Netron
Desktop and browser viewer that draws the structure of trained neural network model files.
- Stars
- 33.6k
- Forks
- 3.2k
- Licence
- MIT
- Language
- JavaScript
- Last update
- 4 days ago
- Type
- App
What it is
Netron opens a trained machine learning model file and shows its layers and connections as an interactive diagram. It supports many formats, including ONNX, TensorFlow Lite, PyTorch, TensorFlow, Core ML and Keras. People use it to inspect a model's architecture without writing code.
- Open a sample ONNX or TensorFlow Lite model from the README in the browser at netron.app
- Open a local model file with the desktop app or with the Python command netron [FILE]
- Inspect the layers of a Keras .h5 or TensorFlow .pb file
- Check the structure of a Core ML .mlmodel file
Great for
- Engineers who need to check a model's layers and tensor connections before deploying it
- Students learning how a network is structured by looking at a real file
- Developers converting a model between formats who want to confirm the result
- People who want to view a model quickly in the browser without installing anything
Not ideal for
- People who need to train or edit models, since the README describes Netron as a viewer
- Users whose model format is not listed, because support is limited to the formats named in the README (some experimental)
- Those expecting a full model debugger or performance profiler; the README does not describe these features
How to get it
- 1
Use it in the browser (or self-host)
Install Python package
pip install netronOpen a model file from Python
netron [FILE] - 2
Run it
- Go to https://netron.app in a browser, or install the desktop app for your system
- Open a model file from your computer, or use one of the sample links in the README
- Click nodes in the diagram to view their attributes and details
- 3
Get the most out of it
- Use the Python package with netron.start('[FILE]') to open a model from a script
- Install the macOS version with brew install --cask netron or the Windows version with winget install -s winget netron
- Try the sample models listed in the README to see how different formats appear
Downloads
Strengths
- Supports a wide range of model formats in one tool
- Available as a browser version, desktop app, and Python package
- Released under the permissive MIT license
Trade-offs
- Some formats are only experimental support, per the README
- The README does not describe editing or running models inside the viewer
- Very large models may be harder to read; the README does not address performance limits
How it compares
| Project | Stars | Type | Level | Updated | Best for |
|---|---|---|---|---|---|
| netron (this) | 33.6k | App | beginner | 4 days ago | Engineers who need to check a model's layers and tensor connections before deploying it |
| onnxruntime | 22.2k | Library | intermediate | 4 days ago | Developers who want to deploy a trained model inside an application or service |
| wandb | 11.3k | Library | beginner | 4 days ago | Machine learning engineers who want to compare training runs over time |
| mit-deep-learning | 10.5k | Learning resource | beginner | 2.8 years ago | Students following MIT deep learning courses |
| nni | 14.4k | Library | intermediate | 2.3 years ago | Researchers comparing hyperparameter tuning algorithms such as TPE, Hyperband, or Bayesian methods |
| ML-For-Beginners | 91.9k | Learning resource | beginner | 5 days ago | Beginners who want a structured, self-paced introduction to machine learning |
Stats synced from GitHub. The assessment above was drafted with AI from the project's README on 2026-10-29 and may miss recent changes — check the README before relying on a command. MIT licence: free to use, modify and distribute, including commercially, as long as you keep the copyright notice.