Scott McKay 40d80cde8f Rework CDist (#3393) ... * Make CDist faster via Eigen squaredNorma and GEMM. * Add call to abs() as the GEMM output may differ slightly due to floating point accuracy and result in a negative distance which returns NaN if sqrt() is applied to it. * Update math::Gemm to use the type for alpha and beta instead of hardcoding to float. Matches the GemmEx definition. * Provide Eigen based replication of the GEMM call on x86 if T=double. * Make test model data deterministic. * Do the GEMM first so we can avoid potentially subtracting two numbers that are very close to each other.		2020-04-09 14:05:25 +10:00
..
cpu	Rework CDist (#3393)	2020-04-09 14:05:25 +10:00
cuda	Add CPU implementation for FastGelu operator (#3398)	2020-04-07 00:19:30 -07:00
cpu_contrib_kernels.cc	Reduce binary size of Slice implementation (#3238)	2020-04-08 07:19:29 +10:00
cpu_contrib_kernels.h	Fix kernel registry bug (#2137)	2019-10-17 23:10:54 -07:00
cuda_contrib_kernels.cc	Revert an op version change (#3026)	2020-02-18 09:43:18 -08:00
cuda_contrib_kernels.h	move all contrib ops to contrib ops namespace (#1190)	2019-06-24 10:19:01 -07:00