As AI continues to integrate into various applications, the ability for different machine learning (ML) models to operate across frameworks is becoming increasingly important. Models are often created and trained using one framework, but must be able to run in different environments, hardware and software engines. The Open Neural Network Exchange (ONNX) is an open standard which addresses this problem. ONNX simplifies the process of moving models between frameworks by offering a common set of operators and a portable file format.

My poster will provide an overview of ONNX and its role in the exchange and deployment of models. We will briefly discuss the architecture of ONNX files, its operator set, which supports a wide range of deep learning models, and the significance of having a standardized model representation in AI development.

ONNX is an open-source project and we are always looking for contributors with good ideas. I want to introduce you to the community behind ONNX and let you know how you can become a part of it. Our community includes representatives of large companies such as Microsoft, Nvidia, IBM and Intel, but also many smaller startups and individuals. Contributing to ONNX is straightforward, based on GitHub pull requests and code reviews. I will tell you about interesting areas where your contribution would be very valueable regardless of your previous expertise level in AI.

The ONNX project thrives on community contributions. Whether you're interested in adding new features, refining existing ones, or improving documentation, there's room for your input. This is an invitation to learn about ONNX and explore ways to contribute to its development.