Skip to content

Latest commit

 

History

History
36 lines (18 loc) · 1.81 KB

README.md

File metadata and controls

36 lines (18 loc) · 1.81 KB
Raw

condensed-binocular

A simple script to report ML metrics to Azure ML and Azure AppInsights at the same time!

🌦️ You might recognize these small annoyances:

  • There is no out-of-the-box way to report metrics to AppInsights and Azure ML simultaneously == more lines of code than necessary.

  • AML currently does not allow you to report metrics to both parent- and childrun simultaneously == more lines of code than necesary.

Therefore, this simple class aims to reduce lines of code for AML metric reporting, so reporting in Azure becomes a tiny bit more fun! Also, it follows some engineering best practices, like tests, environment variables and constants - you know, just as a reminder of what's nice.

⛰️ It's based on the following:

🍭 It's actually very simple:

  1. Make sure you have an AML Workspace

  2. Make sure you have an AppInsights resource

  3. Add the condensed_binocular.py script in your project. Don't forget to add the .env file and the constants.py file if feel like using them this way, and change accordingly.

  4. Integrate in your code. See binocular_sample.py on how to use it.

  5. Wait until your experiment ran, and check if the values show up in both the AML Experiment as in AppInsights.

  6. PROFIT.

💚 Contribute

Not working as expected or nice little additions necessary? PRs are welcome! For major changes, please open an issue first.