Below you will walk through how to create a project. You can check out ready-to-run example projects on our GitHub Repo.
Engine ML uses Git to keep track of your model code and experiment configuration. So, the SSH key that you use with Git must be registered with Engine ML.
You can do this by issuing
engine user add-key /path-to-public-ssh-key.
engine user add-key ~/.ssh/id_rsa.pub Successfully added key: ~/.ssh/id_rsa.pub
If you do not have an SSH public key, then you must generate one.
SSH keys uniquely identify you to Engine ML.
Do not ever share your private key!
List All Projects
Engine ML projects are for grouping experiments (jobs).
Any user can create a project, but it is best to keep the number of projects low.
Before creating a new project, first see what projects already exist.
engine project list my-project
Create a Project or Join an Existing One
Let's create a new project. To do this, first make sure that your current
working directory is within a Git repository. Issue
git status to verify.
engine project create PROJECT_NAME to
connect to an existing project or create a new one.
This will add an appropriately named Git remote.
engine project create my-project Found git repository at /my-project-local Creating git repository for project my-project... Added git remote `email@example.com:my-project.git`
You can see what Git remotes are configured for your Git repository using
git remote -v:
git remote -v engine firstname.lastname@example.org:my-project.git (fetch) engine email@example.com:my-project.git (push)