NVIDIA/flownet2-pytorch issues and pull requests

#100 - Checkpoint tar file problem

Issue - State: closed - Opened by pawandeep2155 about 6 years ago - 3 comments

#99 - How to test flownet2-pytorch???

Issue - State: open - Opened by jiapei100 about 6 years ago - 9 comments

#98 - error: ‘class at::Context’ has no member named ‘getCurrentCUDAStream…

Pull Request - State: closed - Opened by jiapei100 about 6 years ago - 6 comments

#97 - How to freeze the weights of FlowNet2C if I want to train FlowNet2S?

Issue - State: open - Opened by lc4321 about 6 years ago - 1 comment

#96 - Update README.md

Pull Request - State: closed - Opened by jishnujayakumar about 6 years ago

#95 - ModuleNotFoundError: No module named 'correlation_cuda?

Issue - State: open - Opened by long2double about 6 years ago - 8 comments

#94 - Fix channelnorm variable name bug

Pull Request - State: closed - Opened by adriaangraas about 6 years ago

#93 - Experiment

Issue - State: open - Opened by XDhughie about 6 years ago

#92 - NameError: name 'inference_loader' is not defined

Issue - State: closed - Opened by 3togo about 6 years ago - 3 comments

#91 - NameError: name 'validation_loader' is not defined

Issue - State: closed - Opened by 3togo about 6 years ago - 4 comments

#90 - Optimized correlation forward pass CUDA kernel

Pull Request - State: closed - Opened by mkolod about 6 years ago

#89 - Correlation module does not compile

Issue - State: open - Opened by karipulli about 6 years ago - 1 comment

#88 - Compile Error

Issue - State: closed - Opened by amjltc295 about 6 years ago - 9 comments

#87 - Whether the model can detect subtle movement

Issue - State: open - Opened by taylover-pei about 6 years ago

#86 - error in correlation_forward_cuda_kernel: no kernel image is available for execution on the device

Issue - State: open - Opened by laaRaa about 6 years ago - 35 comments

#85 - How to achieve Correlation1d layer???

Issue - State: closed - Opened by long2double about 6 years ago

#84 - Fix broadcasting for batch size > 1

Pull Request - State: open - Opened by tomrunia about 6 years ago

#83 - RuntimeError when 'inference_batch_size' > 1

Issue - State: open - Opened by tomrunia about 6 years ago - 1 comment

#82 - Thanks for sharing the code. I have one small question here. My system is cuda 8.0 with pytorch 0.4.1 installed. There is no problem when I try to run the inference code. However, when I try to run the training mode, it said an error that channelnorm at line25 of networks/channelnorm_package/channelnorm is not defined. I then checked it and found that it should be channelnorm_cuda and fixed it. Am I supposed to fix it that way or did I ignore something? Anyone has the same issue? Need help

Issue - State: closed - Opened by ChaoningZhang about 6 years ago

#81 - Tensor([[[1., 1., 1., 1.], [1., 1., 1., 1.], [1., 1., 1., 1.]]]]) Tensor([[[[0., 1., 2., 3.], [ 4., 5., 6., 7.], [ 8., 9., 10., 11.]]]]) Data test why the output is 0

Issue - State: closed - Opened by long2double about 6 years ago - 2 comments

#80 - ImportError: /home/dell/.conda/envs/py3/lib/python3.6/site-packages/resample2d_cuda-0.0.0-py3.6-linux-x86_64.egg/resample2d_cuda.cpython-36m-x86_64-linux-gnu.so: undefined symbol: _ZN2at5ErrorC1ENS_14SourceLocationESs

Issue - State: closed - Opened by long2double about 6 years ago - 3 comments

#79 - undefined symbol: _ZN2at5ErrorC1ENS_14SourceLocationESs

Issue - State: closed - Opened by long2double about 6 years ago - 7 comments

#78 - Process finished with exit code 139 (interrupted by signal 11: SIGSEGV)

Issue - State: open - Opened by YingjieYin about 6 years ago - 2 comments

#77 - Make FlowNet2 work with PyTorch 0.4.1

Pull Request - State: closed - Opened by mkolod about 6 years ago

#76 - correlation package can't not support pytorch-0.4.1

Issue - State: closed - Opened by lxtGH over 6 years ago - 7 comments

#75 - Error for 'bash install.sh'

Issue - State: closed - Opened by irwinherrmann over 6 years ago - 2 comments

#74 - Unable to run it for custom dataset.

Issue - State: open - Opened by priyam-nd over 6 years ago

#73 - ImportError:undefined symbol:state

Issue - State: open - Opened by xianyuhong over 6 years ago - 1 comment

#72 - flownet2-fine tune

Issue - State: closed - Opened by ericzw over 6 years ago - 2 comments

#71 - ImportError: _correlation.so: undefined symbol: __cudaPopCallConfiguration

Issue - State: open - Opened by omrysendik over 6 years ago - 4 comments

#70 - edit build.py, and install successfully.

Pull Request - State: open - Opened by xuezhisd over 6 years ago

#69 - pretrained weights

Issue - State: closed - Opened by shanyuhu over 6 years ago - 1 comment

#68 - fix relative import

Pull Request - State: closed - Opened by wandering007 over 6 years ago

#67 - Compile error

Issue - State: open - Opened by kristinarakova over 6 years ago - 2 comments

#66 - Complile error

Issue - State: open - Opened by ericsqxd over 6 years ago

#65 - Strange results with inference

Issue - State: closed - Opened by gbarl over 6 years ago - 2 comments

#64 - ImportError: networks/resample2d_package/_ext/resample2d/_resample2d.so: undefined symbol: state

Issue - State: closed - Opened by HopLee6 over 6 years ago - 1 comment

#63 - Puzzles on the comprehension of cuda codes

Issue - State: open - Opened by kingsleyluoxin over 6 years ago

#62 - FlowNetC does not converge

Issue - State: open - Opened by MatthewD1993 over 6 years ago - 1 comment

#61 - Problem with training Flownet2S

Issue - State: closed - Opened by Meghana4 over 6 years ago - 3 comments

#60 - Why are test images forced to be cropped?

Issue - State: open - Opened by TomHeaven over 6 years ago - 26 comments

#59 - Validation_Epochs

Issue - State: closed - Opened by JensPeter92 over 6 years ago - 2 comments

#58 - How to accelerate the Correlation layer?

Issue - State: open - Opened by PkuRainBow over 6 years ago - 2 comments

#57 - Problem with main inference loop

Issue - State: open - Opened by Oscared over 6 years ago

#56 - Cannot run main.py properly

Issue - State: open - Opened by zengzhi2015 over 6 years ago - 1 comment

#55 - Converted Caffe Pre-trained Models

Issue - State: closed - Opened by xiaoweihappy123 over 6 years ago - 2 comments

#54 - flying chairs dataset

Issue - State: closed - Opened by tongpinmo over 6 years ago - 1 comment

#53 - training and validation problem

Issue - State: open - Opened by tongpinmo over 6 years ago - 6 comments

#52 - run bash install.sh error

Issue - State: closed - Opened by Zhuysheng over 6 years ago - 10 comments

#51 - Running a FlowNet-pytorch on a single image pair

Issue - State: open - Opened by hajarSokeh over 6 years ago - 31 comments
Labels: good first issue

#50 - RuntimeError: Given groups=1, weight[64, 3, 7, 7], so expected input[1, 2, 436, 1024] to have 3 channels, but got 2 channels instead

Issue - State: closed - Opened by mikeleatila over 6 years ago - 1 comment

#49 - make.sh for python3

Pull Request - State: closed - Opened by HTLife over 6 years ago

#48 - added documentation to model.py that better reflects content in FlowN…

Pull Request - State: closed - Opened by loganrooks over 6 years ago

#47 - Size of tensors does not match Inference Step.

Issue - State: closed - Opened by TusharNimbhorkar over 6 years ago - 2 comments

#46 - Python3 version

Issue - State: closed - Opened by HTLife over 6 years ago - 10 comments

#45 - confirm！correlation layer backward is extremely slow

Issue - State: closed - Opened by LiangXu123 over 6 years ago - 4 comments

#44 - multi-gpu training

Issue - State: open - Opened by LiliMeng over 6 years ago - 2 comments

#43 - Torch multiprocessing in docker

Issue - State: closed - Opened by 0asa over 6 years ago - 1 comment

#42 - Models

Issue - State: closed - Opened by coderclear over 6 years ago - 5 comments

#41 - trying to run in inference mode

Issue - State: closed - Opened by bkvie over 6 years ago - 11 comments

#40 - correlation function runs very slow

Issue - State: closed - Opened by sczhou over 6 years ago - 2 comments

#39 - Can I remove the "grad_output.is_contiguous() == True"?

Issue - State: closed - Opened by PkuRainBow over 6 years ago - 4 comments

#38 - L1 loss and EPE become extremely large

Issue - State: closed - Opened by shanyuhu over 6 years ago - 3 comments

#37 - FlowNet2SD single output is wrong

Issue - State: closed - Opened by heurainbow over 6 years ago - 1 comment

#36 - ImportError: /flownet2-pytorch/networks/resample2d_package/_ext/resample2d/_resample2d.so: undefined symbol: _Py_Dealloc

Issue - State: closed - Opened by p31415926 almost 7 years ago - 3 comments

#35 - Modify custom ops according to pytorch recommendations

Pull Request - State: closed - Opened by hellock almost 7 years ago - 3 comments

#34 - Why use only the first loss value for weight updating?

Issue - State: closed - Opened by JiamingSuen almost 7 years ago

#33 - error in Correlation_forward_cuda_kernel: invalid device function

Issue - State: open - Opened by erickwan almost 7 years ago - 2 comments
Labels: good first issue

#32 - Is there any plan to develop 1D correlation on x-axis?

Issue - State: closed - Opened by blackjack201312 almost 7 years ago - 3 comments
Labels: good first issue

#31 - I keep getting error saying the list index out of range

Issue - State: closed - Opened by kimj09 almost 7 years ago - 14 comments
Labels: not reproducible

#30 - main.py: error: argument insta--schedule_lr_frequency: invalid int value: 'MpiSintelClean'

Issue - State: closed - Opened by kimj09 almost 7 years ago

#29 - Fail to adapt to Pytorch Version 0.3

Issue - State: closed - Opened by MatthewD1993 almost 7 years ago - 10 comments
Labels: good first issue

#28 - How to understand the "training for multiple flow" and the "div_flow"

Issue - State: closed - Opened by PkuRainBow almost 7 years ago - 4 comments

#27 - Inconsistent tensor sizes

Issue - State: closed - Opened by xukai8 almost 7 years ago - 1 comment

#26 - Augmentation issues?

Issue - State: open - Opened by PkuRainBow almost 7 years ago - 5 comments
Labels: help wanted, good first issue

#25 - About the Super Huge Training Cost

Issue - State: closed - Opened by PkuRainBow almost 7 years ago - 7 comments

#24 - Should convert the input image from RGB to BGR?

Issue - State: closed - Opened by xukai8 almost 7 years ago - 1 comment

#23 - Can't download the MPI-Sintel flow datasets.

Issue - State: closed - Opened by xukai8 almost 7 years ago - 1 comment

#22 - is cropping implemented correct?

Issue - State: closed - Opened by HaiyongJiang almost 7 years ago - 2 comments

#21 - torch.FatalError: aborting at correlation_cuda.c

Issue - State: closed - Opened by PkuRainBow almost 7 years ago - 5 comments

#20 - performance about flownet2

Issue - State: closed - Opened by HaiyongJiang almost 7 years ago - 1 comment

#19 - Why only consider data[0] and target[0] in the forward computation in FlowNet?

Issue - State: closed - Opened by PkuRainBow almost 7 years ago - 1 comment

#18 - About "Convert Official Caffe Pre-trained Models to PyTorch"

Issue - State: closed - Opened by PkuRainBow almost 7 years ago - 11 comments

#17 - Tips for using for the task of Action detection

Issue - State: closed - Opened by aelnouby almost 7 years ago - 4 comments

#16 - fix offsets in cuda kernel for correlatin map

Pull Request - State: closed - Opened by WarBean almost 7 years ago - 1 comment

#15 - Extremely slow for training Flownet2C

Issue - State: closed - Opened by lliuz almost 7 years ago - 10 comments

#14 - Any natural methods to implement 1D Correlation layer with your 2D correlation

Issue - State: closed - Opened by blackjack201312 almost 7 years ago - 1 comment

#13 - Could you provide the visualization code for the*.flo files~

Issue - State: open - Opened by PkuRainBow almost 7 years ago - 4 comments
Labels: good first issue

#12 - Inference with FlowNetC model

Issue - State: closed - Opened by PkuRainBow almost 7 years ago - 2 comments

#11 - About the MPI-Sintel datasets

Issue - State: closed - Opened by PkuRainBow almost 7 years ago - 8 comments

#10 - Is the resample2d function implemented correctly?

Issue - State: closed - Opened by HaiyongJiang almost 7 years ago - 2 comments

#9 - Performance?

Issue - State: closed - Opened by Yuliang-Zou almost 7 years ago - 4 comments

#8 - Runtime Error: Value cannot be converted to type double without overflow: inf

Issue - State: closed - Opened by sczhou almost 7 years ago - 5 comments

#7 - The training EPE error explodes on FlyingChairs dataset.

Issue - State: closed - Opened by PkuRainBow almost 7 years ago - 8 comments

#6 - forward() takes exactly 2 arguments (3 given)

Issue - State: closed - Opened by PkuRainBow almost 7 years ago - 4 comments

#5 - Confused of your inference mode

Issue - State: closed - Opened by qijiezhao almost 7 years ago - 3 comments

GitHub / NVIDIA/flownet2-pytorch issues and pull requests