onnx/onnx
🏗️ Frameworkonnx
Open standard for AI model interchange across frameworks.
ONNX (Open Neural Network Exchange) defines an extensible computation graph model with a set of operators and data types. It supports both deep neural networks and traditional ML models. Key features include operator set versioning, graph optimization, and support for multiple runtime backends. ONNX facilitates model portability and reduces vendor lock-in, allowing developers to train in one framework and deploy in another. The project includes a runtime for efficient inference and conversion tools for popular frameworks.
💡Highlights
- ├─21k GitHub stars
- ├─Supports PyTorch, TensorFlow, scikit-learn
- └─Graph-based model representation
🎯For
- ├─AI engineers
- ├─ML researchers
- └─DevOps teams