#124 · Primary category: MLOps & Evaluation

tensorwatch

ai data-science debug debugging debugging-tool deep-learning deeplearning explainable-ai explainable-ml jupyter jupyter-notebook machine-learning machinelearning model-visualization monitoring python reinforcement-learning saliency

Debugging, monitoring and visualization for Python Machine Learning and Data Science

Project last updated:03/30/26

GitHub Stars

3.5K

Forks

361

Contributors

15

License

MIT

Why we included this project

TensorWatch from Microsoft Research gives you a live view of your training run inside Jupyter Notebook. It renders real-time graphs of your metrics as they are written, so you can watch a loss curve move instead of waiting for the run to finish. Its Lazy Logging mode lets you submit arbitrary Python expressions to the live training process and route the resulting stream to any visualizer, which is handy for probing things you didn't think to log ahead of time. You can also build custom UIs and dashboards to inspect model internals like saliency maps or reinforcement learning behavior. Keep in mind it is a development and debugging tool, not something for production or multi-tenant environments.

Articles for this project

No articles for this project yet.

To suggest a topic or contribute an article, contact us.

Related projects in this category