#2323rd of 37 in ML frameworks and libraries
onnx
Open model format and operator spec for exchanging ML models between frameworks
Documentation ↗
(opens in a new tab)Website ↗
(opens in a new tab)Repository on GitHub ↗
(opens in a new tab)
- Stars
- 21.6k
- License
- Apache-2.0
- Last commit
- Oct 2026
- Last release
- Oct 2026
- Language
- Python
Missed the latest GitHub check; these facts may be older.
Overview
ONNX defines an extensible computation graph format, built-in operators and standard data types for deep learning and traditional ML models, with a focus on inference. The Python package provides model loading and saving, shape and type inference, opset version conversion and a reference implementation. It is a spec and toolkit, not a runtime; execution is left to other engines.
Who it is for: Engineers exporting, converting or validating models across ML frameworks
Strengths
- Framework-neutral model format with documented IR spec, operator list and versioning rules
- Python API includes shape and type inference and opset version conversion
- Supported across many frameworks, tools and hardware, per the project
- abi3 wheels from Python 3.12; releases published to PyPI with weekly builds
Weaknesses
- Does not run models in production; a separate inference runtime is needed
- Scope is inference; training is not a stated focus
- Graph optimizer lives in a separate repository (onnx/optimizer)
- Reference implementation needs the optional onnx[reference] extra
What it needs
- no GPU
- Needs pytest (tests only)
- Models: ONNX, deep learning models, traditional ML models
Also in ML frameworks and libraries
See all 37| Rank | Project | Score |
|---|---|---|
| 1 | transformersPython library defining open models for text, vision, audio and multimodal tasks | 95 out of 100 |
| 2 | pytorchPython tensor library with GPU support and tape-based autograd for deep learning | 94 out of 100 |
| 3 | ultralyticsPython library and CLI for training and running YOLO vision models | 90 out of 100 |
| #4 | supervisionPython toolkit for loading, annotating and counting computer vision detections | 87 out of 100 |
| #5 | pytorch-image-modelsPyTorch library of image models, pretrained weights and reference training scripts | 86 out of 100 |