Skip to main content

Serverless Deep Learning On Private Git Repositories

· 2 min read

You can now run trainML serverless deep learning training jobs on model code hosted in private git repositories that support SSH authentication.

How It Works

In order to provide the trainML platform secure access to your private git repository, you can now create an SSH key on the Account Profile page, which is accessible through the Account dropdown menu in the toolbar. At the bottom of the Third-Party Keys section there is a new section called Git SSH Key. Click the Generate button to create a new key pair. When the generation completes, the public key portion is displayed. The private key is stored securely within the trainML platform and is not viewable.

To enable the use of this key, you must add it to a user account within your git repository that has access to the repository you want to use for the training job. To add the key to your Github account, follow these instructions.

Once the generated key has access to the code repository, you can now create a new training jobs using your private repository url as the Model Code Location on the job form. However, you must specify the repository as an SSH URL. For example, rather than using https://github.com/trainML/examples.git, you must instead specify git@github.com:trainML/examples.git. If you are using Github, you can obtain the SSH URL by following the instructions here and clicking the Use SSH link.