Tracy Sharpe fe0b2b2abd QLinearConv speed up (#3196) ... For x86/x64 builds, change the QLinearConv op to use MLAS for the u8u8=s32 GEMM, then requantize the intermediate buffer to u8.		2020-03-13 16:54:55 -07:00
..
ci_build	QLinearConv speed up (#3196)	2020-03-13 16:54:55 -07:00
nuget	Enable DML Nuget Package for x64 or x86 architectures (#3120)	2020-03-02 20:18:46 -08:00
python	Fix cgmanifest.json generating script (#2770)	2020-01-14 14:59:07 -08:00