yonghenglh6/DepthwiseConvolution

Issues

• windows caffe编译

  #24 opened 4 years ago by rlczddl
  0

• just a copy of conv

  #35 opened 5 years ago by igurantee
  1

• affter add the 3 piece into caffe file, when make caffe, crt1.o error happen, how to solve it?

  #34 opened 6 years ago by PiseyYou
  0

• only for dalition 1?

  #33 opened 6 years ago by ys198918
  0

• Which GPU did you run the test with?

  #32 opened 6 years ago by AlexofNTHU
  0

• Did somesone train some models with this DepthwiseConvolution layer?

  #19 opened 7 years ago by Ai-is-light
  6

• 与convd的不同

  #31 opened 6 years ago by wzj5133329
  6

• Caffe (Windows) - version, build, usage?

  #30 opened 6 years ago by Edison0902
  0

• About group in deploy.prototxt

  #29 opened 6 years ago by wang5566
  0

• Movidius ncsdk not support this

  #26 opened 6 years ago by qingshutie
  1

• Except for the speed, does it perform like the original caffe convolution?

  #28 opened 6 years ago by jiangyangzhou
  1

• why your deploy not have batch_norm_param { use_global_stats: true}???

  #21 opened 6 years ago by Ai-is-light
  1

• only for stride 1？

  #25 opened 6 years ago by OMG59E
  1

• Unknown layer type: DepthwiseConvolution

  #27 opened 6 years ago by ShawKai666
  0

• Why slower than tensorflow

  #16 opened 7 years ago by twmht
  1

• Dose the restriction 'only for stride1' still hold?

  #20 opened 7 years ago by kfxw
  2

• which vison of caffe did you use? When I use it on my caffe ,it got wrong.

  #23 opened 7 years ago by daixiaogang
  0

• is this implementation support cudnn？

  #22 opened 7 years ago by lvhao7896
  1

• how to change the two super-parameters ?And, trian time

  #17 opened 7 years ago by Ai-is-light
  11

• error == cudaSuccess (77 vs. 0 ) an illegal memory access was encountered

  #18 opened 7 years ago by Ai-is-light
  4

• train slowly

  #11 opened 7 years ago by wander2017
  12

• Number of parameters not reduced

  #12 opened 7 years ago by zhaokai5
  1

• About group number error

  #14 opened 7 years ago by rickchen147258
  5

• depthwise layer is the same with original group convolution？

  #15 opened 7 years ago by lbwang2006
  1

• about some details

  #13 opened 7 years ago by dlyldxwl
  5

• Do you have implemented the "cpp" file of DepthwiseConvolution for CPU?

  #4 opened 8 years ago by csyking
  10

• Unknow layer type: DepwiseConvolution

  #10 opened 7 years ago by Derekglk
  4

• the solver.prototxt

  #9 opened 7 years ago by andeyeluguo
  1

• Unknown layer type: DepthwiseConvolution

  #7 opened 7 years ago by firefox1031
  4

• Check failed: registry.count(type) == 1 (0 vs. 1) Unknown layer type: DepthwiseConvolution

  #8 opened 7 years ago by walegahaha
  3

• Does the DepthwiseConvolutionLayer need the parameter(group) ?

  #6 opened 7 years ago by birdwcp
  1

• group convoltion

  #5 opened 7 years ago by camel007
  1

• How about proto file changes?

  #3 opened 8 years ago by mychina75
  1

• About the benchmark device

  #2 opened 8 years ago by austingg
  5

• Is this different with the original caffe's conv_layer.cpp?

  #1 opened 8 years ago by lyk125
  1