nihui/ncnn - ncnn - 开源协同云脑生态支撑系统

Commit Graph

Author	SHA1	Message	Date
nihui	aed05aa851	pnnx fuse more function to module (#4351 ) * pnnx fuse more function to module * rename some pass name * fuse adjacent reshape, fuse pad conv2d * fuse pad conv1d	3 years ago
nihui	ec1b07c9fe	pnnx fp16 option for ncnn and onnx weight type (#4350 )	3 years ago
nihui	6967baaccc	pnnx convert torch bitwise left_shift right_shift (#4349 )	3 years ago
nihui	eceac35a7f	implement MultiheadAttention kdim vdim (#4347 )	3 years ago
nihui	498ca7341b	squeeze and expanddims 4d (#4346 )	3 years ago
Lry89757	6a47f8d15c	gridsample op support (#4288 ) Co-authored-by: LRY89757 <LRY89757@users.noreply.github.com> Co-authored-by: nihuini <nihuini@tencent.com> Co-authored-by: nihui <shuizhuyuanluo@126.com>	3 years ago
nihui	a2af6369d9	match inplace slice copy pattern, rewrite copy uses (#4338 )	3 years ago
nihui	a7e3c62a1b	save foldable constants in file for reducing memory usage (#4337 )	3 years ago
nihui	cb88e16fdf	pnnx save onnx zero (#4077 )	3 years ago
WuJinxuan	abb28435d6	fix:pnnx-softmax (#4333 )	3 years ago
nihui	92da26be79	pnnx load gpu torchscript and reset device (#4330 )	3 years ago
nihui	5b28c1730e	implement ncnn fold and unfold (#4326 )	3 years ago
nihui	a12c24d328	pnnx convert fold unfold (#4325 )	3 years ago
nihui	b8d40a960f	pnnx convert nn.Softmax2d (#4324 )	3 years ago
nihui	bcf06bd1c0	fold new_full and full_like (#4323 )	3 years ago
nihui	0f9a3bb9dc	fix slice end index, fix fp16 model weight alignment (#4317 )	3 years ago
nihui	6e49fa30dc	groupnorm 1d/2d/4d (#4312 )	3 years ago
nihui	fdf129fcd7	pnnx pytorch 1.13 (#4314 )	3 years ago
nihui	8edc03c515	pnnx skip dynamic size evaluation (#4291 )	3 years ago
nihui	f770987267	pnnx glu batchindex aware conversion (#4285 )	3 years ago
nihui	777e4ef0a0	fix pnnx softmax/normalize/slice negative axis conversion to ncnn (#4284 )	3 years ago
Fangjun Kuang	5281d51535	implement GLU and pnnx conversion (#4283 )	3 years ago
nihui	77eda4c19f	implement lstm proj_size (#4263 )	3 years ago
LiuYi-Up	902954d737	modified the param axes of expanddims in modelwriter (#4259 )	3 years ago
nihui	cef95f64a4	pnnx math operation (#4251 )	3 years ago
Zhang Ge	3fce00b260	fix param parsing issue when layer/blob name exceeds 255 (#4236 ) * fix param parsing issue when layer/blob name exceeds 255 * apply code-format changes Co-authored-by: ZhangGe6 <ZhangGe6@users.noreply.github.com>	3 years ago
miemie2013	720f3c9aab	Add DeformableConv2D (#4070 ) * Add DeformableConv2D * add unittest and docs * pnnx torchvision deformconv2d conversion Co-authored-by: miemie2013 <miemie2013@users.noreply.github.com> Co-authored-by: nihui <shuizhuyuanluo@126.com>	3 years ago
nihui	402001128d	eliminate math expression with torch.zeros/ones input (#4083 )	3 years ago
nihui	4f414c1806	implement 4d memorydata (#4074 ) * implement 4d memorydata * fix ncnnoptimize memorydata 4d	3 years ago
nihuini	e33c85ca82	fix pnnx crash on dynamic pads, do not eliminate upsample for dynamic shape	3 years ago
nihui	1d0917c83b	fix build with very old gcc (#4048 ) * clear bom marker, avoid vector data function	3 years ago
nihui	9b39691cc8	pnnx handle unrecognized file format (#4028 )	3 years ago
nihui	322667a2ab	pnnx fix fused tensor_split operator insert order (#4006 )	3 years ago
nihui	1892c25360	pnnx fuse megvii style shufflechannel slice (#3999 )	3 years ago
nihui	94786308bd	pnnx fuse binaryop eltwise as weighted sum (#4000 )	3 years ago
nihui	d4a704de0e	pnnx eliminate noop upsample (#3991 )	3 years ago
nihui	b4bae2c9e4	pnnx convert torch.tensor_split, fuse full dim size slice to tensor_split (#3988 )	3 years ago
nihui	044467b2c6	pnnx support torch 1.12 (#3981 )	3 years ago
nihui	531506d602	improve pattern value match, always treat inplace operator as non-inplace version (#3970 )	3 years ago
nihui	8dbedf8a19	use cmake gnuinstalldirs for install destination (#3968 )	3 years ago
moozae	94ec06a8a5	delete unused variables (#3965 )	3 years ago
nihui	f597619e94	pnnx export weight inside moduleop (#3902 )	3 years ago
nihui	5ae827c745	convert inplace relu6, match more hardswish pattern (#3952 ) * convert inplace relu6, match more hardswish pattern * upgrade requests	3 years ago
nihui	bc8e939c85	insert reshape for nn.Linear with 4d/5d inputs (#3959 ) * insert reshape for nn.Linear with 4d/5d inputs * batch index aware reshape, add test	3 years ago
nihui	08e2dd5a0c	pnnx ncnn fp16 storage conversion as a pass (#3939 )	3 years ago
Evgeny Proydakov	1c8f1ba7c6	Fixed LGTM warnings (#3891 ) src/stb_image.h src/layer/detectionoutput.cpp tools/quantize/imreadwrite.cpp	4 years ago
nihui	5400028e12	fix ncnn2table crash in lightmode (#3898 ) * fix crash in ncnn2table for missing weight data in lightmode * Update modelwriter.h	4 years ago
nihui	b8caeadee3	fix crash in modelwriter	4 years ago
nihui	612806a919	pnnx convert torch fft family (#3892 ) * disable mkl avx/avx2 optimization for potential memory issue in dft	4 years ago
nihui	53fa52e78a	pnnx support complex data type, convert torch complex/imag/real/fft.rfft2/fft.irfft2 (#3890 )	4 years ago

1 2 3 4 5 ...

605 Commits (057b5bb515d551fa64decdb7350422c19feba447)