Star 1
Version 1.0.1

Maintainers
support[at]europeanweather.cloud
License
Support level
EWC
Category
Machine Learning & AI
GPU-accelerated
Technology
Ansible Playbook
ECMWF Anemoi
Create a conda environment with the Anemoi framework for weather forecasting based on machine learning and associated software stack.
Requirements
This ansible role depends on ewc-ansible-role-conda and needs to be applied on a GPU-powered instance if training or inference are used.
Role Variables
anemoi_env_wipe
: Boolean to decide whether to wipe the environment if exists prior to a reinstallation. Default: noanemoi_env_name
: Name of the environment containing the software stack. Default: aifs-single-mseanemoi_env_path
: Installation path for the environment. Default: "{{ conda_prefix }}/envs/{{ anemoi_env_name }}"anemoi_create_ipykernel
: Create the jupyter kernel for this environment. Default: yesconda_prefix
: Prefix where conda is installed. Default:/opt/conda
conda_user
: User owning the conda installation. Default:root
Example Playbook
- hosts: all
roles:
- ewc-ansible-role-anemoi
License
Apache 2.0.
Author Information
ECMWF for the European Weather Cloud

Other
Deployable
EWCCLI-compatible