Comment on page
Anaconda
Anaconda provides Python, R and other packages for scientific computing including data sciences, machine learning, etc.
There are several anaconda modules available on Oscar. To list all anaconda modules, run
module avail anaconda. The anaconda/2022.5 module is recommended.$ module avail anaconda
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ name: anaconda*/* ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
anaconda/2-4.3.0 anaconda/2020.02 anaconda/3-4.3.0
anaconda/2-5.3.0 anaconda/2022.05 anaconda/3-5.2.0
Do not activate a conda environment before submitting a batch job if the batch job activates a conda environment. Otherwise, the batch job will not be able to activate the conda environment and hence fail.
This is the newest anaconda module on Oscar. The first time you load the anaconda/2022.05 module, you need initialize the environment by running the following command:
$conda init bash
no change /gpfs/runtime/opt/anaconda/2022.05/condabin/conda
no change /gpfs/runtime/opt/anaconda/2022.05/bin/conda
no change /gpfs/runtime/opt/anaconda/2022.05/bin/conda-env
no change /gpfs/runtime/opt/anaconda/2022.05/bin/activate
no change /gpfs/runtime/opt/anaconda/2022.05/bin/deactivate
no change /gpfs/runtime/opt/anaconda/2022.05/etc/profile.d/conda.sh
no change /gpfs/runtime/opt/anaconda/2022.05/etc/fish/conf.d/conda.fish
no change /gpfs/runtime/opt/anaconda/2022.05/shell/condabin/Conda.psm1
no change /gpfs/runtime/opt/anaconda/2022.05/shell/condabin/conda-hook.ps1
no change /gpfs/runtime/opt/anaconda/2022.05/lib/python3.7/site-packages/xontrib/conda.xsh
no change /gpfs/runtime/opt/anaconda/2022.05/etc/profile.d/conda.csh
modified /users/yliu385/.bashrc
==> For changes to take effect, close and re-open your current shell. <==
Do not load the module in your .modules or .bashrc file. Otherwise, your VNC session cannot start.
If you load the module in your .modules or .bashrc file, you may need to have the following lines in your .bashrc as well to be able to start a VNC session:
unset SESSION_MANAGER
unset DBUS_SESSION_BUS_ADDRESS
Anaconda uses conda to install packages and manage their depenencies. You can use conda to manage conda environments. To access conda, you need to load an Anaconda module. For example,
module load anaconda/2022.05
or
module load anaconda/3-5.2.0
A user may install all needed software packages for a project in a conda environment. A conda environment can be
- shared among all users if the environment is installed in a shared directory
- private to one user if the environment is installed in a user's private directory
The command 'conda info' shows important configurations for conda environment.
$ conda info
active environment : None
user config file : /users/yliu385/.condarc
populated config files :
conda version : 4.5.4
conda-build version : 3.10.5
python version : 3.6.5.final.0
base environment : /gpfs/runtime/opt/anaconda/3-5.2.0 (read only)
channel URLs : https://repo.anaconda.com/pkgs/main/linux-64
https://repo.anaconda.com/pkgs/main/noarch
https://repo.anaconda.com/pkgs/free/linux-64
https://repo.anaconda.com/pkgs/free/noarch
https://repo.anaconda.com/pkgs/r/linux-64
https://repo.anaconda.com/pkgs/r/noarch
https://repo.anaconda.com/pkgs/pro/linux-64
https://repo.anaconda.com/pkgs/pro/noarch
package cache : /gpfs/runtime/opt/anaconda/3-5.2.0/pkgs
/users/yliu385/.conda/pkgs
envs directories : /users/yliu385/anaconda
/users/yliu385/.conda/envs
/gpfs/runtime/opt/anaconda/3-5.2.0/envs
platform : linux-64
user-agent : conda/4.5.4 requests/2.18.4 CPython/3.6.5 Linux/3.10.0-957.5.1.el7.x86_64 rhel/7.3 glibc/2.17
UID:GID : 20175:601
netrc file : None
offline mode : False
Below are some important configurations:
- envs directories: a list of directories where a conda environment is installed by default. In the output of 'conda info' above, the first default directory to install a conda environment is a $HOME/anaconda.
- package cache: a list of directories where downloaded packages are stored.
To create a new conda environment in a default directory, run the following command:
conda create -n conda_environment_name
To create a new conda environment in a different directory, run the following command:
conda create -p /path/to/install/conda_environment_name
After creating a conda environment, users can activate a conda environment to install or access packages in the environment. The command is slightly different for different anaconda modules.
For the anaoconda/2022.05 module, users can activate an environment with the following command:
conda activate conda_environment_name
For the anaconda/3-5.2.0 module, users can activate an environment with the following command:
source activate conda_environment_name
The commands above will only work if:
- A conda environment with the specified name (
conda_environment_namein the example) exists - The appropriate anaconda module has been loaded (if you are unsure about this one, consult this documentation)
If you need to activate a conda environment in a bash script, you need to source the conda.sh as shown in the following example bash script:
#!/usr/bin/bash
module load anaconda/2022.05
source /gpfs/runtime/opt/anaconda/2022.05/etc/profile.d/conda.sh
conda activate my_env
If you are using anaconda/3-5.2.0, replace the instances of
2022.05 with 3-5.2.0.After installing packages in an active environment (instructions below), you do not need to load or install those packages in the bash script; any packages installed in the conda environment (before the script even starts) will be available through the environment after it is activated (line 4 in the code above).
To deactivate a conda environment, simply use the following command:
conda deactivate
To install a package, we need to first activate a conda environment, and then run
conda install package_name=version
The "=version" is optional. By default, conda install a package from the anaconda channel. To install a package from a different channel, run
conda install with the -c option. For example, to install a package from the conda_forge channel, runconda install -c conda_forge package_name
To delete a conda environment, run
conda remove -n conda_environment_name --all
Conda may download lots of additional packages when installing a package. A user may use up all quota due to these downloaded packages. To remove the downloaded packges, run
conda clean --all
Last modified 7mo ago