openvinotoolkit
Added torchdynamo backend entry point in package setup
With this change, users don't have to add "import openvino.torch" in their applications. Providing torch.compile(model, backend="openvino") automatically imports the necessary modules. This simplifies the user experience and enables PT2 based benchmarking OOB without any changes. .
More information can be found here: https://github.com/pytorch/pytorch/pull/124536
JIRA: CVS-139370
import openvino.torch
should we deprecated this approach then? CC @jiwaszki what are the recommendations how to deprecate such approach?
What if OpenVINO is installed as archive? where we don't have setup.py involved to installation process.
import openvino.torch
should we deprecated this approach then? CC @jiwaszki what are the recommendations how to deprecate such approach?
What if OpenVINO is installed as archive? where we don't have
setup.pyinvolved to installation process.
In this case old approach can be used, right?
