This is a patch release containing following changes to v1.6:
- Fixed performance regression for convolutions with 1x1 filter on Intel AVX2 (8186817)
- Fixed invalid memory access issue for bfloat16 1D grouped convolutions (9ebda65)
- Fixed
RuntimeError: label is redefined
for convolutions with large filter size on Intel AVX512 (f974b50) - Suppressed MSBuild warning MSB8065 (f91e641)
- Restricted support for shared virtual memory (SVM) to OpenCL 2.0 and later (fa6bbf4)