nihui/ncnn - ncnn - 开源协同云脑生态支撑系统

History

nihui 40a69a2dd3 discard riscv weight memory (#3874 ) * discard riscv innerproduct weight * drop riscv conv convdw weight * drop riscv deconv deconvdw weight		4 years ago
..
avx512_mathfun.h	Add SSE&AVX optimized for tan (#3765)	4 years ago

avx_mathfun.h	Add SSE&AVX optimized for tan (#3765)	4 years ago

batchnorm_x86.cpp	x86 avx512 optimization (#3581)	4 years ago

batchnorm_x86.h	Added AVX swish/lrn/batchnorm (#1897)	6 years ago

bias_x86.cpp	relu3d, batchnorm3d, reshape4d, flatten4d, permute4d (#3397)	4 years ago

bias_x86.h	LSTM arm/x86 + fp16 innerproduct arm (#1881)	6 years ago

binaryop_x86.cpp	binaryop type specialization (#3830)	4 years ago

binaryop_x86.h	X86 Elempack 8 AVX implementations. (#1853)	6 years ago

bnll_x86.cpp	Add unittest and SSE&AVX optimized for BNLL (#3759)	4 years ago

bnll_x86.h	Add unittest and SSE&AVX optimized for BNLL (#3759)	4 years ago

cast_fp16.h	x86 f16c infrastructure (#3577)	4 years ago

cast_x86.cpp	x86 f16c infrastructure (#3577)	4 years ago

cast_x86.h	Optimize FP32/FP16 conversion with AVX intrinsic. (#1545)	6 years ago

cast_x86_f16c.cpp	x86 f16c infrastructure (#3577)	4 years ago

clip_x86.cpp	x86 avx512 optimization (#3581)	4 years ago

clip_x86.h	X86 Elempack 8 AVX implementations. (#1853)	6 years ago

concat_x86.cpp	x86 avx512 optimization (#3581)	4 years ago

concat_x86.h	SSE2 optimization pack (#2123)	5 years ago

convolution1d_x86.cpp	x86 avx512 optimization (#3581)	4 years ago

convolution1d_x86.h	dynamic convolution weight (#3408)	4 years ago

convolution_1x1.h	discard weight memory for x86 arm vulkan (#3865)	4 years ago

convolution_1x1_int8.h	x86 avx2 optimization for convolution gemm int8 (#3489)	4 years ago

convolution_1x1_pack1to4.h	x86 sse2 optimization for conv1x1/3x3 pack4 and general sgemm pack4/pack4to1 (#3538)	4 years ago

convolution_1x1_pack1to4_int8.h	x86 avx2 optimization for convolution gemm int8 (#3489)	4 years ago

convolution_1x1_pack1to8.h	x86 avx fma optimization (#3543)	4 years ago

convolution_1x1_pack1to16.h	x86 avx512 optimization (#3581)	4 years ago

convolution_1x1_pack4.h	x86 sse2 optimization for conv1x1/3x3 pack4 and general sgemm pack4/pack4to1 (#3538)	4 years ago

convolution_1x1_pack4to1.h	x86 sse2 optimization for conv1x1/3x3 pack4 and general sgemm pack4/pack4to1 (#3538)	4 years ago

convolution_1x1_pack4to8.h	x86 avx fma optimization (#3543)	4 years ago

convolution_1x1_pack4to16.h	x86 avx512 optimization (#3691)	4 years ago

convolution_1x1_pack8.h	x86 avx fma optimization (#3543)	4 years ago

convolution_1x1_pack8to1.h	x86 avx fma optimization (#3543)	4 years ago

convolution_1x1_pack8to1_int8.h	x86 avx2 optimization for convolution gemm int8 (#3489)	4 years ago

convolution_1x1_pack8to4.h	always build tightly packed weight, fix #3545 (#3547)	4 years ago

convolution_1x1_pack8to4_int8.h	x86 avx2 optimization for convolution gemm int8 (#3489)	4 years ago

convolution_1x1_pack8to16.h	x86 avx512 optimization (#3581)	4 years ago

convolution_1x1_pack16.h	x86 avx512 optimization (#3581)	4 years ago

convolution_1x1_pack16to1.h	x86 avx512 optimization (#3691)	4 years ago

convolution_1x1_pack16to4.h	x86 avx512 optimization (#3581)	4 years ago

convolution_1x1_pack16to8.h	x86 avx512 optimization (#3581)	4 years ago

convolution_2x2_pack8.h	Added ability to switch AVX/AVX2 during runtime (#3076)	5 years ago

convolution_3x3.h	optimize x86 winograd input transform transpose (#3818)	4 years ago

convolution_3x3_int8.h	initial data structure changes for 3dcnn, conv3d, pooling3d (#3378)	4 years ago

convolution_3x3_pack1to4.h	added a number of optimized sse layers (#3302)	4 years ago

convolution_3x3_pack1to4_int8.h	x86 avx2 optimization for convolution gemm int8 (#3489)	4 years ago

convolution_3x3_pack1to8.h	Added ability to switch AVX/AVX2 during runtime (#3076)	5 years ago

convolution_3x3_pack4.h	fix winograd function name (#3820)	4 years ago

convolution_3x3_pack4to1.h	fix winograd function name (#3820)	4 years ago

convolution_3x3_pack8.h	fix winograd function name (#3820)	4 years ago

convolution_3x3_pack8to1.h	fix winograd function name (#3820)	4 years ago

convolution_3x3_pack8to1_int8.h	fix winograd function name (#3820)	4 years ago

convolution_3x3_pack8to4_int8.h	fix winograd function name (#3820)	4 years ago

convolution_3x3_pack16.h	fix winograd function name (#3820)	4 years ago

convolution_3x3_pack16to1.h	fix winograd function name (#3820)	4 years ago

convolution_5x5.h	rewrite convolution x86 sgemm pack1 (#3544)	4 years ago

convolution_7x7_pack1to4_int8.h	x86 avx2 optimization for convolution gemm int8 (#3489)	4 years ago

convolution_int8.h	architecture changes for int8 packing (#2771)	5 years ago

convolution_pack1to4.h	x86 avx fma optimization (#3543)	4 years ago

convolution_pack1to4_int8.h	some x86 sse2 optimization for convolution int8	5 years ago

convolution_pack1to8.h	x86 avx fma optimization (#3543)	4 years ago

convolution_pack1to16.h	x86 avx512 optimization (#3581)	4 years ago

convolution_pack4.h	x86 avx fma optimization (#3543)	4 years ago

convolution_pack4to1.h	x86 avx fma optimization (#3543)	4 years ago

convolution_pack4to8.h	x86 avx fma optimization (#3543)	4 years ago

convolution_pack4to16.h	x86 avx512 optimization (#3581)	4 years ago

convolution_pack8.h	x86 avx fma optimization (#3543)	4 years ago

convolution_pack8to1.h	x86 avx fma optimization (#3543)	4 years ago

convolution_pack8to1_int8.h	some x86 sse2 optimization for convolution int8	5 years ago

convolution_pack8to4.h	x86 avx fma optimization (#3543)	4 years ago

convolution_pack8to4_int8.h	some x86 sse2 optimization for convolution int8	5 years ago

convolution_pack8to16.h	x86 avx512 optimization (#3581)	4 years ago

convolution_pack16.h	x86 avx512 optimization (#3581)	4 years ago

convolution_pack16to1.h	fix build x86 avx512 source with old gcc (#3705)	4 years ago

convolution_pack16to4.h	x86 avx512 optimization (#3581)	4 years ago

convolution_pack16to8.h	x86 avx512 optimization (#3581)	4 years ago

convolution_sgemm.h	rewrite convolution x86 sgemm pack1 (#3544)	4 years ago

convolution_sgemm_int8.h	x86 f16c infrastructure (#3577)	4 years ago

convolution_sgemm_pack1to4.h	x86 sse2 optimization for conv1x1/3x3 pack4 and general sgemm pack4/pack4to1 (#3538)	4 years ago

convolution_sgemm_pack1to4_int8.h	x86 f16c infrastructure (#3577)	4 years ago

convolution_sgemm_pack1to8.h	x86 avx fma optimization (#3543)	4 years ago

convolution_sgemm_pack1to16.h	x86 avx512 optimization (#3581)	4 years ago

convolution_sgemm_pack4.h	x86 sse2 optimization for conv1x1/3x3 pack4 and general sgemm pack4/pack4to1 (#3538)	4 years ago

convolution_sgemm_pack4to1.h	x86 sse2 optimization for conv1x1/3x3 pack4 and general sgemm pack4/pack4to1 (#3538)	4 years ago

convolution_sgemm_pack4to8.h	always build tightly packed weight, fix #3545 (#3547)	4 years ago

convolution_sgemm_pack4to16.h	x86 avx512 optimization (#3691)	4 years ago

convolution_sgemm_pack8.h	always build tightly packed weight, fix #3545 (#3547)	4 years ago

convolution_sgemm_pack8to1.h	x86 avx fma optimization (#3543)	4 years ago

convolution_sgemm_pack8to1_int8.h	x86 f16c infrastructure (#3577)	4 years ago

convolution_sgemm_pack8to4.h	always build tightly packed weight, fix #3545 (#3547)	4 years ago

convolution_sgemm_pack8to4_int8.h	x86 f16c infrastructure (#3577)	4 years ago

convolution_sgemm_pack8to16.h	x86 avx512 optimization (#3581)	4 years ago

convolution_sgemm_pack16.h	x86 avx512 optimization (#3581)	4 years ago

convolution_sgemm_pack16to1.h	fix build x86 avx512 source with old gcc (#3705)	4 years ago

convolution_sgemm_pack16to4.h	x86 avx512 optimization (#3691)	4 years ago

convolution_sgemm_pack16to8.h	x86 avx512 optimization (#3581)	4 years ago

convolution_winograd_transform.h	optimize x86 winograd input transform transpose (#3818)	4 years ago

convolution_winograd_transform_pack4.h	fix winograd function name (#3820)	4 years ago

convolution_winograd_transform_pack8.h	fix winograd function name (#3820)	4 years ago

convolution_winograd_transform_pack16.h	fix winograd function name (#3820)	4 years ago

convolution_x86.cpp	discard weight memory for x86 arm vulkan (#3865)	4 years ago

convolution_x86.h	discard weight memory for x86 arm vulkan (#3865)	4 years ago

convolution_x86_avx2.cpp	fix winograd function name (#3820)	4 years ago

convolution_x86_avx512vnni.cpp	fix winograd function name (#3820)	4 years ago

convolution_x86_avxvnni.cpp	fix winograd function name (#3820)	4 years ago

convolution_x86_xop.cpp	fix winograd function name (#3820)	4 years ago

convolutiondepthwise_3x3.h	format code style and setup restyled.io (#1840)	6 years ago

convolutiondepthwise_3x3_int8.h	architecture changes for int8 packing (#2771)	5 years ago

convolutiondepthwise_3x3_pack4.h	x86 sse2 optimization for conv1x1/3x3 pack4 and general sgemm pack4/pack4to1 (#3538)	4 years ago

convolutiondepthwise_3x3_pack8.h	x86 avx fma optimization (#3543)	4 years ago

convolutiondepthwise_3x3_pack16.h	x86 avx512 optimization (#3581)	4 years ago

convolutiondepthwise_5x5_pack4.h	x86 sse2 optimization for conv1x1/3x3 pack4 and general sgemm pack4/pack4to1 (#3538)	4 years ago

convolutiondepthwise_5x5_pack8.h	x86 avx fma optimization (#3543)	4 years ago

convolutiondepthwise_5x5_pack16.h	x86 avx512 optimization (#3581)	4 years ago

convolutiondepthwise_x86.cpp	discard riscv weight memory (#3874)	4 years ago

convolutiondepthwise_x86.h	discard weight memory for x86 arm vulkan (#3865)	4 years ago

crop_x86.cpp	x86 avx512 optimization (#3581)	4 years ago

crop_x86.h	X86 Elempack 8 AVX implementations. (#1853)	6 years ago

deconvolution_pack1to4.h	x86 avx512 optimization for convolution winograd pack16to1 and deconvolution family, increase simpleomp argv count (#3694)	4 years ago

deconvolution_pack1to8.h	x86 avx512 optimization for convolution winograd pack16to1 and deconvolution family, increase simpleomp argv count (#3694)	4 years ago

deconvolution_pack1to16.h	x86 avx512 optimization for convolution winograd pack16to1 and deconvolution family, increase simpleomp argv count (#3694)	4 years ago

deconvolution_pack4.h	x86 avx512 optimization for convolution winograd pack16to1 and deconvolution family, increase simpleomp argv count (#3694)	4 years ago

deconvolution_pack4to1.h	x86 avx512 optimization for convolution winograd pack16to1 and deconvolution family, increase simpleomp argv count (#3694)	4 years ago

deconvolution_pack4to8.h	x86 avx512 optimization for convolution winograd pack16to1 and deconvolution family, increase simpleomp argv count (#3694)	4 years ago

deconvolution_pack4to16.h	x86 avx512 optimization for convolution winograd pack16to1 and deconvolution family, increase simpleomp argv count (#3694)	4 years ago

deconvolution_pack8.h	x86 avx512 optimization for convolution winograd pack16to1 and deconvolution family, increase simpleomp argv count (#3694)	4 years ago

deconvolution_pack8to1.h	x86 avx512 optimization for convolution winograd pack16to1 and deconvolution family, increase simpleomp argv count (#3694)	4 years ago

deconvolution_pack8to4.h	x86 avx512 optimization for convolution winograd pack16to1 and deconvolution family, increase simpleomp argv count (#3694)	4 years ago

deconvolution_pack8to16.h	x86 avx512 optimization for convolution winograd pack16to1 and deconvolution family, increase simpleomp argv count (#3694)	4 years ago

deconvolution_pack16.h	x86 avx512 optimization for convolution winograd pack16to1 and deconvolution family, increase simpleomp argv count (#3694)	4 years ago

deconvolution_pack16to1.h	fix build x86 avx512 source with old gcc (#3705)	4 years ago

deconvolution_pack16to4.h	x86 avx512 optimization for convolution winograd pack16to1 and deconvolution family, increase simpleomp argv count (#3694)	4 years ago

deconvolution_pack16to8.h	x86 avx512 optimization for convolution winograd pack16to1 and deconvolution family, increase simpleomp argv count (#3694)	4 years ago

deconvolution_x86.cpp	discard weight memory for x86 arm vulkan (#3865)	4 years ago

deconvolution_x86.h	discard weight memory for x86 arm vulkan (#3865)	4 years ago

deconvolutiondepthwise_x86.cpp	discard weight memory for x86 arm vulkan (#3865)	4 years ago

deconvolutiondepthwise_x86.h	discard weight memory for x86 arm vulkan (#3865)	4 years ago

dequantize_x86.cpp	x86 avx512 optimization (#3581)	4 years ago

dequantize_x86.h	architecture changes for int8 packing (#2771)	5 years ago

dropout_x86.cpp	x86 avx512 optimization (#3581)	4 years ago

dropout_x86.h	X86 Elempack 8 AVX implementations. (#1853)	6 years ago

eltwise_x86.cpp	x86 avx512 optimization (#3691)	4 years ago

eltwise_x86.h	X86 Elempack 8 AVX implementations. (#1853)	6 years ago

flatten_x86.cpp	x86 avx512 optimization (#3581)	4 years ago

flatten_x86.h	architecture changes for int8 packing (#2771)	5 years ago

hardsigmoid_x86.cpp	x86 avx512 optimization (#3581)	4 years ago

hardsigmoid_x86.h	X86 Elempack 8 AVX implementations. (#1853)	6 years ago

hardswish_x86.cpp	x86 avx512 optimization (#3581)	4 years ago

hardswish_x86.h	avx2 infrastructure (#1943)	6 years ago

innerproduct_x86.cpp	discard weight memory for x86 arm vulkan (#3865)	4 years ago

innerproduct_x86.h	discard weight memory for x86 arm vulkan (#3865)	4 years ago

interp_bicubic.h	x86 optimization for interp (#3546)	4 years ago

interp_bicubic_pack4.h	x86 optimization for interp (#3546)	4 years ago

interp_bicubic_pack8.h	x86 optimization for interp (#3546)	4 years ago

interp_bicubic_pack16.h	x86 avx512 optimization (#3581)	4 years ago

interp_bilinear.h	x86 optimization for interp (#3546)	4 years ago

interp_bilinear_pack4.h	x86 optimization for interp (#3546)	4 years ago

interp_bilinear_pack8.h	x86 optimization for interp (#3546)	4 years ago

interp_bilinear_pack16.h	x86 avx512 optimization (#3581)	4 years ago

interp_x86.cpp	x86 avx512 optimization (#3581)	4 years ago

interp_x86.h	x86 optimization for interp (#3546)	4 years ago

lrn_x86.cpp	x86 avx512 optimization (#3581)	4 years ago

lrn_x86.h	Added AVX swish/lrn/batchnorm (#1897)	6 years ago

lstm_x86.cpp	multi-threading rnn/lstm/gru with openmp (#3834)	4 years ago

lstm_x86.h	drop x86 avx2 fp16 (#3568)	4 years ago

mish_x86.cpp	x86 avx512 optimization (#3581)	4 years ago

mish_x86.h	X86 Elempack 8 AVX implementations. (#1853)	6 years ago

packing_x86.cpp	x86 avx512 optimization (#3581)	4 years ago

packing_x86.h	architecture changes for int8 packing (#2771)	5 years ago

padding_pack4.h	x86 sse2 optimization for conv1x1/3x3 pack4 and general sgemm pack4/pack4to1 (#3538)	4 years ago

padding_pack8.h	x86 avx fma optimization (#3543)	4 years ago

padding_pack8_int8.h	test flatten packing padding int8	5 years ago

padding_pack16.h	x86 avx512 optimization (#3581)	4 years ago

padding_x86.cpp	x86 avx512 optimization (#3581)	4 years ago

padding_x86.h	architecture changes for int8 packing (#2771)	5 years ago

pooling_2x2.h	X86 Elempack 8 AVX implementations. (#1853)	6 years ago

pooling_2x2_pack4.h	added a number of optimized sse layers (#3302)	4 years ago

pooling_2x2_pack8.h	X86 Elempack 8 AVX implementations. (#1853)	6 years ago

pooling_2x2_pack16.h	x86 avx512 optimization (#3581)	4 years ago

pooling_3x3_pack4.h	added a number of optimized sse layers (#3302)	4 years ago

pooling_3x3_pack8.h	X86 Elempack 8 AVX implementations. (#1853)	6 years ago

pooling_3x3_pack16.h	x86 avx512 optimization (#3581)	4 years ago

pooling_x86.cpp	x86 avx512 optimization (#3581)	4 years ago

pooling_x86.h	support PyTorch AdaptiveAvgPool2d and AdaptiveMaxPool2d (#2546)	5 years ago

prelu_x86.cpp	x86 avx512 optimization (#3581)	4 years ago

prelu_x86.h	X86 Elempack 8 AVX implementations. (#1853)	6 years ago

quantize_x86.cpp	x86 avx512 optimization (#3581)	4 years ago

quantize_x86.h	architecture changes for int8 packing (#2771)	5 years ago

relu_x86.cpp	x86 avx512 optimization (#3581)	4 years ago

relu_x86.h	architecture changes for int8 packing (#2771)	5 years ago

requantize_x86.cpp	x86 avx512 optimization (#3581)	4 years ago

requantize_x86.h	architecture changes for int8 packing (#2771)	5 years ago

reshape_x86.cpp	x86 avx512 optimization (#3581)	4 years ago

reshape_x86.h	reshape x86 pack4	5 years ago

roialign_x86.cpp	fix roialign_x86.cpp integer multiply may overflow (#2211)	5 years ago

roialign_x86.h	Improve ROIAlign (accelerate ROIAlign, support sampling ratio and aligned ROIAlign) (#1820)	6 years ago

scale_x86.cpp	Edit _bias128 in scale_x86.cpp for useless if (#3821)	4 years ago

scale_x86.h	LSTM arm/x86 + fp16 innerproduct arm (#1881)	6 years ago

sigmoid_x86.cpp	x86 avx512 optimization (#3581)	4 years ago

sigmoid_x86.h	X86 Elempack 8 AVX implementations. (#1853)	6 years ago

slice_x86.cpp	x86 avx512 optimization (#3581)	4 years ago

slice_x86.h	x86 slice pack4	5 years ago

softmax_x86.cpp	pnnx save ncnn bin with fp16 storage (#3715)	4 years ago

softmax_x86.h	x86 sse/avx/avx512 optimization for softmax (#3712)	4 years ago

sse_mathfun.h	Add SSE&AVX optimized for tan (#3765)	4 years ago

swish_x86.cpp	x86 avx512 optimization (#3581)	4 years ago

swish_x86.h	Added AVX swish/lrn/batchnorm (#1897)	6 years ago

tanh_x86.cpp	add tanh avx512 optimize (#3770)	4 years ago

tanh_x86.h	LSTM arm/x86 + fp16 innerproduct arm (#1881)	6 years ago

unaryop_x86.cpp	fix x86 unaryop bug on gcc-4.4 (#3838)	4 years ago

unaryop_x86.h	Add unaryop x86 (#3579)	4 years ago

x86_activation.h	fix build x86 avx512 source with old gcc (#3705)	4 years ago

x86_usability.h	x86 sse/avx/avx512 optimization for softmax (#3712)	4 years ago

yolov3detectionoutput_x86.cpp	style(src/layer/x86): fix build warning (#3699)	4 years ago

yolov3detectionoutput_x86.h	Add yolov3detectionoutput test and AVX optimization (#1994)	5 years ago

No Description

C++ C Python Text Protocol Buffer other

nihuini@tencent.com shuizhuyuanluo@126.com imzhuo@foxmail.com 248857878@qq.com zhiliu6@gmail.com 67086033+ncnnnnn@users.noreply.github.com khj.application@aliyun.com caishanli25@gmail.com e.proydakov@gmail.com 49699333+dependabot[bot]@users.noreply.github.com Timen@users.noreply.github.com 314377460@qq.com 19603468+cmdbug@users.noreply.github.com leo@nullptr.com.cn mingzilaochongtu@gmail.com myhouseng@gmail.com nihui@users.noreply.github.com 774074168@qq.com gemfield@civilnet.cn me@lstlx.com youhanmir@gmail.com ShuangLiu1992@users.noreply.github.com daquexian566@gmail.com ericliu612@gmail.com hansh-sz@hotmail.com

Contributors (25+) All

Contributors (25+)
All