You can not select more than 25 topics Topics must start with a chinese character,a letter or number, can include dashes ('-') and can be up to 35 characters long.
 
 
 
 
 
 
Dong Xu 28154dcb29 fix vst1.f32 of coeff sum at eltwise_arm layer 8 years ago
..
arm fix vst1.f32 of coeff sum at eltwise_arm layer 8 years ago
x86 call Innerproduct for convolution on flattened blob 8 years ago
absval.cpp forward reuse forward_inplace routine, reduce binary size with little memcpy overhead in non-light mode 8 years ago
absval.h forward reuse forward_inplace routine, reduce binary size with little memcpy overhead in non-light mode 8 years ago
argmax.cpp new param load api 8 years ago
argmax.h new param load api 8 years ago
batchnorm.cpp element type storage support in Mat, move data member the first so that a pointer to Mat is a pointer to data, convenient index access for float vector 8 years ago
batchnorm.h new model load api, fix #215 8 years ago
bias.cpp element type storage support in Mat, move data member the first so that a pointer to Mat is a pointer to data, convenient index access for float vector 8 years ago
bias.h new model load api, fix #215 8 years ago
binaryop.cpp inplace binaryop with scalar 8 years ago
binaryop.h inplace binaryop with scalar 8 years ago
bnll.cpp forward reuse forward_inplace routine, reduce binary size with little memcpy overhead in non-light mode 8 years ago
bnll.h forward reuse forward_inplace routine, reduce binary size with little memcpy overhead in non-light mode 8 years ago
concat.cpp concat slice shufflechannel honor elemsize 8 years ago
concat.h merge concatv2 dropoutv2 softmaxv2 8 years ago
convolution.cpp call Innerproduct for convolution on flattened blob 8 years ago
convolution.h new model load api, fix #215 8 years ago
convolutiondepthwise.cpp decouple convolutiondepthwise and convolution, reduce binary size by 10%, fix #254 8 years ago
convolutiondepthwise.h decouple convolutiondepthwise and convolution, reduce binary size by 10%, fix #254 8 years ago
crop.cpp new param load api 8 years ago
crop.h new param load api 8 years ago
deconvolution.cpp decouple convolutiondepthwise and convolution, reduce binary size by 10%, fix #254 8 years ago
deconvolution.h new model load api, fix #215 8 years ago
deconvolutiondepthwise.cpp decouple convolutiondepthwise and convolution, reduce binary size by 10%, fix #254 8 years ago
deconvolutiondepthwise.h decouple convolutiondepthwise and convolution, reduce binary size by 10%, fix #254 8 years ago
detectionoutput.cpp element type storage support in Mat, move data member the first so that a pointer to Mat is a pointer to data, convenient index access for float vector 8 years ago
detectionoutput.h new param load api 8 years ago
dropout.cpp forward reuse forward_inplace routine, reduce binary size with little memcpy overhead in non-light mode 8 years ago
dropout.h forward reuse forward_inplace routine, reduce binary size with little memcpy overhead in non-light mode 8 years ago
eltwise.cpp element type storage support in Mat, move data member the first so that a pointer to Mat is a pointer to data, convenient index access for float vector 8 years ago
eltwise.h new param load api 8 years ago
elu.cpp forward reuse forward_inplace routine, reduce binary size with little memcpy overhead in non-light mode 8 years ago
elu.h forward reuse forward_inplace routine, reduce binary size with little memcpy overhead in non-light mode 8 years ago
embed.cpp enable embed layer 8 years ago
embed.h new model load api, fix #215 8 years ago
exp.cpp forward reuse forward_inplace routine, reduce binary size with little memcpy overhead in non-light mode 8 years ago
exp.h forward reuse forward_inplace routine, reduce binary size with little memcpy overhead in non-light mode 8 years ago
expanddims.cpp new param load api 8 years ago
expanddims.h new param load api 8 years ago
flatten.cpp element type storage support in Mat, move data member the first so that a pointer to Mat is a pointer to data, convenient index access for float vector 8 years ago
flatten.h add ncnn source qwq 9 years ago
innerproduct.cpp innerproduct produce continous blob, fix #236 8 years ago
innerproduct.h new model load api, fix #215 8 years ago
input.cpp change input param order to w h c, replace caffe MemoryData to Input 8 years ago
input.h change input param order to w h c, replace caffe MemoryData to Input 8 years ago
instancenorm.cpp implement instancenorm 8 years ago
instancenorm.h implement instancenorm 8 years ago
interp.cpp forward reuse forward_inplace routine, reduce binary size with little memcpy overhead in non-light mode 8 years ago
interp.h forward reuse forward_inplace routine, reduce binary size with little memcpy overhead in non-light mode 8 years ago
log.cpp forward reuse forward_inplace routine, reduce binary size with little memcpy overhead in non-light mode 8 years ago
log.h forward reuse forward_inplace routine, reduce binary size with little memcpy overhead in non-light mode 8 years ago
lrn.cpp load LRN bias from param 8 years ago
lrn.h load LRN bias from param 8 years ago
lstm.cpp element type storage support in Mat, move data member the first so that a pointer to Mat is a pointer to data, convenient index access for float vector 8 years ago
lstm.h new model load api, fix #215 8 years ago
memorydata.cpp new model load api, fix #215 8 years ago
memorydata.h new model load api, fix #215 8 years ago
mvn.cpp element type storage support in Mat, move data member the first so that a pointer to Mat is a pointer to data, convenient index access for float vector 8 years ago
mvn.h new param load api 8 years ago
normalize.cpp implement spatial norm, convert L2Normalization 8 years ago
normalize.h implement spatial norm, convert L2Normalization 8 years ago
padding.cpp new param load api 8 years ago
padding.h new param load api 8 years ago
permute.cpp [fix-compile-warnings] fix compiler warnings, and add werror in CMakeLists.txt (#217) 8 years ago
permute.h new param load api 8 years ago
pooling.cpp 1.fix pad tail bug in commit d1ea2a3 at pooling layer 8 years ago
pooling.h rewrite pooling pad scheme, global pooling return continous blob 8 years ago
power.cpp forward reuse forward_inplace routine, reduce binary size with little memcpy overhead in non-light mode 8 years ago
power.h forward reuse forward_inplace routine, reduce binary size with little memcpy overhead in non-light mode 8 years ago
prelu.cpp prelu on vector and image 8 years ago
prelu.h new model load api, fix #215 8 years ago
priorbox.cpp element type storage support in Mat, move data member the first so that a pointer to Mat is a pointer to data, convenient index access for float vector 8 years ago
priorbox.h new param load api 8 years ago
proposal.cpp element type storage support in Mat, move data member the first so that a pointer to Mat is a pointer to data, convenient index access for float vector 8 years ago
proposal.h new param load api 8 years ago
reduction.cpp element type storage support in Mat, move data member the first so that a pointer to Mat is a pointer to data, convenient index access for float vector 8 years ago
reduction.h new model load api, fix #215 8 years ago
relu.cpp forward reuse forward_inplace routine, reduce binary size with little memcpy overhead in non-light mode 8 years ago
relu.h forward reuse forward_inplace routine, reduce binary size with little memcpy overhead in non-light mode 8 years ago
reshape.cpp new param load api 8 years ago
reshape.h new param load api 8 years ago
rnn.cpp element type storage support in Mat, move data member the first so that a pointer to Mat is a pointer to data, convenient index access for float vector 8 years ago
rnn.h new model load api, fix #215 8 years ago
roipooling.cpp new param load api 8 years ago
roipooling.h new param load api 8 years ago
scale.cpp innerproduct produce continous blob, fix #236 8 years ago
scale.h new model load api, fix #215 8 years ago
shufflechannel.cpp concat slice shufflechannel honor elemsize 8 years ago
shufflechannel.h code style changes 8 years ago
sigmoid.cpp forward reuse forward_inplace routine, reduce binary size with little memcpy overhead in non-light mode 8 years ago
sigmoid.h forward reuse forward_inplace routine, reduce binary size with little memcpy overhead in non-light mode 8 years ago
slice.cpp slice (#269) 8 years ago
slice.h slice dim 8 years ago
softmax.cpp element type storage support in Mat, move data member the first so that a pointer to Mat is a pointer to data, convenient index access for float vector 8 years ago
softmax.h forward reuse forward_inplace routine, reduce binary size with little memcpy overhead in non-light mode 8 years ago
split.cpp add ncnn source qwq 9 years ago
split.h add ncnn source qwq 9 years ago
spp.cpp element type storage support in Mat, move data member the first so that a pointer to Mat is a pointer to data, convenient index access for float vector 8 years ago
spp.h new param load api 8 years ago
squeeze.cpp new param load api 8 years ago
squeeze.h new param load api 8 years ago
tanh.cpp forward reuse forward_inplace routine, reduce binary size with little memcpy overhead in non-light mode 8 years ago
tanh.h forward reuse forward_inplace routine, reduce binary size with little memcpy overhead in non-light mode 8 years ago
threshold.cpp forward reuse forward_inplace routine, reduce binary size with little memcpy overhead in non-light mode 8 years ago
threshold.h forward reuse forward_inplace routine, reduce binary size with little memcpy overhead in non-light mode 8 years ago
tile.cpp new param load api 8 years ago
tile.h new param load api 8 years ago
unaryop.cpp element type storage support in Mat, move data member the first so that a pointer to Mat is a pointer to data, convenient index access for float vector 8 years ago
unaryop.h forward reuse forward_inplace routine, reduce binary size with little memcpy overhead in non-light mode 8 years ago