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.

Qiao Longfei 96d5500934 optimize code		8 years ago
..
detail	Refine activation for GRU operator (#13275)	8 years ago
CMakeLists.txt	Merge pull request #14048 from velconia/change_sequence_pool_to_cpu	8 years ago
algorithm.h	test=develop	8 years ago
blas.cc	Follow comments and polish code names	8 years ago
blas.h	fix bugs	8 years ago
blas_impl.cu.h	matmul support float16/double	8 years ago
blas_impl.h	fix bugs	8 years ago
compound_functors.h	Feature/op_fuse_pass (#12440)	8 years ago
concat.hip.cu	…
concat_and_split.cc	Refine Split op (#13967)	8 years ago
concat_and_split.cu	Refine Split op (#13967)	8 years ago
concat_and_split.h	Refine Split op (#13967)	8 years ago
concat_test.cc	Refine Split op (#13967)	8 years ago
context_project.cc	…
context_project.cu	…
context_project.h	Clean and extract blas	8 years ago
cos_sim_functor.cc	…
cos_sim_functor.cu	Feature/cuda9 cudnn7 (#10140)	8 years ago
cos_sim_functor.h	…
cpu_vec.h	refine vmul	8 years ago
cpu_vec_test.cc	refine vmul	8 years ago
cross_entropy.cc	Add ignore_index in cross_entropy op (#13217)	8 years ago
cross_entropy.cu	Add ignore_index in cross_entropy op (#13217)	8 years ago
cross_entropy.h	Add ignore_index in cross_entropy op (#13217)	8 years ago
depthwise_conv.cu	optimize depthwise conv by register memory (#13778)	8 years ago
depthwise_conv.h	Optimization of Kernels that related to DeepLabv3+ (#13534)	8 years ago
fc_compute.h	refine and add eltadd_relu unit test	8 years ago
functors.h	Feature/op_fuse_pass (#12440)	8 years ago
gru_compute.cc	Clean and extract blas	8 years ago
gru_compute.cu	Clean and extract blas	8 years ago
gru_compute.h	…
im2col.cc	complete im2col with padding==1 and speedup filter width==1	8 years ago
im2col.cu	change im2sequence for ctc batch inference (#11696)	8 years ago
im2col.h	…
im2col_cfo_cpu.h	further optimize im2col using variables	8 years ago
im2col_test.cc	complete im2col with padding==1 and speedup filter width==1	8 years ago
jit_kernel.cc	thread local jit kernels	8 years ago
jit_kernel.h	add fusion gru jit kernel	8 years ago
jit_kernel_blas.cc	refine and add eltadd_relu unit test	8 years ago
jit_kernel_exp.cc	clean code exp avx	8 years ago
jit_kernel_macro.h	refine code and fix	8 years ago
jit_kernel_rnn.cc	fix avx error	8 years ago
jit_kernel_test.cc	refine and add eltadd_relu unit test	8 years ago
lstm_compute.cc	…
lstm_compute.cu	…
lstm_compute.h	…
math_function.cc	"fix compile error" (#13579)	8 years ago
math_function.cu	Feature/template (#13093)	8 years ago
math_function.h	"fix compile error" (#13579)	8 years ago
math_function_impl.h	add hsigmoid	8 years ago
math_function_test.cc	add gemm_warp test	8 years ago
math_function_test.cu	Clean and extract blas	8 years ago
matrix_bit_code.cc	Fix the HierarchicalSigmoidGradOpKernel and refine the codes. Now hsigmoid_op is same with V2 implementation and can pass gradient check.	8 years ago
matrix_bit_code.h	fix style (#13142)	8 years ago
maxouting.cc	…
maxouting.cu	Feature/cuda9 cudnn7 (#10140)	8 years ago
maxouting.h	fix style (#13142)	8 years ago
padding.h	Add pad_constant_like_op (#12943)	8 years ago
pooling.cc	fix cpplint error (#10329)	8 years ago
pooling.cu	fix cpplint error (#10329)	8 years ago
pooling.h	fix style (#13142)	8 years ago
sampler.cc	…
sampler.h	…
selected_rows_functor.cc	optimize code	8 years ago
selected_rows_functor.cu	optimize code	8 years ago
selected_rows_functor.h	Merge branch 'develop' of https://github.com/PaddlePaddle/Paddle into optimize-sum-seq-pooling-op	8 years ago
selected_rows_functor_test.cc	sum op handle empty input	8 years ago
selected_rows_functor_test.cu	sum op handle empty input	8 years ago
sequence2batch.cc	refine seq2batch	8 years ago
sequence2batch.cu	…
sequence2batch.h	refine cpu gru batch mode	8 years ago
sequence_padding.cc	update function	8 years ago
sequence_padding.cu	update function	8 years ago
sequence_padding.h	bug fix	8 years ago
sequence_padding_test.cc	Complete sequence_padding GPU kernel	8 years ago
sequence_pooling.cc	Move code to SumSeqPoolGradFunctor	8 years ago
sequence_pooling.cu	refine seq_pool	8 years ago
sequence_pooling.h	…
sequence_pooling_test.cc	Fix copy wrong pos bug	8 years ago
sequence_scale.cc	…
sequence_scale.cu	Feature/cuda9 cudnn7 (#10140)	8 years ago
sequence_scale.h	…
softmax.cc	…
softmax.cu	Revert ""cherry picked operators changes" (#12184)" (#12747)	8 years ago
softmax.h	…
softmax_impl.h	…
unpooling.cc	…
unpooling.cu	Feature/cuda9 cudnn7 (#10140)	8 years ago
unpooling.h	…
vol2col.cc	…
vol2col.cu	Feature/cuda9 cudnn7 (#10140)	8 years ago
vol2col.h	…
vol2col_test.cc	…