netron
Visualizer for neural network, deep learning and machine learning models
What is netron?
Visualizer for neural network, deep learning and machine learning models
Overview of netron
What it is
and what it does, in plain terms: Netron is a viewer for neural network, deep learning and machine learning models. It is available as a browser version, as desktop applications for macOS, Linux and Windows, and as a Python package. The source code lives in the lutzroeder/netron repository, and the browser version is hosted at netron.app. The project describes itself as a visualizer for neural network, deep learning and machine learning models and lists AI among its categories.
Key features
- Supports ONNX, TensorFlow Lite, PyTorch, torch.export, ExecuTorch, TorchScript, TensorFlow, Core ML, OpenVINO, Keras, Caffe, Darknet, Safetensors and NumPy.
- Provides experimental support for MLIR, JAX, GGUF, RKNN, ncnn, MNN, PaddlePaddle and scikit-learn.
- Runs in the browser, so a model can be opened without installing a desktop application.
- Ships desktop builds for macOS (
.dmgor the Homebrew cask), Linux (.debor.rpm) and Windows (.exeinstaller or winget). - Offers a Python package that launches the viewer from the command line with
netron [FILE]or from code withnetron.start('[FILE]'). - The README lists sample model files for ONNX, TorchScript, TensorFlow Lite, TensorFlow, Keras, MLIR, Core ML and Darknet that can be downloaded or opened in the browser version.
Use cases
- Opening a model file to view its structure in one of the supported formats.
- Inspecting models in the browser without installing anything locally.
- Launching the viewer for a specific file from a Python environment.
- Trying the viewer on the sample models listed in the README, such as squeezenet (ONNX), yamnet (TensorFlow Lite), mobilenet (Keras) or yolo (Darknet).
Getting started
pip install netron netron [FILE] On macOS the README also lists brew install --cask netron; on Windows it lists winget install -s winget netron. Desktop downloads are provided as .dmg (macOS), .deb or .rpm (Linux) and .exe (Windows).
License and pricing
Netron is released under the MIT license and is listed as open source. The README does not state any pricing.
Based on the project's public README and repository metadata.