News

VS Code Improves ML Model Training with Python

Using Python in Visual Studio Code for machine learning model training and experimentation is easier in the February 2021 update to the tool that fosters Python programming in Microsoft's popular, open source, cross-platform code editor.

That ease comes with new integration with TensorBoard, the visualization toolkit for TensorFlow, a leading open source ML platform used for developing and training ML models created by Google. Its primary use is for visualizing model graphs, metrics and other data patterns.

Working with TensorBoard in VS Code
[Click on image for larger view.] Working with TensorBoard in VS Code (source: Microsoft).

"TensorBoard is a data science companion dashboard that helps PyTorch and TensorFlow developers visualize their dataset and model training," said Jeffrey Lew of the VS Code Python team in announcing the February update. " With TensorBoard directly integrated in VS Code, you can spot check your models' predictions, view the architecture of your model, analyze you model's loss and accuracy over time, profile your code to find out where it's the slowest, and much more!"

The new update also improves the Pylance extension (specifically, docstring readability) that serves as the language server for VS Code, leveraging the Language Server Protocol to provide Python-specific "smarts" in the editor such as autocomplete and smart completions (IntelliSense), error-checking (diagnostics), jump-to-definition, linting and corrections, find all references and so on.

Several other fixes and new features, such as tweaks to improve code navigation by streamlining go to definition and go to declaration behavior.

By far the most popular extension in the Visual Studio Code Marketplace, the Python tool has been downloaded more than 31.5 million times.

About the Author

David Ramel is an editor and writer at Converge 360.

comments powered by Disqus

Featured

  • GitHub Previews Agentic AI in VS Code Copilot

    GitHub announced a raft of improvements to its Copilot AI in the Visual Studio Code editor, including a new "agent mode" in preview that lets developers use the AI technology to write code faster and more accurately.

  • Copilot Engineering in the Cloud with Azure and GitHub

    Who better to lead a full-day deep dive into this tech than two experts from GitHub, which introduced the original "AI pair programmer" and spawned the ubiquitous Copilot moniker?

  • Uno Platform Wants Microsoft to Improve .NET WebAssembly in Two Ways

    Uno Platform, a third-party dev tooling specialist that caters to .NET developers, published a report on the state of WebAssembly, addressing some shortcomings in the .NET implementation it would like to see Microsoft address.

  • Random Neighborhoods Regression Using C#

    Dr. James McCaffrey from Microsoft Research presents a complete end-to-end demonstration of the random neighborhoods regression technique, where the goal is to predict a single numeric value. Compared to other ML regression techniques, advantages are that it can handle both large and small datasets, and the results are highly interpretable.

  • As Some Orgs Restrict DeepSeek AI Usage, Microsoft Offers Models and Dev Guidance

    While some organizations are restricting employee usage of the new open source DeepSeek AI from a Chinese company due to data collection concerns, Microsoft has taken a different approach.

Subscribe on YouTube

Upcoming Training Events