nihui/ncnn - ncnn - 开源协同云脑生态支撑系统

Commit Graph

Author	SHA1	Message	Date
nihuini	6c2cee8186	fix mat clone with atypical source cstep	4 years ago
nihuini	a490f8a533	fix layernorm with affine	4 years ago
Tijmen Verhulsdonck	ac5dc23ccc	added a number of optimized sse layers (#3302 ) * added a number of optimized sse layers, specifically to increase performance of mobilenet style networks	4 years ago
zhiliu6	a08f700775	Optimize avx convolution activation (#3299 ) * use general fmadd * forceline x86 fmadd for better performance * fix msvc compile warning * simplify swish implementation * Use activation layer for better performance * Optimize x86 ConvolutionDepthWise activation	4 years ago
Tijmen Verhulsdonck	e3aa893dfb	move custom_layer_to_index to public (#3294 )	4 years ago
nihui	aa9753b2f0	detach mat from local blob allocator so net instance could be destroyed much earlier (#3287 )	4 years ago
zhiliu6	814f89ef1a	Fuse HardSwish activation into Convolution and InnerProduct (#3233 ) * add general fused activation * add NCNN_FORCE_INLINE option	4 years ago
nihui	4313d23355	update ci swiftshader 20211002 (#2366 ) * build macos vulkan * drop rt * load swiftshader * workaround swiftshader fltmax fp16 nan Co-authored-by: zhuo@mbp <imzhuo@foxmail.com>	4 years ago
Tijmen Verhulsdonck	4270b5c502	Fix broken codepaths with AVX only (#3254 ) * Fix codepaths for fp16 weights when only AVX is enabled * Disable opt overrides * Update SDK url * Update vulkan SDK download version * Debugging risv pad * apply code-format changes * fix padding test * fix mips slice test * fix lrn test * implement mish swish image shader, fix pooling adaptive image storage support, drop debug output * update ci ubuntu 18.04 Co-authored-by: nihui <shuizhuyuanluo@126.com>	4 years ago
Zhuo Zhang	492297d2f6	add A15 and M1 macro definitions (#3263 )	4 years ago
nihui	57ad2c138c	fix build on c906, fix #3230	4 years ago
nihui	ceec22cd46	Update convolution_riscv.cpp (#3050 )	4 years ago
Xavier Hsinyuan	a2f89e7392	RVV fp16/fp32 optimized Dropout, GRU and Softmax (#3200 ) * RVV optimzied DropOut * RVV optimized GRU, fp32 * RVV optimized GRU, fp16 * RVV optimzed Softmax	4 years ago
TianZer	c44a6c7f47	Remove two potential warnings for VisualStudio (#3188 )	4 years ago
nihui	5a35c2b11e	load model memory by reference (#3179 )	4 years ago
nihui	e9b5bbcd2d	fix armv7 roundmode (#3176 )	4 years ago
yaobyPerfxlab	ec561736a5	Riscv64 c906 d1 (#3159 ) * Use RVV spec 0.7.1 for C906. * Fix code style issue. * Update convolution_sgemm_packn_fp16s.h RVV_SPEC_0_7 update * apply code-format changes Co-authored-by: Zhang Xianyi <xianyi@perfxlab.com> Co-authored-by: yaobyPerfxlab <yaobyPerfxlab@users.noreply.github.com>	4 years ago
nihui	51652a2280	NCNN_MALLOC_OVERREAD for potential kernel reading data out of allocation size (#3155 )	4 years ago
nihuini	a99cc13611	ncnnoptimize generate proper weight for int8 scales, fix non-neon innerproduct int8, fix #3157	4 years ago
nihui	c6cda8d07c	arm neon optimization for requantize leakyrelu (#3144 ) * arm neon optimization for requantize leakyrelu * add missing changes * Update test_requantize.cpp * more test coverage	4 years ago
nihuini	169614f732	fix build with NCNN_STDIO off	5 years ago
nihui	da7b64b833	fix build on c906	5 years ago
nihuini	11794675f3	apple a11 and a12 do not support armv8.2 dotprod, restore the fp16-only optimized path	5 years ago
nihuini	b9460c6e8f	fix armv7 requantize relu, fix #3122	5 years ago
Xavier Hsinyuan	2a5c672787	Add unittest and RVV optimized for SELU (#3114 )	5 years ago
Xavier Hsinyuan	d78add2acd	RVV optimized PReLU, with fp16 support (#3113 )	5 years ago
nihui	f9a16ea1ec	fix build on c906	5 years ago
Xavier Hsinyuan	9933cc776a	RVV optimized HardSwish and HardSigmoid (#3108 ) * RVV optimized HardSwish, with fp16 support * RVV optimized HardSigmoid, with fp16 support * apply code-format changes Co-authored-by: thelastlin <thelastlin@users.noreply.github.com>	5 years ago
nihui	d91cccfb55	apply code-format changes	5 years ago
Xavier Hsinyuan	99440e67f7	RVV optimized binaryop, with fp16 support (#3097 )	5 years ago
nihui	2c4ae09604	fix #2961 (#3095 )	5 years ago
nihuini	affbefe311	some space cleanup, blob clone from allocator	5 years ago
lsdustc	61af40cfbd	Fix Innerproduct gemm forgot to add an offset of w in int8 forward (#3084 )	5 years ago
zhiliu6	190207a173	fix AVX2 compiled as AVX problem (#3090 )	5 years ago
nihui	052b2a1653	apply code-format changes	5 years ago
chenxiemin	9bd8a50b96	fix vulkan memory leak issue (#3088 )	5 years ago
nihui	cdf45a6512	cmake option NCNN_BF16 (#3068 )	5 years ago
Tijmen Verhulsdonck	eaa7e24db6	Added ability to switch AVX/AVX2 during runtime (#3076 )	5 years ago
nihui	9391fae741	optimize arm neon tanh (#3079 )	5 years ago
deepage	5841f667a6	optimize arm neon tanh fma (#3078 ) Co-authored-by: chao.tang <chao.tang@lynxi.com>	5 years ago
nihui	06a7086daa	optimize arm neon exp log sincos fma (#3077 )	5 years ago
nihui	b413fd3a3d	auto code-format bot and disable restyled (#3075 )	5 years ago
Tijmen Verhulsdonck	a7f301a99d	Add clang compatiblity (#3071 ) * Add clang compatiblity Add ability to build NCNN lib on windows with clang GNU * Restyled/pull 3071 (#16) * [skip ci] Restyled by clang-format * [skip ci] Restyled by astyle * [skip ci] Restyled by clang-format * [skip ci] Restyled by astyle Co-authored-by: Restyled.io <commits@restyled.io> Co-authored-by: Restyled.io <commits@restyled.io>	5 years ago
Zhuo Zhang	5049aaeabd	fix invalid read for yuv420sp2rgb on arm32 (#3066 )	5 years ago
nihui	927e34278c	mips msa optimization part2 (#3063 ) * mips msa optimization for convolution 3x3 pack1to4, 7x7s2 pack1to4, dropout, eltwise, hardsigmoid, hardswish, packing, flatten, prelu * prefetch convdw3x3 and convdw5x5 * fix mips convolution sgemm pack4to1 * more prefetch helps	5 years ago
nihui	49cda73420	mips msa optimization for convolution sgemm and winograd (#3055 ) * mips msa optimization for convolution sgemm and 3x3s1 winograd, use msa fmadd * mips msa optimization for convolution sgemm pack4to1 * mips msa optimization for swish, improve sgemm kernel * unroll 12x4, use prefetch	5 years ago
nihui	8a40b98348	mips optimization (#3051 ) * mips msa optimization for crop deconvolution deconvolutiondepthwise general cases * mips msa activation support pack4 * mips msa optimization for concat slice interp * mips msa optimization for binaryop unaryop	5 years ago
nihui	90b38ba145	fix mips convolutiondepthwise	5 years ago
nihui	a133f4be92	mips msa optimization for pooling general cases, fix test on ls2k	5 years ago
nihui	1f49fc4b67	mips msa optimization for padding packing flatten innerproduct convolution convolutiondepthwise general cases	5 years ago

1 2 3 4 5 ...

1380 Commits (6c2cee818668bc19ce3fbcd68e3964283050554b)