You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

Jeng Bai-Cheng fc93266b0a Improve qkv transpose performance (#23919) ... Use vector instruction (LDG.128) to improve qkv transpose. It provides 1.4X speedup at same GPU base frequency. test=develop		5 years ago
..
framework	add the support for pipeline (#24560)	5 years ago
imperative	modify into eager_tmp of Base Class test=develop (#25323)	5 years ago
inference	Improve qkv transpose performance (#23919)	5 years ago
memory	Refine error message in memory folder (#25095)	5 years ago
operators	randperm API: remove out, devive, stop_gradient; add name (#25410)	5 years ago
platform	remove useless property, test=develop (#25461)	5 years ago
pybind	add the c++ part of Imperative QAT. test=develop (#25446)	5 years ago
string	add timeout and http store in communication (#23436)	5 years ago
train	[Inference] [unittest] Inference unit tests rely on dynamic libraries (2) (#24859)	5 years ago
.clang-format	…
API.spec	modify pipeline optimizer to only support the mode of sync pipeline training (#25065)	5 years ago
CMakeLists.txt	merge develop to solve conflict, also fix API doc, test=develop (#18823)	6 years ago