WANNA-ML is a CLI tool that helps researchers, data scientists, and ML Engineers quickly adapt to Google Cloud Platform (GCP) and get started on the cloud in almost no time.
It makes it easy to start a Jupyter notebook, run training jobs and pipelines, build a Docker container, export logs to Tensorboards, and much more.
We build on top of Vertex-AI managed services and integrate with other GCP services like Cloud Build and Artifact Registry to provide you with a standardized structure for managing ML assets on GCP.
See the documentation for more details.
pip install -U wanna-ml.
For more information on the installation process and requirements, visit out installation page in documentation
WANNA-ML relies on
gcloud for user authentication.
- Install the
gcloudCLI – follow official guide
- Authenticate with the
- Set you Google Application Credentials
gcloud auth application-default login
You can use a local Docker daemon to build Docker images, but it is not required. You are free to choose between local building on GCP Cloud Build. If you prefer local Docker image building, install Docker Desktop.
GCP IAM Roles and Permissions
Different WANNA-ML calls require different GCP permissions to create given resources on GCP. Our documentation page
lists recommended GCP IAM roles for each
Jump to the samples to see a complete solution for various use cases.
Please report issues to GitHub.
Your contributions are always welcome, see CONTRIBUTING.md for more information. If you like WANNA-ML, don’t forget to give our project a star!
Distributed under the MIT License – see LICENSE.