wkentaro/pytorch-fcn

Issues

• How to change batch_size？

  #168 opened a year ago by Lukashsx
  1

• Gradient explosion occurs after modifying the weight initialization method

  #167 opened 2 years ago by eadstry
  0

• What h[:, :, 31:31 + x.size()[2], 31:31 + x.size()[3]] mean?

  #161 opened 3 years ago by lyw615
  6

• Running custom datasets

  #152 opened 4 years ago by fredpeertje
  2

• I cannot find a name (osp)

  #166 opened 2 years ago by zsxzs
  1

• 这是怎么处理的？

  #165 opened 3 years ago by Shajiu
  0

• no module named "fcn"

  #163 opened 3 years ago by liul-ing
  1

• fcn.utils.visualize_segmentation,where the visualize_segmentation function?

  #164 opened 3 years ago by liul-ing
  1

• why all my output results of the fcn32 is ZERO

  #113 opened 5 years ago by daiyp3360xxxxxisxxxxxxmine
  16

• I have the same problem, all output results are zeros. Anyone can help ?

  #144 opened 5 years ago by prfans
  5

• Pretrained model 404

  #160 opened 3 years ago by BlackDriver
  4

• Load FCN8 weights

  #145 opened 4 years ago by luistelmocosta
  1

• Hello, I got some problems when I was training.

  #159 opened 3 years ago by dmndxld
  0

• Final layer deconvolutional filters are not fixed

  #155 opened 3 years ago by JJ-res101
  1

• I can't run the train_fcn32s.

  #153 opened 3 years ago by rgd66
  3

• Why the padding of first conv layer in FCN32s is 100?

  #157 opened 4 years ago by msy1412
  3

• requests.exceptions.ConnectionError

  #154 opened 4 years ago by Wenenen
  5

• what's this -1 meaning??

  #156 opened 4 years ago by who1129
  0

• Small question about score_pool3 and score_pool4 in FCN8

  #151 opened 4 years ago by k4kuz0
  1

• feature map size after upsample is 2h or 2(h+1)?

  #141 opened 5 years ago by Lucksong
  3

• I use vgg16 pretrained model to train ,but the result of fcn32s is zero, can you help me?

  #150 opened 4 years ago by hylovezhang
  0

• ValueError: Unexpected module: FCN8sAtOnce

  #123 opened 5 years ago by chentao2016
  3

• excelent work

  #149 opened 4 years ago by lqxisok
  0

• The begining index to crop output images in fcn32s.py

  #147 opened 4 years ago by gniknoil
  1

• Where is torchfcn‘s DOCs?

  #148 opened 4 years ago by LIYHUI
  0

• About download datasets

  #143 opened 5 years ago by BangguWu
  1

• "pip install torchfcn "have error!!--------error: can't copy 'torchfcn/ext/fcn.berkeleyvision.org': doesn't exist or not a regular file

  #127 opened 5 years ago by XGuider
  5

• Why the parameter of bias is set False to nn.ConvTranspose2d() in fcn8s.py?

  #146 opened 4 years ago by songyuc
  0

• padding=100

  #138 opened 5 years ago by LynnHg
  3

• TypeError: expected np.ndarray (got numpy.ndarray)

  #124 opened 5 years ago by Iamyourbaba
  3

• weight returned by get_upsampling_weight?

  #142 opened 5 years ago by Lucksong
  0

• The weight in deconv doesn`t update?

  #140 opened 5 years ago by Lucksong
  3

• CPU memory usage is high

  #139 opened 5 years ago by LiheYoung
  1

• question:can i use my own date to train?

  #126 opened 5 years ago by zhengkai17
  2

• Question about "untransform" method in "VOCClassSegBase" class

  #134 opened 5 years ago by BarryBA
  2

• How to train my own dataset?

  #133 opened 5 years ago by QbaixueQ
  1

• IncompatibleKeys(missing_keys=[], unexpected_keys=[]) for FCN32

  #131 opened 5 years ago by AlexTS1980
  1

• Background problem

  #132 opened 5 years ago by AlexTS1980
  1

• Why didn't you normalize image data?

  #125 opened 5 years ago by sgmath12
  1

• Cannot allocate memory (train my data)

  #130 opened 5 years ago by luostars
  2

• how to test my model？

  #122 opened 5 years ago by jamesbondzhou
  1

• Citing this work

  #120 opened 5 years ago by bhashithe
  2

• Hi,I have a problem in running?

  #115 opened 5 years ago by BCWang93
  1

• Hi,I have a question about the loss?

  #119 opened 5 years ago by BCWang93
  0

• Error while running the training part....TypeError: resize_images() got an unexpected keyword argument 'preserve_aspect_r atio'

  #112 opened 5 years ago by OmorFarukRakib
  1

• Cant not download the prtrained model with 'torchfcn.models.VGG16(pretrained=True)'

  #116 opened 5 years ago by henanjun
  1

• Hi,I have a problem?

  #117 opened 5 years ago by BCWang93
  6

• Hi,I have a question about this line?

  #118 opened 5 years ago by BCWang93
  1

• A question about the transform on the label

  #114 opened 5 years ago by Balabala-Hong
  4

• train for specific tasks

  #111 opened 5 years ago by Fly2flies
  0