facebookresearch/OctConv

Issues

• about enlarges the receptive field

  #17 opened 4 years ago by PJJie
  0

• When apply depthwise conv, why discard frequency exchange?

  #16 opened 4 years ago by shoutOutYangJie
  0

• about the pre-trained model of the OctResNet-26 or 18

  #15 opened 4 years ago by louvinci
  0

• How to divide low frequency and high frequency？

  #1 opened 6 years ago by ys-dpc
  2

• is it possible to finetune from original weights?

  #14 opened 5 years ago by yyfcc17
  3

• Which resnet version you use in the paper, pre-active or post-active?

  #9 opened 5 years ago by vMagina
  2

• Oct-I3D

  #12 opened 5 years ago by JinyangGuo
  1

• Octave Transposed Convolution

  #11 opened 5 years ago by makbari7
  0

• I have demonstrated the effectiveness of octave conv on detection tasks.

  #10 opened 5 years ago by Stinky-Tofu
  1

• Hi, I want to re-implement the result on ImageNet. Where can I get the hyperparameters, ie. batch_size.

  #8 opened 5 years ago by vMagina
  1

• octave conv cannot support input at any size

  #7 opened 5 years ago by Stinky-Tofu
  0

• Change Input Size

  #6 opened 5 years ago by VictorGump
  0

• How to use OctConv in mobileNet v2

  #5 opened 5 years ago by shutdownup
  0

• Details about GPU run times in Table 2 of the paper

  #4 opened 5 years ago by peri044
  7

• How to get the first `data_l ` with depth_wise conv?

  #3 opened 5 years ago by PistonY
  1

• Third-party Implementation in PyTorch

  #2 opened 6 years ago by d-li14
  2