lutzroeder/netron

This system is a desktop and browser-based application for visualizing and inspecting machine learning models, built on Electron. It provides tools to parse and render model graphs and tensor values across a wide variety of ML frameworks, including Arm NN, JAX, and Keras. The project includes comprehensive end-to-end testing infrastructure and supports distribution as native packages for Windows, macOS, and Linux.

The Tools directory now includes build scripts and supporting code to generate JavaScript schema and metadata files for several new machine learning frameworks: Arm NN, BigDL, CatBoost, Circle, CNTK, ESP-DL, ExecuTorch, GGUF, JAX, KaNN, Keras, LiteRT-LM, MediaPipe, and MegEngine. These additions enable the library to parse and understand the model formats and operator definitions specific to these frameworks.

Tools · high confidence

The repository has been initialized with essential configuration files, including a .gitignore to exclude build artifacts and caches, an .npmrc to silence npm logs, a CITATION.cff for academic referencing, a CONTRIBUTING.md with development instructions, and a MIT LICENSE. Additionally, an eslint.config.js establishes code quality rules, while package.js and package.py provide build and packaging scripts for the application and Python distribution.

(repo-wide) · high confidence

The Netron desktop application features a refreshed icon design with updated gradients and shadows, available in both standard and macOS-specific SVG formats. The \publish/electron-builder.json\ configuration has been updated to include file associations for a wider range of model formats (including JAX, MindIR, TOSA, and LiteRT-LM) and configures Linux packages to build as .deb and .rpm, while Windows builds are now set up for Azure Trusted Signing.

publish · high confidence

Added Playwright-based end-to-end tests for the browser and desktop applications, verifying model loading, graph rendering, and tensor value inspection. Introduced a new test worker (\test/worker.js\) and mock environment (\test/mock.js\) to support isolated model loading and validation tests, along with a Python backend test script (\test/backend.py\) and a measures analysis script (\test/measures.py\) for performance tracking.

test · high confidence

The project's dependencies have been updated to Electron 44.4.5, @playwright/test 1.63.0, and ESLint 10.11.0, alongside electron-builder 26.16.1 and @electron/notarize 3.1.1. These updates ensure compatibility with the latest Electron runtime and testing tooling while maintaining the existing electron-updater dependency for application updates.

(dependencies) · high confidence

The application runtime has been upgraded to Electron 42.0.1. This update brings the underlying Chromium and Node.js versions current, which may affect browser compatibility, performance, and security posture for the desktop application.

source · high confidence

Written by watchdog.canine.dev from the codebase's own history, inside the signed delivery this page is composed from.

• CAI 32 → 34 (+2.0)
• Rubric changed (rubric-2026.08.15 → rubric-2026.09.15) — scores are not directly comparable.

• Code Health 20 → 14 (-5.2)
• Architecture 95 → 97 (+1.1)
• Maturity 57 → 59 (+1.8)
• Readiness 29 → 51 (+21.4)
• Security 42 → 69 (+26.5)
• Accessibility 40 (new)

• (anonymous) (cognitive 16) (source/view.js)
• (anonymous) (cognitive 17) (source/caffe.js)
• (anonymous) (cognitive 17) (source/onnx.js)
• (anonymous) (cognitive 18) (source/circle.js)
• (anonymous) (cognitive 18) (source/tflite.js)
• (anonymous) (cognitive 20) (source/coreml.js)
• (anonymous) (cognitive 20) (source/megengine.js)
• (anonymous) (cognitive 23) (source/paddle.js)
• (anonymous) (cognitive 26) (source/caffe2.js)
• (anonymous) (cognitive 26) (source/keras.js)
• (anonymous) (cognitive 35) (source/cntk.js)
• (anonymous) (cognitive 36) (source/tf.js)
• (anonymous) (cognitive 57) (source/acuity.js)
• (anonymous) (cyclomatic 18) (source/onnx.js)
• (anonymous) (cyclomatic 19) (source/caffe2.js)
• (anonymous) (cyclomatic 22) (source/cntk.js)
• (anonymous) (cyclomatic 28) (source/acuity.js)
• (anonymous) (cyclomatic 33) (source/paddle.js)
• (anonymous) (cyclomatic 39) (source/tf.js)
• BatchNormalization (cognitive 21) (source/cntk.js)
• …and 783 more

• Change coupling: caffe-proto.js ↔ tf-proto.js (source/caffe-proto.js)
• Change coupling: pytorch.py ↔ pytorch_script.py (source/pytorch.py)
• Change-coupling hub: uff-proto.js → caffe-proto.js, caffe2-proto.js, onnx-proto.js, paddle-proto.js, protoc.js (source/uff-proto.js)
• Documentation: no usage examples (README.md)
• High CVE: [GHSA redacted] (package-lock.json)
• High CVE: [GHSA redacted] (package-lock.json)
• High CVE: [GHSA redacted] (package-lock.json)
• High: security finding (details withheld)
• High: security finding (details withheld)
• High: security finding (details withheld)
• High: security finding (details withheld)
• High: security finding (details withheld)
• High: security finding (details withheld)
• High: security finding (details withheld)
• High: security finding (details withheld)
• High: security finding (details withheld)
• High: security finding (details withheld)
• High: security finding (details withheld)
• High: security finding (details withheld)
• High: security finding (details withheld)
• …and 10 more

• 101 commits — 101 feature/other, 0 fixes

• (root) — 27 commits
• source/mlir-metadata.json — 19 commits
• source/onnx-metadata.json — 16 commits
• source/gguf-metadata.json — 9 commits
• source/litertlm-proto.js — 7 commits
• source/sentencepiece-proto.js — 3 commits
• source/base.js — 2 commits
• source/pytorch-metadata.json — 2 commits
• source/coreml.js — 1 commit
• source/desktop.mjs — 1 commit
• source/executorch-schema.js — 1 commit
• source/gguf.js — 1 commit
• source/keras.js — 1 commit
• source/litertlm-schema.js — 1 commit
• source/mnn-schema.js — 1 commit
• source/onnx-proto.js — 1 commit
• source/onnx.js — 1 commit
• source/sklearn-metadata.json — 1 commit
• source/tf-metadata.json — 1 commit
• source/tf-proto.js — 1 commit

• change: Add GGUF support (#1209)
• change: Add GGUF support (#1209)
• change: Add GGUF support (#1209)
• change: Add GGUF support (#1209)
• change: Add GGUF support (#1209)
• change: Add GGUF support (#1209)
• change: Add GGUF support (#1209)
• change: Add GGUF support (#1209)
• change: Add MLIR support (#1044)
• change: Add MLIR support (#1044)
• change: Add MLIR support (#1044)
• change: Add MLIR support (#1044)
• change: Add MLIR support (#1044)
• change: Add MLIR support (#1044)
• change: Add MLIR support (#1044)
• change: Add MLIR support (#1044)
• change: Add MLIR support (#1044)
• change: Add MLIR support (#1044)
• change: Add MLIR support (#1044)
• change: Add MLIR support (#1044)
• …and 81 more

Written by watchdog.canine.dev from the codebase's own history, inside the signed delivery this page is composed from.

lutzroeder/netron was measured the same way every project in this corpus was: the same rubric, at a pinned commit, with the result published in full. Point a surveyor at a repository you know and see whether you agree with it.