Change the compute type for the F16 wrapper around cublas GEMMEx.
The original code had CUDA_R_32F but was changed to CUBLAS_16F inadvertently during the rocblas->hipblas changes in PyTorch. Hence, the bug was introduced.
This change will allow the unit test for test_fused_dense.py to pass sucessfully.
Change the compute type for the F16 wrapper around cublas GEMMEx.
The original code had
CUDA_R_32F
but was changed toCUBLAS_16F
inadvertently during the rocblas->hipblas changes in PyTorch. Hence, the bug was introduced.This change will allow the unit test for
test_fused_dense.py
to pass sucessfully.