nihui
/
ncnn
Not watched
1
0
Fork 0
Code
Releases 46 Wiki evaluate Activity
Issues 0 Pull Requests 0 Datasets Model Cloudbrain HPC
2296 Commits
5 Branches
354 MB
1 Download
Tree: 525df8bcc5
Branches Tags
${ item.name }
Create branch ${ searchTerm }
from '525df8bcc5'
${ noResults }
Commit Graph

472 Commits (525df8bcc5b63ae009afde83284bb7eb6aefbfeb)

Author SHA1 Message Date
  nihuini 43cf616000
ncnnoptimize embed, fix #3335 4 years ago
  YuShiquan 0b12f5ee07
update readme of quantize (#3283) 4 years ago
  zhiliu6 814f89ef1a
Fuse HardSwish activation into Convolution and InnerProduct (#3233) 4 years ago
  Jimmy 52c6502ae7
Feature/ncnntool/add install (#3212) 4 years ago
  nihui 3d5f447260
fix onnx negative concat axis, fix #3194 4 years ago
  q.yao 41aa601413
fix fuse expand (#3189) 4 years ago
  nihuini a99cc13611
ncnnoptimize generate proper weight for int8 scales, fix non-neon innerproduct int8, fix #3157 4 years ago
  nihuini c5d003001c
preserve lstm hidden output 4 years ago
  nihuini 9afbfdd9dd
ncnnoptimize fuse rsub rdiv with scalar 4 years ago
  nihuini 618f0f70ac
fuse 1x1xc broadcasting add into convolution 4 years ago
  zhiliu6 c6bb70622b
fix convolution padding when pad=0 (#3087) 4 years ago
  lsdustc fd1d5bb2ba
Update ncnn2table.cpp (#3081) 4 years ago
  li mengyang be054aacb0
some misspell (#3046) 4 years ago
  nihui 4f135e07bf
implement convolution1d and pooling1d (#3035) 4 years ago
  zhiliu6 1a9bb708f1
darknet2ncnn: add sam layer support, add letter_box detection. (#3037) 4 years ago
  Jimmy 8fad97cc68
[ncnn2table]do not always resize image (#3023) 5 years ago
  Jimmy def081e845
fix interp miss args (#3021) 5 years ago
  Jimmy e47ab30cf5
fix ncnn2table specification (#3006) 5 years ago
  nihuini 72282f8b1a fix blob count when there are orphaned weight nodes, fix #2969 5 years ago
  nihuini f207001643 recreate layer pipeline before shape inference, fix #2964 5 years ago
  nihui 71bc617a05 better ncnn2table multithreading, print parsed parameters, print progress 5 years ago
  nihuini 312c7372c1 fix norm parsing for long decimal 5 years ago
  nihuini 535f5ed037 fix convolutiondepthwise int8 scale weight saving 5 years ago
  nihuini 1ad37fae1e generate random weight when null bin passed 5 years ago
  nihuini 00b0094b2f copy-pasta makes msvc happy 5 years ago
  nihuini 9a12597dff build ncnn quantization tool even opencv not found 5 years ago
  nihuini 26dc9820e4 custom mlir ncnn optimize pass, add efficientnetv2_b0 benchmark 5 years ago
  nihuini b7959a4fc7 fuse mlir swish and conv-addv2 5 years ago
  nihuini e34edfb672 fix caffe2ncnn input shape for 1-dim and 2-dim blob 5 years ago
  nihuini 34bd5ef161 update eq quant info 5 years ago
  nihuini 72ef77a469 fix build with NCNN_STRING off and NCNN_VULKAN on 5 years ago
  nihuini d021079cd3 implement ncnn2table eq method 5 years ago
  nihui 7d10e5593b fix build 5 years ago
  nihuini 45e548e342 rewrite ncnn2table tool, with better multithreading processing 5 years ago
  nihuini f9e42d4ee4 drop fp16 and int8 mode from caffe2ncnn, ncnnoptimize and ncnn2int8 should be used instead 5 years ago
  nihuini d095149f59 workaround mxnet op non-unique name issue, fix #2855 5 years ago
  nihuini 79927a9eeb update quantized int8 inference doc 5 years ago
  nihuini 31d436c627 more verbose load failure, ncnn2int8 write int8 data properly 5 years ago
  nihuini 56d74580a0 fix uninitialized input and weight 5 years ago
  nihuini 97903bbe85 fix alerts 5 years ago
  nihuini 2b1f8bea25 fix float locale issue in ncnn2mem, fix long line parsing in ncnn2int8 5 years ago
  nihui 7e1aaa5828
cmake option NCNN_INT8 (#2839) 5 years ago
  nihuini 6c41822d35 do not remove optional hidden output for lstm/gru/rnn 5 years ago
  nihuini 5e85f447e6 convert mxnet channel padding, fix some lgtm warnings 5 years ago
  nihui 5fe75f19ef
architecture changes for int8 packing (#2771) 5 years ago
  nihui d4a7abc218 fix onnx2ncnn clip without max blob, fix #2788 5 years ago
  restyled-io[bot] 5f00ba89d2
feat(ncnnoptimize): replace denormals to zero on layers with weights (#2690) 5 years ago
  nihui 19c183cacc fix ncnnoptimize default cut range 5 years ago
  restyled-io[bot] 5e565cfa8a
Restyle add net Cut Function (#2763) 5 years ago
  nihuini 15d63ec0f5 fuse onnx multiheadattention with same qkv blob 5 years ago