Onnx Runtime: “Cross-Platform Accelerated Machine Learning”

Re:Onnx Runtime: “Cross-Platform Accelerated Machine Learning” by FloatArtifact on Tuesday July 25, 2023 @08:17PM (#36871508)

Would it not be better to use https://github.com/tinygrad/tinygrad as an intermediary framework?

[ Parent | on HN ]

Re:Onnx Runtime: “Cross-Platform Accelerated Machine Learning” by new_user55 on Tuesday July 25, 2023 @08:59PM (#36871829)

Tinygrad is python only right? Can it provide gradients during C++ runtime as well? ONNX runtime have multiple language backends for inference.

[ Parent | on HN ]

Re:Onnx Runtime: “Cross-Platform Accelerated Machine Learning” by figomore on Tuesday July 25, 2023 @10:10PM (#36872313)

One option to your case is OpenVino. It's written in C++ and has Python Bindings. Also, it can be used to train new nets. You can use ONNX files with OpenVino too.

[ Parent | on HN ]