Fix for FakeQuantize-1 conformance tests

[OleksiiZaderykhinCapgemini]

The convert node was extended to support the conversion from Precision::BIN precision. The issue with FakeQuantize-1 conformance tests failures is related to the convert node. In the case of binarization, the FakeQuantize node is transformed into a binarization node and the convert node. But the convert node didn't support conversion from Precision::BIN precision.

Ticket: 89297

[ceciliapeng2011]

@dmitry-gorokhov anyone in Russia team could have a look at this PR?

[dmitry-gorokhov]

BTW there are test cases for BIN precision conversion, but disable at the moment: https://github.com/openvinotoolkit/openvino/blob/master/src/tests/functional/plugin/cpu/shared_tests_instances/skip_tests_config.cpp#L142. Please try to enable them, hope test infra works correct with binary data type.

[dmitry-gorokhov]

@OleksiiZaderykhinAltran Could please also measure binary size increase for intel_cpu plugin library? We have a requirement to keep binary size as small as possible, but here there are a lot of instances of the same conversion function. So it is better to understand the impact.

[OleksiiZaderykhinCapgemini]

Hi, @dmitry-gorokhov, sorry for the delay. Actually, I tried to enable the single-layer tests for the convert operation for the BIN precision. But I had several errors that are not related to the convert operation. As I understand the BIN precision is not fully supported yet. Also, I measured the binary size for libopenvino_intel_cpu_plugin.so. It was 38031104 bytes before. And it is 38059776 bytes after. (+28672 bytes)

[dmitry-gorokhov]

/azp run openvino-win

[azure-pipelines[bot]]

Azure Pipelines successfully started running 1 pipeline(s).