Paddle

Author	SHA1	Message	Date
QI JUN	61ec0b9516	Refine device context (#6433 ) There are mainly following fixes: - take `DeviceContext` as the template parameter of math functors and OpKernel instead of `Place` - remove `eigen_device` interface in base class `DeviceContext` - remove `GetEigenDevice` interface in `ExecutionContext` and base class `DeviceContext` - remove unused `platform::EigenDeviceConverter` - rename `REGISTER_OP_GPU_KERNEL` to `REGISTER_OP_CUDA_KERNEL` - rename `USE_GPU_ONLY_OP` to `USE_CUDA_ONLY_OP`	7 years ago
guosheng	3e552cdcac	Fix gru_op related code style	7 years ago
guosheng	aa83e19e24	Remove lstm_op including in gru_op	7 years ago
guosheng	afd1f36186	Merge branch 'develop' of https://github.com/PaddlePaddle/paddle into fix-H0-GRUOp-dev	7 years ago
guosheng	75426e013a	Refine GRU Operator	7 years ago
dangqingqing	2673657684	Move RowwiseAdd functor to math_funcion and Add ColwiseSum functor.	7 years ago
guosheng	b103072dc8	Fix data order of H0 in GRU Operator	7 years ago
dangqingqing	f5e367655e	Use G++ to compile some cu operators.	7 years ago
guosheng	bb75381444	Clean code of GRU Operator	7 years ago
guosheng	53d8165f53	Make GRU Operator adapt to sequence2batch	7 years ago
guosheng	b87eabae56	Add GRU Operator	7 years ago