feinanshan/TDNet

Issues

• Download Pre-trained Model Failed.

  #57 opened 9 months ago by YangWulve
  0

• download pretrained backbone - access denied

  #56 opened a year ago by filipberanek
  0

• scale conflict with layer_norm

  #55 opened 2 years ago by lpetflo
  2

• what is the meaning of path_num?

  #54 opened 2 years ago by zhanghaochen-817
  7

• Can you share NYUDv2 dataset with me? Thanks！

  #50 opened 2 years ago by LinZhuoChen
  0

• accuracy cant achieve 79.9 in using td2_fa

  #53 opened 2 years ago by zhanghaochen-817
  0

• runtimeError: If your Function modifies inplace an input that is a view of another Tensor, your Functionannot return more than one Tensor, This is not supported by the current autograd engine, You shouldither make sure the input is not a view (using .clone() for example) or make your Function only returnone Tensor (potentially splitting it into two Functions: one doing the inplace that returns a singleTensor and a second one that does the other operations).You can ask on the forum https: //discuss.pytorch.org/ if you need help to do this change.

  #52 opened 2 years ago by zhanghaochen-817
  1

• train and validate and test

  #20 opened 4 years ago by 872544050
  8

• A question concerning pretrained model

  #49 opened 3 years ago by mskmei
  2

• Where are TD2-Bise34 and TD4-Bise18?

  #30 opened 4 years ago by Alive1024
  1

• Using TDnet on my dataset

  #46 opened 3 years ago by Yulian-Xie
  0

• Do we actually need attention downsampling ?

  #10 opened 5 years ago by theevann
  2

• FileNotFoundError: xxx_gtFine_labelIds.png

  #45 opened 3 years ago by Yulian-Xie
  2

• ninja: build stopped: subcommand failed.

  #43 opened 4 years ago by alhamidi
  0

• ImportError: No module named 'enclib_cpu'

  #41 opened 4 years ago by Yulian-Xie
  1

• cityscapes dataset

  #40 opened 4 years ago by serenityly
  1

• url for downloading resnet101 not working

  #39 opened 4 years ago by serenityly
  2

• The NYUDv2 dataset problem

  #37 opened 4 years ago by dd1ge
  2

• Camvid原始视频数据

  #36 opened 4 years ago by dd1ge
  3

• training with single gpu

  #28 opened 4 years ago by lucci17
  1

• Which Nvidia GPU card do you use for training?

  #27 opened 4 years ago by J-JunChen
  5

• Will GPU always run at 100% while training?

  #26 opened 4 years ago by J-JunChen
  2

• No module named 'enclib_cpu'

  #23 opened 4 years ago by shanjiuvspikaqiu
  8

• About the speed in Table1 & 3

  #32 opened 4 years ago by Jiangtong-Li
  1

• NYUD pretrained weights

  #34 opened 4 years ago by danigr99727
  1

• Why do you use fours paths in the code?

  #35 opened 4 years ago by liqiokkk
  0

• model file for psp101

  #29 opened 4 years ago by PenPen-ThuSilence
  1

• NYUD less training images

  #18 opened 4 years ago by brdav
  1

• leftimg8bit_sequence_trainvaltest.zip

  #25 opened 4 years ago by mZhenz
  1

• how can I get the pretrained resnet-101 for teacher model?

  #24 opened 4 years ago by J-JunChen
  2

• can you help me download the leftImg8bit_allFrames_frankfurt.zip and provide the download url, l cannot download this from https://www.cityscapes-dataset.com/downloads/ thanks!

  #22 opened 4 years ago by hmmlencat
  1

• The config file for the results reported in the paper

  #15 opened 4 years ago by pinzi1989
  1

• how to create TD-Bise34 and TD-Bise18?

  #6 opened 4 years ago by gitunit
  5

• train on my own dataset

  #21 opened 4 years ago by 872544050
  0

• License

  #16 opened 5 years ago by DavideA
  1

• pretrain ResNet not availabel

  #19 opened 5 years ago by Minotaur-CN
  1

• Possible bug in attention mechanism

  #17 opened 5 years ago by DavideA
  1

• Can this code run on SCANNET dataset?How to modify then?

  #14 opened 5 years ago by EchoAmor
  2

• Pretrained model URLs

  #4 opened 5 years ago by MahdiehP
  2

• Per-frame FLOPs

  #12 opened 5 years ago by MahdiehP
  1

• Problem about evaluation.

  #13 opened 5 years ago by Midkey
  1

• ninjia error

  #5 opened 5 years ago by littlestar231
  5

• Is it crucial to use 4 different instances of the segmentation networks?

  #8 opened 5 years ago by ponjoru
  1

• Train on cityscapes original size

  #11 opened 5 years ago by MahdiehP
  3

• Training Runtime Error: value cannot be converted to type float without overflow:

  #9 opened 5 years ago by MahdiehP
  2

• maxpooling问题

  #7 opened 5 years ago by lisc199
  5

• 推理阶段的问题

  #2 opened 5 years ago by finger-monkey
  3

• Question about the PID

  #3 opened 5 years ago by theevann
  1

• when to release the code?

  #1 opened 5 years ago by zimenglan-sysu-512
  1