No. 9 in the 20 October edition · Developer tools

tensorViz

Picked by the editor

Visualize PyTorch models and inspect layers in VS Code

Visit site
Made this? Claim it

We check that a product is real and live before listing it, not that it works as claimed. Do your own research before signing up or paying. What we check

Follow @dailyshipped on Instagram or Facebook for every edition.

tensorViz homepage
Homepage, as captured 20 October · tap to zoom

In brief · compiled from tensorviz.dev

What is it?
tensorViz is a VS Code extension for visualizing PyTorch models as interactive graphs. It shows nested modules, tensor shapes from captured runs, failures, and linked Python source.
Who is it for?
It is for people working with PyTorch models in VS Code, including developers reviewing model changes and coding-agent workflows.
How does it work?
Install the extension, open a model graph, expand modules, inspect shapes from a captured run, and select layers to view their Python source. It can also compare Git revisions, review pull requests, share graphs, and connect to agents through MCP.

tensorViz is a VS Code extension for exploring PyTorch models as interactive graphs. Developers can expand nested modules, inspect shapes from captured runs, find execution failures, and jump from layers to their Python source. It also supports architecture comparisons, Git and pull request reviews, graph sharing, and optional MCP connections for coding agents.

Also in this edition20 October →

More in Developer toolsAll Developer tools →