Skip to main content

CICD with Gitlab, GCP, Terraform, Ansible and Cloudrun

In this workshop, you will learn how to:

  • Use the transformers package of huggingface in a Python FastAPI project in order to load pre-entrained AI models such as GPT2.
  • Provision some infrastructure on GCP using terraform and terragrunt
  • Make some CI/CD automation with gitlab-ci that will:
    • build OCI/docker images and deliver to a Google Artifact Registry.
    • Deploy containers on Google Cloud Run.

cicd

Workshop's material

All this workshop files are found on https://gitlab.comwork.io/comwork_public/training/cicd-gpt-gcp.

Workshop [part 1]​

In this first part: package transformers from Hugging Face in Python (FastAPI), infrastructure provisioning on GCP with terragrunt, terraform and ansible.

Workshop [part 2]​

In this second part: CI/CD pipelines with docker and gitlab CI, automated deployment on GCP artifact registry and cloud run and comparison with an alternative infrastructure on comwork cloud, Scaleway with a Harbor registry on OVH.