sniklaus/pytorch-liteflownet

Issues

• Train code

  #53 opened 4 months ago by saigopalm
  1

• Would you please provide the pretrained model? Thanks

  #52 opened 5 months ago by Xt117
  1

• AttributeError: module 'correlation' has no attribute 'FunctionCorrelation'

  #51 opened 2 years ago by kwu14victor
  5

• is there any way to infer on cpu?

  #50 opened 2 years ago by HSILA
  1

• run liteflownet on a sequence of pairs (aka video) offline/real time.

  #49 opened 3 years ago by ilinii
  1

• How does the function "backwarp" work?

  #48 opened 3 years ago by sherlocksnowin
  1

• 'NoneType' object has no attribute 'get_function'

  #47 opened 3 years ago by sjtuytc
  1

• Correlation function

  #26 opened 5 years ago by ShrutheeshIR
  3

• batch processing

  #46 opened 3 years ago by rose-jinyang
  3

• 网络模型文件

  #45 opened 4 years ago by dawei03896
  3

• what‘s pretrained models name?

  #44 opened 4 years ago by WangPeng19981204
  1

• In place operation error when backpropagating

  #42 opened 4 years ago by surajsy
  1

• Question for liteflownet

  #41 opened 4 years ago by XQLuck
  1

• EPE on Sintel clean

  #38 opened 4 years ago by Queenyy
  2

• Custom images of different sizes

  #40 opened 4 years ago by BonJovi1
  2

• Greyscale Images

  #39 opened 4 years ago by andreitapch
  1

• Real-time inference?

  #37 opened 4 years ago by FedericoVasile1
  1

• Can this be extended to multi-GPU inference?

  #36 opened 5 years ago by mailvijayasingh
  2

• liteflownet2 pytorch implementation

  #35 opened 5 years ago by steve13durant
  1

• dou

  #34 opened 5 years ago by acse-wz19
  3

• What are the differences between these 3 pre-trained models？

  #33 opened 5 years ago by DonaldKam
  1

• The out.flo might be wrong

  #32 opened 5 years ago by steve13durant
  5

• Need guidance regarding writing the training code

  #31 opened 5 years ago by MFarooqAit
  4

• PyTorch 1.1.0

  #30 opened 5 years ago by ncherel
  1

• NameError: name 'Stream' is not defined

  #29 opened 5 years ago by mdyao
  1

• Check for pytorch >= 0.4.1 fails for pytorch built from git

  #28 opened 5 years ago by abrock
  1

• Training on FlyingChairs

  #27 opened 5 years ago by julien-mille
  2

• How can I change the correlation maxdistance from 7 to 9?

  #24 opened 5 years ago by CXMANDTXW
  1

• I cannot download the pretrained model

  #25 opened 5 years ago by theodoruszq
  1

• Help in modification of estimate () to deal batches

  #23 opened 5 years ago by MFarooqAit
  1

• Data augmentation using mean value

  #22 opened 5 years ago by abrosua
  4

• AttributeError: module 'correlation' has no attribute 'FunctionCorrelation'

  #21 opened 5 years ago by MSLAwan
  2

• Converting pre-trained models from Caffe to PyTorch

  #19 opened 5 years ago by abrosua
  2

• How to see output of out.flo?

  #20 opened 5 years ago by boheyse
  1

• Could you please explain the meaning of the number you minus（0.411618, etc)？

  #18 opened 5 years ago by Le1Elayne
  1

• What is the exact meaning of argument?

  #17 opened 5 years ago by xuantengh
  1

• Question about multiplying 20 to the final tensorFlow

  #16 opened 5 years ago by codeslake
  1

• NaN after the first iteration of training.

  #15 opened 5 years ago by codeslake
  7

• Testing in new image pair

  #14 opened 5 years ago by Yihan-Zhang
  1

• Question on flow normalizing

  #13 opened 5 years ago
  6

• Doubt about Backward Function

  #11 opened 6 years ago by arc130
  1

• It seems to not has a trainning implementation in the repo?

  #12 opened 6 years ago by PeiqinSun
  1

• loss don't declined

  #10 opened 6 years ago by guoyanzi
  1

• Why is the resolution of the output of the optical flow inferred with liteflownet half of the original image pair?

  #9 opened 6 years ago by wuwenhuan
  2

• Is Backward implement warp operation

  #8 opened 6 years ago by CXMANDTXW
  1

• why input_tensor need to minus 0.4XX

  #7 opened 6 years ago by CXMANDTXW
  1

• What is dblBackward?

  #6 opened 6 years ago by wuwenhuan
  2

• _pickle.UnpicklingError: invalid load key, '<'.

  #5 opened 6 years ago by LJNDDUP
  1

• hello how ca i use this code to train a new datasets ?

  #4 opened 6 years ago by fikri2992
  2

• It seems the lite-flownet work poor on gray images like UCSD

  #3 opened 6 years ago by fjchange
  2