GitHub stars
3.5K
PyPI downloads / week
155
Commits, last 12 months
6
Contributors
13
Last commit

Overview

TensorWatch is an open-source ML visualization tool released under the MIT license. Its GitHub repository has 3,474 stars, 361 forks, and 13 contributors. It is maintained with 6 commits in the last 12 months; the most recent commit was on Mar 30, 2026. On PyPI it is downloaded about 155 times per week.

Its GitHub description: Debugging, monitoring and visualization for Python Machine Learning and Data Science

GitHub topics: aideep-learningdeeplearningmachine-learningmachinelearningreinforcement-learningdebuggingdebugging-tooldebugmonitoringexplainable-aiexplainable-ml

Install

PyPI
pip install tensorwatch
v0.9.1

Development activity

Commits per month to the default branch over the last 12 complete months.

0246 Oct 2025 Nov Dec Jan 2026 Feb Mar Apr May Jun Jul Aug Sep
0246 Oct 2025 Jan 2026 Apr Jul
6 commits from Oct 2025 to Sep 2026.
Show data as a table
TensorWatch commits per month
MonthCommits
Oct 20250
Nov 20250
Dec 20250
Jan 20260
Feb 20260
Mar 20266
Apr 20260
May 20260
Jun 20260
Jul 20260
Aug 20260
Sep 20260

Popularity among Machine learning visualization tools

GitHub stars for TensorWatch and the most-starred alternatives.
Show data as a table
GitHub stars
ToolStars
Netron33,544
PlotNeuralNet25,009
Opik22,373
Phoenix11,704
FiftyOne11,143
Visdom10,318
TensorWatch3,474

Alternatives to TensorWatch

  • Netron

    Viewer for neural network, deep learning and machine learning models.

    33.5K starsJavaScriptMIT

  • PlotNeuralNet

    LaTeX code for drawing neural network architecture diagrams.

    25K starsTeXMIT

  • Opik

    Formerly CometLLM. Debug, evaluate, and monitor LLM applications with tracing and dashboards.

    22.4K starsPythonApache-2.0

  • Phoenix

    ML observability in a notebook with UMAP visualizations

    11.7K starsPythonElastic-2.0

  • FiftyOne

    Tool for visualizing, curating and evaluating computer vision datasets and models.

    11.1K starsTypeScriptApache-2.0

  • Visdom

    Tool for real-time visualization and monitoring of live data such as ML experiments.

    10.3K starsPythonApache-2.0

TensorWatch FAQ

Is TensorWatch open source?

Yes. TensorWatch is released under the MIT license, and its source code is on GitHub.

Is TensorWatch still maintained?

Mostly. TensorWatch had 6 commits in the last 12 months; the latest was on Mar 30, 2026.

How do I install TensorWatch?

From PyPI: pip install tensorwatch.

What are the best alternatives to TensorWatch?

The most popular alternatives listed in Machine learning visualization tools are Netron, PlotNeuralNet, and Opik. See the full ranking.

What language is TensorWatch written in?

TensorWatch is primarily written in Jupyter Notebook, according to GitHub.

Data: GitHub API, PyPI. Updated Oct 4, 2026.