Executorch initial support

[cavusmustafa]

Details:

• OV side changes for initial ExecuTorch OV backend

Tickets:

• ticket-id

[rkazants]

I have several question about this feature:

1. Will is support ExecuTorch binary compiled to OV ExecuTorch backend?
2. What other ExecuTorch binaries will be supported: default, xnnpack?
3. We won't support ExecuTorch binary inference using native OpenVINO workflow, right?
4. Do you have PR/branch with ExecuTorch OV backend implementation?

Best regards,
Roman

[cavusmustafa]

I have several question about this feature:

1. Will is support ExecuTorch binary compiled to ExecuTorch backend?
2. What other ExecuTorch binaries will be supported: default, xnnpack?
3. We won't support ExecuTorch binary inference using native OpenVINO workflow, right?
4. Do you have PR/branch with ExecuTorch OV backend implementation?

Best regards, Roman

1. Yes, the changes included in this PR to support ExecuTorch OpenVINO backend.
2. POC will support OpenVINO backend only but we are also working on enabling fallback to default backend for unsupported ops.
3. ExecuTorch binary inference will be using OpenVINO workflow similar to inference execution with torch.compile. The partition compiled for OpenVINO execution (which is stored in .pte file) is executed in ExecuTorch OpenVINO backend which uses OpenVINO C++ API.
4. Branch: https://github.com/ynimmaga/executorch/tree/openvino_backend

Please let me know for additional questions and clarifications needed.