtual environment for your deep learning projects to avoid conflicts between different libraries. You can create a new environment using the following command:

conda create -n dplrn python=3.12

This command creates a new environment named dplrn with Python 3.12 installed.

Step 3: Activate the Virtual Environment

To start using the virtual environment you just created, you need to activate it with the following command:

conda activate dplrn

Step 4: Install Deep Learning Libraries

Once you have activated your virtual environment, you can install the necessary libraries for deep learning.

Pytorch: Please follow the instructions on the official PyTorch website to install PyTorch with CUDA support if you have an NVIDIA GPU.