Integrate ONNX Runtime with nDPI for AI model execution

[mmanoj]

Incorporation of ONNX Runtime:
Utilizing the ONNX Runtime can facilitate the deployment of machine learning models across various platforms, enhancing our system's flexibility and performance.

https://onnxruntime.ai/docs/install/

Adoption of libonnx: Employing libonnx, a lightweight, portable C99 ONNX inference engine, can optimize our operations on embedded devices, especially those with hardware acceleration support.

https://github.com/xboot/libonnx

Why this is important:

AI/ML models can be develop using various technologies and easy to integrate with nDPI without conversions and can run top of ONNX

@IvanNardi This is as per our initial discussion, lets discuss in more detail and fine tune the idea to work with more portable and modular manner

[lucaderi]

See also as reference https://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=10355065

[mmanoj]

@lucaderi thanks for the article, I will study and update for clarifications.

P.S. I found interesting techniques and few research papers related to the new optimized techniques we can use with Tiny ML. I will update it ASAP once I finalize.Based on that we can plan next steps.