- 26 12月, 2017 1 次提交
-
-
由 dzhwinter 提交于
* "fix threadpool style" * "remove header"
-
- 25 12月, 2017 12 次提交
-
-
由 dzhwinter 提交于
* "add data layout" * "need kernel registry support" * "fix data layout" * "reorder include headers" * "change enum to enum class" * "fix CI"
-
由 dangqingqing 提交于
-
由 qingqing01 提交于
-
由 Qiao Longfei 提交于
* init kernel hint * fix typo * rm unused code * add include in op_kernel.h * restore op_kernel since it will be moved to op_kernel_type * change force_cpu to use_cpu * fix compilation
-
由 QI JUN 提交于
* remove unused place * fix ci
-
由 Yang Yu 提交于
-
由 Luo Tao 提交于
-
由 dzhwinter 提交于
* "fix CopyFrom parameters" * "fix unused Place argument" * "fixed based on comment"
-
由 dzhwinter 提交于
-
由 Yang Yu 提交于
-
由 Yancey 提交于
* implement a simple threadpool * unlock before cv.notify * add done function * add lock with GetAvailable function * delete done_ * using call_once in GetInstance * update by comment * update comment * enhance unit test for multi threads task
-
由 qiaolongfei 提交于
-
- 24 12月, 2017 5 次提交
-
-
由 qiaolongfei 提交于
-
由 qiaolongfei 提交于
-
由 dzhwinter 提交于
-
由 QI JUN 提交于
* refine OpKernelKey * refine codes * fix code style * follow comments
-
由 dzhwinter 提交于
* "change operator interface" * "move devicepool to device_context" * "fix operator test" * "fix op_registry Run interface" * "net op passed. Need to fix nccl multi-Context" * "add nccl group function" * "add nccl group function" * "fix gpu count exceed 32 error" * "fix recurrent op, nccl op" * "change the other operators interface with Place" * "fix typo" * "fix pybind" * "fix device in python side" * "fix pybind failed" * "add init for test" * "fix CI"
-
- 23 12月, 2017 3 次提交
- 22 12月, 2017 6 次提交
-
-
由 Luo Tao 提交于
-
由 QI JUN 提交于
* add data layout * fix ci
-
由 QI JUN 提交于
-
由 Yang Yu 提交于
-
由 dzhwinter 提交于
* "remove GPU Sync Interface" * "fix typo" * "fix type cast error" * "fix related Copy with stream" * "fix failed tests with DevicePool" * "fix stupid removed position error"
-
由 xuwei06 提交于
For input argument with a list of variables, drop_empty_grad is not allowed because it makes the correspondence bewteen a variable and its gradient ambiguous. Use REGISTER_OP_EX to register the op or call InputGrad(?,false) in GradOpDescMaker.
-
- 21 12月, 2017 13 次提交
-
-
由 typhoonzero 提交于
-
由 typhoonzero 提交于
-
由 caoying03 提交于
-
由 typhoonzero 提交于
-
由 Yu Yang 提交于
* Remove unnecessary reshape in ColwiseSum Speed up 12s -> 10s. * Hand write ColwiseAdd in CPU
-
由 fengjiayi 提交于
-
由 Yang Yu 提交于
-
由 fengjiayi 提交于
-
由 fengjiayi 提交于
-
由 Yu Yang 提交于
* Rename XXDescBind --> XXDesc * Fix Compile
-
由 dzhwinter 提交于
-
由 Yang Yu 提交于
It is useful to reorder RNN memory block.
-
由 whs 提交于
-