Source
yaml
id: gpu-modal
namespace: company.team
tasks:
- id: python_repository
type: io.kestra.plugin.core.flow.WorkingDirectory
tasks:
- id: clone_repository
type: io.kestra.plugin.git.Clone
branch: main
url: https://github.com/kestra-io/scripts
- id: modal
type: io.kestra.plugin.modal.cli.ModalCLI
commands:
- modal run modal/gpu.py
docker:
image: ghcr.io/kestra-io/modal:latest
env:
MODAL_TOKEN_ID: "{{ secret('MODAL_TOKEN_ID') }}"
MODAL_TOKEN_SECRET: "{{ secret('MODAL_TOKEN_SECRET') }}"
About this blueprint
CLI Python
This flow clones a repository, pulls a public container image with the required dependencies and runs a Python script in a Docker container.
The Python script uses Modal Python Client to run the script on a GPU-enabled server.
- To learn more about GPU support on modal.com, check the Modal documentation
- The Python script is available in the repository: kestra-io/scripts
This flow assumes that you have a Modal account and that you stored the Modal credentials as secrets.
More Related Blueprints