10 Commits (8894c67d7168dfa5f5dc8e57ec2b5c60f24e368c)

Author	SHA1	Message	Date
chengduoZH	782ddc5f59	follow comments	7 years ago
chengduoZH	0468422d06	follow comments	7 years ago
chengduoZH	259858b41b	modify doc	7 years ago
chengduoZH	578d60bf4a	code refine	7 years ago
chengduoZH	2edc136c5f	add 4-d for matmul_op	7 years ago
Luo Tao	761b329793	unify the indentation of license	7 years ago
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
dangqingqing	f5e367655e	Use G++ to compile some cu operators.	7 years ago
Yu Yang	c532b96741	Remove template parameter for Tensor methods (#4937) ... * Remove template parameter for Tensor methods * Also check the type is correct when data() * Simplize holder_ * Fix accuracy_op * Register Code	7 years ago
Markus Kliegl	164898277c	MatMul operator (#4856) ... * initial matmul operator Similar to np.matmul, but also has transpose_X and transpose_Y flags, and only supports tensors from rank 1 to 3 inclusive. For GPU, uses cublas?gemmStridedBatched. For CPU, uses cblas_?gemm_batch if available via MKL; otherwise a simple serial implementation that loops over the batch dimension is employed for now.	7 years ago