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.

GaoWei8 5af0c7ba89 Modify padding strategy: remove weight copy in fc padding (#21650) ... test=develop		5 years ago
..
detail	Add dependency for error_codes.proto (#21084)	5 years ago
CMakeLists.txt	…
algorithm.h	…
beam_search.cc	…
beam_search.cu	…
beam_search.h	…
beam_search_test.cc	…
blas.cc	…
blas.h	…
blas_impl.cu.h	…
blas_impl.h	add dequantize_abs_max op and modify lookup_table op (#20899)	5 years ago
bloomfilter.h	refine murmurhash3_x64_128 for bloom_filter (#20996)	5 years ago
compound_functors.h	…
concat.hip.cu	…
concat_and_split.cc	…
concat_and_split.cu	…
concat_and_split.h	…
concat_test.cc	…
context_project.cc	…
context_project.cu	…
context_project.h	…
cos_sim_functor.cc	…
cos_sim_functor.cu	…
cos_sim_functor.h	…
cpu_vec.h	…
cpu_vec_test.cc	fix -Wno-error=sign-compare warning in gcc8 (#21434)	5 years ago
cross_entropy.cc	…
cross_entropy.cu	…
cross_entropy.h	…
depthwise_conv.cu	fix the bug of conv_transpose:compatible with Anylayout setting, test=develop (#20897)	5 years ago
depthwise_conv.h	…
fc.cc	Modify padding strategy: remove weight copy in fc padding (#21650)	5 years ago
fc.cu	Add fc padding to improve mkl GEMM's performance when N and K are multiple of 128. (#20972)	5 years ago
fc.h	Add fc padding to improve mkl GEMM's performance when N and K are multiple of 128. (#20972)	5 years ago
functors.h	…
gru_compute.cc	…
gru_compute.cu	Fix ce ocr_recognition test fails (#20987)	5 years ago
gru_compute.h	…
im2col.cc	fix the bug of conv_transpose:compatible with Anylayout setting, test=develop (#20897)	5 years ago
im2col.cu	fix the bug of conv_transpose:compatible with Anylayout setting, test=develop (#20897)	5 years ago
im2col.h	…
im2col_cfo_cpu.h	…
im2col_test.cc	…
lstm_compute.cc	…
lstm_compute.cu	…
lstm_compute.h	…
math_function.cc	…
math_function.cu	…
math_function.h	…
math_function_impl.h	…
math_function_test.cc	…
math_function_test.cu	…
matrix_bit_code.cc	…
matrix_bit_code.h	…
maxouting.cc	maxout supports channel_last input (#20846)	5 years ago
maxouting.cu	maxout supports channel_last input (#20846)	5 years ago
maxouting.h	maxout supports channel_last input (#20846)	5 years ago
padding.h	remove -Wno-error=sign-compare, make warning as error (#21358)	5 years ago
pooling.cc	…
pooling.cu	…
pooling.h	…
prelu.cu	fix the computation for dx (grad for x) for prelu operation. (#20949)	5 years ago
prelu.h	fix the computation for dx (grad for x) for prelu operation. (#20949)	5 years ago
sample_prob.cc	…
sample_prob.cu	…
sample_prob.h	…
sampler.cc	…
sampler.h	…
selected_rows_functor.cc	…
selected_rows_functor.cu	…
selected_rows_functor.h	…
selected_rows_functor_test.cc	…
selected_rows_functor_test.cu.cc	…
sequence2batch.cc	…
sequence2batch.cu	…
sequence2batch.h	…
sequence_padding.cc	…
sequence_padding.cu	…
sequence_padding.h	Fix warpctc in padding mode. (#21033)	5 years ago
sequence_padding_test.cc	…
sequence_pooling.cc	…
sequence_pooling.cu	…
sequence_pooling.h	…
sequence_pooling_test.cc	…
sequence_scale.cc	…
sequence_scale.cu	…
sequence_scale.h	…
softmax.cc	…
softmax.cu	…
softmax.h	…
softmax_impl.h	…
tree2col.cc	…
tree2col.cu	…
tree2col.h	…
unpooling.cc	…
unpooling.cu	…
unpooling.h	…
vol2col.cc	…
vol2col.cu	…
vol2col.h	…
vol2col_test.cc	…