Disabel GroupwiseQuantizedConvolution/0 for OpenCL backend (pytorch#3909

)

Summary:
Due to recent change to GroupwiseQuantizedConv (pytorch#3877), we need to change the implementation for OpenCL backend. This PR disable the test until the impl is changed since this backend isn't being actively used now.

Documentation:

[Optional Fixes #issue]
Pull Request resolved: pytorch#3909

Test Plan: Please see a detailed explanation of how to fill out the fields in the relevant sections in PULL_REQUEST.md.

Differential Revision: D19062388

Pulled By: yinghai

fbshipit-source-id: 9aae67655360f6ecb596685ecd512d2641343a7a

lib/Backends/OpenCL/tests/OpenCLOperatorTest.cpp

@@ -278,5 +278,6 @@ std::set<std::string> glow::backendTestBlacklist = {
     "RepeatedSLSWithPartialTensors/0",
     "GatherWithInt32PartialTensors/0",
     "GatherWithInt64PartialTensors/0",
+    "GroupwiseQuantizedConvolution/0",
     "ParallelBatchMatMul_Float16/0",
 };