sampepose/flownet2-tf

Issues

• Questions about compiling

  #61 opened 6 years ago
  3

• About the correlation layer

  #60 opened 6 years ago
  1

• Data augmentation on CPU？

  #59 opened 6 years ago
  2

• I have succeeded compiling correlation.so. However, I got the zero output when testing

  #57 opened 6 years ago
  0

• the shape of the flow file

  #56 opened 6 years ago
  0

• Compiling problem: invalid redeclaration of type name & cannot overload functions distinguished by return type alone

  #55 opened 6 years ago
  3

• so many bugs while compiling

  #54 opened 6 years ago
  0

• Maintainer Update

  #53 opened 6 years ago
  2

• out of memory error

  #52 opened 6 years ago
  1

• fatal error: nsync_cv.h: No such file or directory

  #51 opened 6 years ago
  3

• ImportError: cannot import name FlowNetC

  #50 opened 6 years ago
  0

• from ..net import Mode ValueError: Attempted relative import in non-package

  #49 opened 6 years ago
  2

• Unable to do make with Tensorflow1.2.1, Python 2.7, Cuda 8.0, CuDNN 6.0

  #48 opened 6 years ago
  1

• convert image data to tfrecord file

  #47 opened 6 years ago
  1

• LossTensor is inf or nan

  #46 opened 6 years ago
  0

• make all error

  #45 opened 6 years ago
  1

• OutOfRangeError (see above for traceback): Read less bytes than requested

  #44 opened 6 years ago
  0

• undefined symbol: _ZN10tensorflow16CorrelationGrad

  #43 opened 6 years ago
  16

• nvcc fatal : Unknown option 'std'

  #42 opened 6 years ago
  1

• Solution to " undefined symbol: _ZN10tensorflow8 .... "

  #41 opened 6 years ago
  5

• correlation.so : Undefined symbols (after succesfull compilation)

  #40 opened 6 years ago
  3

• Estimated flow boarder is showing abnormal results

  #39 opened 6 years ago
  0

• Data conversion code

  #38 opened 6 years ago
  1

• Segmentation fault (core dumped)

  #37 opened 7 years ago
  5

• weights and loss become nan in the training

  #36 opened 7 years ago
  1

• Input preprocessing to use pretrained weights

  #35 opened 7 years ago
  1

• The edge of the image generated by flow warp is burr

  #34 opened 7 years ago
  0

• Is there a good method or labeling tool to create ground truth data to fine tune flownet2 on new video datasets?

  #33 opened 7 years ago
  1

• Train only one batch?

  #32 opened 7 years ago
  1

• How to convert the result to Gray Scale Image

  #31 opened 7 years ago
  13

• Convergence problem with augmentation

  #30 opened 7 years ago
  1

• fine-tune

  #29 opened 7 years ago
  1

• Issue compiling on Ubuntu and tensorflow 1.4

  #28 opened 7 years ago
  24

• test a batch

  #27 opened 7 years ago
  2

• checkpoints corrupted in download file?

  #26 opened 7 years ago
  0

• deal with different resolution pictures as input

  #25 opened 7 years ago
  5

• Help compiling on Windows

  #24 opened 7 years ago
  6

• TypeError: __init__() got an unexpected keyword argument 'reader_kwargs'

  #23 opened 7 years ago
  1

• Training data used for pre-trained model.

  #22 opened 7 years ago
  13

• Corrupted OF output

  #21 opened 7 years ago
  2

• Small mistake?

  #20 opened 7 years ago
  1

• Missing 1-th output from gradients_5/FlowNet2/FlowWarp_grad/FlowWarpGrad

  #19 opened 7 years ago
  5

• nsync_cv.h

  #18 opened 7 years ago
  3

• InvalidArgumentError: No OpKernel was registered to support Op 'FlowWarp' with these attrs.

  #17 opened 7 years ago
  5

• unmatched shapes in tf.concat()

  #16 opened 7 years ago
  4

• process killed during training

  #15 opened 7 years ago
  0

• Problem with data augmentation

  #14 opened 7 years ago
  5

• Question about porting from caffe to Tensorflow

  #13 opened 7 years ago
  2

• Fatal error: cuda/cuda_config.h

  #12 opened 7 years ago
  5

• Intermedia npy file

  #11 opened 7 years ago
  2