Skip to content

Latest commit

 

History

History
60 lines (39 loc) · 2.15 KB

README.md

File metadata and controls

60 lines (39 loc) · 2.15 KB
Raw

Run Jupyter notebooks in background

This extension supports running and monitoring notebooks in the background from the comfort of Jupyter Labs.

The use cases are -

  • you want to continue working on an approach and try several other approaches at the same time in background
  • you are sharing expensive resources like GPU with others in your team or univ

Requirements

  • Jupyter Labs (>2.0.0)
  • hyperML (>0.9.0)

Install

jupyter labextension install hyperml-submit-notebooks

Get Started

Setup the following OS environment variables:

  • hyperML:
    • HYPERML_SERVER_ENDPOINT
    • HYPERML_API_KEY
  • AWS S3: The source Notebook and the processed notebook will be stored on S3 or Minio
    • HYPERML_S3_ACCESS_KEY
    • HYPERML_S3_SECRET_KEY
    • HYPERML_S3_BUCKET (defaults to hyperML)
    • HYPERML_S3_URL (e.g. s3-us-west-2.amazonaws.com)
    • HYPERML_S3_SECURE (default true)

Scheduling notebooks

  1. Locate Run in Background button in notebook toolbar

check screens/run-in-background.png

  1. Enter Resource Plan (must be setup on hyperML) and Container Image

check screens/choose-params.png

  1. Click OK
  2. Continue working on the current notebook or monitor the background request on background-notebooks tab.

check screens/background-notebooks.png

  1. Download and Open the processed notebook

Note: You can also open background notebooks from command search 'background-notebooks:open'.

Limitations

Only Dark mode styling as of now

Issue Reporting

Welcome any issues. We are a small team to expect a reponse time of 2-3 days

About hyperML

hyperML is a radically simplifies on-cloud machine learning for teams/developers. Scale your training jobs right from jupyter labs session or launch a notebook with click of button. Read more details at https://www.hyperml.com