Home

Awesome

An example of setting up remote cloud environment for training a machine learning model using Terraform Provider Iterative.

See the blog post at https://dvc.org/blog/local-experiments-to-cloud-with-tpi for a full explanation.

The main.tf file contains two different tasks:

  1. basic scenario for running a script remotely, and
  2. training a model on a GPU device

To run this tutorial, make sure to have a cloud account (AWS, Azure, GCP, or K8s) with authentication credentials stored as environment variables.

  1. Install Terraform
  2. terraform init: setup dependencies
  3. terraform apply: provision cloud infrastructure & upload task
  4. terraform refresh && terraform show: check status
  5. terraform destroy: download results and terminate cloud infrastructure