#2626th of 37 in ML frameworks and libraries
onnxruntime
Cross-platform inference engine for ONNX models with hardware accelerator support
Documentation ↗
(opens in a new tab)Website ↗
(opens in a new tab)Repository on GitHub ↗
(opens in a new tab)
- Stars
- 22k
- License
- MIT
- Last commit
- Oct 2026
- Last release
- Sep 2026
- Language
- C++
Missed the latest GitHub check; these facts may be older.
Overview
ONNX Runtime runs ONNX models across different hardware, drivers and operating systems, applying graph optimizations and using hardware accelerators where available. It takes models exported from PyTorch, TensorFlow/Keras and classical libraries such as scikit-learn, LightGBM and XGBoost. A separate training mode accelerates multi-node NVIDIA GPU training of transformer models with a one-line change to existing PyTorch scripts.
Who it is for: Engineers deploying trained models on varied hardware
Strengths
- Accepts models from PyTorch, TensorFlow/Keras, scikit-learn, LightGBM and XGBoost via ONNX
- Runs on multiple hardware targets and operating systems with graph optimizations
- MIT license, with docs, roadmap and companion example repositories
- Training acceleration needs one added line in existing PyTorch scripts
Weaknesses
- Library, not a server: README describes no built-in HTTP API
- Models must be converted to ONNX format first
- README lists no execution providers, language bindings or system requirements
- May collect usage data and send it to Microsoft; see privacy statement
What it needs
- GPU optional
- Models: ONNX, PyTorch, TensorFlow/Keras, scikit-learn, LightGBM
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 |