Overview
Netron makes exported model structure inspectable. It helps verify input names, dimensions, and operators, but opening a model does not prove runtime compatibility or numerical correctness. Some formats have experimental support.
Key features
- Graph viewer
- Tensor metadata
- Multiple formats
- Desktop/browser
- Python launch
Requirements, installation and quick start
Usage
How it works
Format parsers expose graph and tensor metadata in an interactive viewer. Weight-only containers may expose less structure than full graphs.
Audience and requirements
Training and deployment engineers. Supported model file and a viewer runtime.
Practical use cases
Export inspection; interface handoff; model education.
Limitations and selection
Does not execute benchmarks or validate numerical equivalence. Format support varies.
Related projects and selection
google-ai-edge/mediapipe：Complement: inspect edge model interfaces before task-level runtime testing.
pytorch/pytorch：Pipeline: train/export with PyTorch and inspect the exported artifact with Netron.
Source review
Editorial analysis of upstream sources, without runtime or benchmark testing. Proposed workflows are editorial suggestions.
Model compatibility and use cases
Views formats including ONNX, TensorFlow Lite, and PyTorch; some formats such as GGUF are experimental.
License and risk notes
The repository page identifies MIT. Read LICENSE; model weights and datasets may have separate terms.
Editorial source review 2026-09-09T05:00:00.950Z. README and live repository page verified; current stars/forks from GitHub HTML. Last-push metadata retained from 2026-09-05 discovery snapshot. No runtime benchmark. Integration proposals are editorial analysis.
Release and maintenance
Reviewed 2026-09-09. Counters come from repository pages; features are based on upstream documentation. See Releases in the source links. Editorial analysis of upstream sources, without runtime or benchmark testing. Proposed workflows are editorial suggestions.