openVino vs PyTorch Tracing

Dear Community,

i followed the examples for PyTorch C++ export and Tracing, which works really nicely.
However, there is also openVino, which claims to “support heterogeneous execution across an Intel® CPU, Intel® Integrated Graphics, Intel® Neural Compute Stick 2 […]”.

Does PyTorch Tracing also support CPU with Integrated Graphics/built-in AI Acceleration? If yes, are there additional steps needed from the dev side to make it run?


PS: Does someone know a good source to compare openVino and PyTorch tracing?