clovaai/WCT2

Issues

• Create

  #41 opened 2 months ago by Vivekkumar4136
  0

• About training source

  #23 opened 4 years ago by xurong1981
  4

• Without Seg

  #37 opened 2 years ago by ycjing
  1

• About SSIM metircs

  #33 opened 3 years ago by diaodeyi
  0

• some questions about the figure 9

  #20 opened 3 years ago by hello-trouble
  2

• Details about training decoder

  #30 opened 3 years ago by wuqiling97
  0

• How does it work for an indoor dataset?

  #25 opened 4 years ago by RJ2702
  1

• LL components vs. avg. pooling

  #27 opened 4 years ago by felixplum
  2

• QS about the depth map of transfered images

  #22 opened 4 years ago by Jemmagu
  1

• why the encoder has the layer conv0

  #24 opened 4 years ago by YXXX
  1

• How to add temporal consistency

  #21 opened 4 years ago by fantasy-fish
  2

• Does it support large-size photos style transfer?

  #19 opened 4 years ago by zxq2233
  1

• Qs about the Dataset

  #18 opened 4 years ago by EmilyLike
  2

• Question about the loss functions

  #26 opened 4 years ago by ptran1203
  0

• Segmentation maps precision

  #15 opened 4 years ago by Angtrim
  1

• Some questiona about network?

  #12 opened 4 years ago by xuzhongyou
  4

• Query regarding "conv0" layer

  #13 opened 4 years ago by divyagaddipati
  1

• In "cat5" skip network, only the last layer matters

  #17 opened 5 years ago by wanghxcis
  2

• Can I train with my data?

  #11 opened 5 years ago by skh0619
  1

• WCT2 seems cannot transfer styles without segmentation maps? Could you please indicate how to make it works without segmentation maps since you claimed in your paper that it supports transfer directly and without segmentation.

  #10 opened 5 years ago by pkuanjie
  1

• RuntimeError: got 4 channels instead

  #9 opened 5 years ago by galoisgroupcn
  3

• Could you release the pretrained model?

  #8 opened 5 years ago by jianlong-yuan
  1

• Fixed VGG encoder weights query

  #7 opened 5 years ago by SidShenoy
  2

• How do I generate segmentation maps?

  #3 opened 5 years ago by AdarshMJ
  1

• Could you please explain the idea of passing on the information captured by low frequency filters of wavelet pooling to encoder layer while the high frequency components are skip connected to decoder module directly?

  #5 opened 5 years ago by AdarshMJ
  1

• How to generate video stylisation results?

  #4 opened 5 years ago by AdarshMJ
  9

• Error with installing requirements

  #2 opened 5 years ago by pniedzwiedzinski
  1

• trying to port wct transfer to numpy.

  #1 opened 5 years ago by iperov
  5