ML Automation Solutions (MAS)

A simplified and automated orchestration workflow to perform ML end-to-end (E2E) model tests and benchmarking on Cloud VMs across different frameworks.

Getting Started

1. Follow the instruction to create a Cloud Composer environment using Terraform. This step may take about 30min to complete.
2. Identify your dags folder. See the instructions to access the bucket of your environment.
3. In the root directory of the repository, run the following command to upload tests and utilities to the dags folder you identified in the previous step (gsutil command-line tool is required).

bash scripts/upload-tests.sh gs://<your_bucket_name>/dags

4. After the automatically scheduled tests start running, integrate Looker Studio or any other dashboard with BigQuery to monitor metrics.

If you have a use case that MAS does not cover, please email ml-auto-solutions-users@googlegroups.com. We're here to help!

Contributing

Thank you for your interest in contributing to this project!

Please review the contribution guidelines, and note that all contributions must adhere to the code of conduct.

License

Apache License 2.0