Colab pro error Interpolation

[AlexU225]

https://colab.research.google.com/github/AhabbscienceStudioPak/DAIN/blob/master/DAIN_Colab.ipynb#scrollTo=LH7EmLT2gA4l
colab PRO assign GPU
name, driver_version, memory.total [MiB]
Tesla V100-SXM2-16GB, 418.67, 16130 MiB

#interpolation

/content/DAIN
revise the unique id to a random numer 91876
Namespace(SAVED_MODEL=None, alpha=[0.0, 1.0], arg='./model_weights/91876-Thu-Sep-03-17-38/args.txt', batch_size=1, channels=3, ctx_lr_coe=1.0, datasetName='Vimeo_90K_interp', datasetPath='', dataset_split=97, debug=False, depth_lr_coe=0.001, dtype=<class 'torch.cuda.FloatTensor'>, end_frame=137, epsilon=1e-06, factor=0.2, filter_lr_coe=1.0, filter_size=4, flow_lr_coe=0.01, force=False, frame_input_dir='/content/DAIN/input_frames', frame_output_dir='/content/DAIN/output_frames', log='./model_weights/91876-Thu-Sep-03-17-38/log.txt', lr=0.002, netName='DAIN_slowmotion', no_date=False, numEpoch=100, occ_lr_coe=1.0, patience=5, rectify_lr=0.001, save_path='./model_weights/91876-Thu-Sep-03-17-38', save_which=1, seed=1, start_frame=1, time_step=0.2997002997002997, uid=None, use_cuda=True, use_cudnn=1, weight_decay=0, workers=8)
cudnn is used
Interpolate 2 frames
error in correlation_forward_cuda_kernel: no kernel image is available for execution on the device
Warning: Legacy autograd function with non-static forward method is deprecated and will be removed in 1.3. Please use new-style autograd function with static forward method. (Example: https://pytorch.org/docs/stable/autograd.html#torch.autograd.Function) (THPFunction_do_forward at /pytorch/torch/csrc/autograd/python_function.cpp:622)
Traceback (most recent call last):
File "colab_interpolate.py", line 112, in
y_s, offset, filter = model(torch.stack((X0, X1),dim = 0))
File "/usr/local/lib/python3.6/dist-packages/torch/nn/modules/module.py", line 532, in call
result = self.forward(*input, **kwargs)
File "/content/DAIN/networks/DAIN_slowmotion.py", line 148, in forward
self.forward_flownets(self.flownets, cur_offset_input, time_offsets=time_offsets),
File "/content/DAIN/networks/DAIN_slowmotion.py", line 212, in forward_flownets
temp = model(input) # this is a single direction motion results, but not a bidirectional one
File "/usr/local/lib/python3.6/dist-packages/torch/nn/modules/module.py", line 532, in call
result = self.forward(*input, **kwargs)
File "/content/DAIN/PWCNet/PWCNet.py", line 221, in forward
corr6 = self.corr(c16, c26)
File "/usr/local/lib/python3.6/dist-packages/torch/nn/modules/module.py", line 532, in call
result = self.forward(*input, kwargs)
File "/content/DAIN/PWCNet/correlation_package_pytorch1_0/correlation.py", line 59, in forward
result = CorrelationFunction(self.pad_size, self.kernel_size, self.max_displacement,self.stride1, self.stride2, self.corr_multiply)(input1, input2)
File "/content/DAIN/PWCNet/correlation_package_pytorch1_0/correlation.py", line 27, in forward
self.pad_size, self.kernel_size, self.max_displacement,self.stride1, self.stride2, self.corr_multiply)
RuntimeError: CUDA call failed (correlation_forward_cuda at correlation_cuda.cc:80)
frame #0: c10::Error::Error(c10::SourceLocation, std::string const&) + 0x33 (0x7fc469e26193 in /usr/local/lib/python3.6/dist-packages/torch/lib/libc10.so)
frame #1: correlation_forward_cuda(at::Tensor&, at::Tensor&, at::Tensor&, at::Tensor&, at::Tensor&, int, int, int, int, int, int) + 0x628 (0x7fc46625ab38 in /usr/local/lib/python3.6/dist-packages/correlation_cuda-0.0.0-py3.6-linux-x86_64.egg/correlation_cuda.cpython-36m-x86_64-linux-gnu.so)
frame #2: + 0x1bd4a (0x7fc46626ad4a in /usr/local/lib/python3.6/dist-packages/correlation_cuda-0.0.0-py3.6-linux-x86_64.egg/correlation_cuda.cpython-36m-x86_64-linux-gnu.so)
frame #3: + 0x18890 (0x7fc466267890 in /usr/local/lib/python3.6/dist-packages/correlation_cuda-0.0.0-py3.6-linux-x86_64.egg/correlation_cuda.cpython-36m-x86_64-linux-gnu.so)
frame #4: python3() [0x50a7f5]

frame #7: python3() [0x594b01]
frame #9: THPFunction_do_forward(THPFunction, _object) + 0x4ac (0x7fc4b2e37d4c in /usr/local/lib/python3.6/dist-packages/torch/lib/libtorch_python.so)
frame #11: python3() [0x54ac61]
frame #13: python3() [0x50a783]
frame #16: python3() [0x594b01]
frame #19: python3() [0x507f24]
frame #21: python3() [0x594b01]
frame #22: python3() [0x54ac61]
frame #24: python3() [0x50a783]
frame #26: python3() [0x507f24]
frame #28: python3() [0x594b01]
frame #31: python3() [0x507f24]
frame #33: python3() [0x594b01]
frame #34: python3() [0x54ac61]
frame #36: python3() [0x50a783]
frame #38: python3() [0x507f24]
frame #39: python3() [0x509c50]
frame #40: python3() [0x50a64d]
frame #42: python3() [0x507f24]
frame #44: python3() [0x594b01]
frame #47: python3() [0x507f24]
frame #49: python3() [0x594b01]
frame #50: python3() [0x54ac61]
frame #52: python3() [0x50a783]
frame #54: python3() [0x507f24]
frame #56: python3() [0x634dd2]
frame #61: __libc_start_main + 0xe7 (0x7fc4be047b97 in /lib/x86_64-linux-gnu/libc.so.6)

please tell me how to deal with the error?

[tianchengdw]

I have the same problem.

[AlphaGit]

Hi there! You seem to be using an old version of the colab file. I believe that also the repository has changed minor things about the interpolation so if I was in your situation, I'd give it a try with the new version. You can find it here: https://github.com/baowenbo/DAIN/blob/master/Colab_DAIN.ipynb

[AlexU225]

: https://github.com/baowenbo/DAIN/blob/master/Colab_DAIN.ipynb

Using this Colab, an error occurred in the fps detection block,
and Google drive was successfully connected.
sorry for my English, I'm using a translator

cp: cannot stat '/content/gdrive/My Drive//content/gdrive/My': No such file or directory
cp: cannot stat 'Drive/Pexels': No such file or directory
cp: cannot stat 'Videos': No such file or directory
cp: cannot stat '2759484.mp4': No such file or directory

---

CalledProcessError Traceback (most recent call last)

in ()
1 # Detecting FPS of input file.
----> 2 get_ipython().magic('shell yes | cp -f /content/gdrive/My\ Drive/{INPUT_FILEPATH} /content/DAIN/')
3
4 import os
5 filename = os.path.basename(INPUT_FILEPATH)

3 frames

/usr/local/lib/python3.6/dist-packages/google/colab/_system_commands.py in check_returncode(self)
136 if self.returncode:
137 raise subprocess.CalledProcessError(
--> 138 returncode=self.returncode, cmd=self.args, output=self.output)
139
140 def repr_pretty(self, p, cycle): # pylint:disable=unused-argument

CalledProcessError: Command 'yes | cp -f /content/gdrive/My\ Drive//content/gdrive/My Drive/Pexels Videos 2759484.mp4 /content/DAIN/' returned non-zero exit status 1.

[AlphaGit]

@AlexU225 Hi, the error is simply that it's not finding the file path. See the error you got:

cp: cannot stat '/content/gdrive/My Drive//content/gdrive/My': No such file or directory
cp: cannot stat 'Drive/Pexels': No such file or directory
cp: cannot stat 'Videos': No such file or directory
cp: cannot stat '2759484.mp4': No such file or directory

So, in parameters, instead of /content/gdrive/My Drive/Pexels... you should use Pexels...

[AlexU225]

@AlexU225 Hi, the error is simply that it's not finding the file path. See the error you got:

cp: cannot stat '/content/gdrive/My Drive//content/gdrive/My': No such file or directory
cp: cannot stat 'Drive/Pexels': No such file or directory
cp: cannot stat 'Videos': No such file or directory
cp: cannot stat '2759484.mp4': No such file or directory

So, in parameters, instead of /content/gdrive/My Drive/Pexels... you should use Pexels...

Thank you for your advice! But now there is an error in this block

File "/usr/local/lib/python3.6/dist-packages/torch/nn/modules/module.py", line 532, in call
result = self.forward(*input, **kwargs)
File "/usr/local/lib/python3.6/dist-packages/torch/nn/modules/conv.py", line 345, in forward
return self.conv2d_forward(input, self.weight)
File "/usr/local/lib/python3.6/dist-packages/torch/nn/modules/conv.py", line 342, in conv2d_forward
self.padding, self.dilation, self.groups)
RuntimeError: cuDNN error: CUDNN_STATUS_NOT_SUPPORTED. This error may appear if you passed in a non-contiguous input.

[AlphaGit]

Hey @AlexU225 I'm glad you made it that far! Unfortunately, I cannot help you there. That seems like a problem with the image processing itself.

[niuhuojian]

I have same problem with Tesla V100-SXM2-16GB,but P100-PCIE-16GB is work.
RuntimeError: CUDA call failed (correlation_forward_cuda at correlation_cuda.cc:80)
I tried to use new version,but it still happened.
please tell me how to deal with the error?

[mpriessner]

Hello,
I have the same problem as @niuhuojian also with the Tesla V100-SXM2-16GB. Which I am using on Google Colab.
Here I get the following error when running the notebook from: https://github.com/baowenbo/DAIN/blob/master/Colab_DAIN.ipynb

I tried already some things to fix it.
using different combinations of Cuda, gcc and torch versions . (Cuda 9.0, gcc 6.5, torch 1.0.0/Cuda 9.0, gcc 6.5, torch 1.1.0/ Cuda 9.0, gcc 4.8, torch, pytorch 0.4.1 /Cuda 10.0 gcc 7.5, torch 1.4/Cuda 10.1 gcc 7.5, torch 1.6.)
But non of them worked for me.

I also tried to use the solution from from CyFeng16 from issue#44 in but this also seems to stop working.

When I use Cuda 9.0 with gcc-4.8 g++-4.8 which used to work around 4 month ago. This one as well as some of the other combinations gave me the FilterInterpolation Module error from the my_packages folder. see below:

Traceback (most recent call last):
File "train.py", line 15, in
import networks
File "/content/DAIN/networks/init.py", line 1, in
from .DAIN import DAIN
File "/content/DAIN/networks/DAIN.py", line 4, in
from my_package.FilterInterpolation import FilterInterpolationModule
File "/content/DAIN/my_package/FilterInterpolation/init.py", line 1, in
from .FilterInterpolationModule import *
File "/content/DAIN/my_package/FilterInterpolation/FilterInterpolationModule.py", line 6, in
from .FilterInterpolationLayer import FilterInterpolationLayer,WeightLayer, PixelValueLayer,PixelWeightLayer,ReliableWeightLayer
File "/content/DAIN/my_package/FilterInterpolation/FilterInterpolationLayer.py", line 4, in
import filterinterpolation_cuda as my_lib
ModuleNotFoundError: No module named 'filterinterpolation_cuda'

I am slowly running out of ideas to fix that. Does anyone have a working notebook, or an idea what else I could try to do?
That would be great!

[iBobbyTS]

Hi there, I think that's caused by the building process of DAIN packages
ModuleNotFoundError: No module named 'filterinterpolation_cuda'
This means the "filterinterpolation_cuda" package is not installed. Did you run build.sh?
Since you have V100 with the compute compatibility of 7.0, you should uncomment the line
# '-gencode', 'arch=compute_70,code=sm_70',
at DAIN/my_package/compiler_args.py. Then run the build.sh at my_package and PWCNet.
AD. For eaiser installation and usage, refer to iBobbyTS/VFIN, this is kind of like a Video interpolation toolkit, of cause DAIN is in it. I have a colab notebook and you can store the whole built VFIN in drive, every time you only need to extract the files to Colab Runtime and you can start using it.

[semel1]

I have same problem with Tesla V100-SXM2-16GB,
RuntimeError: CUDA call failed (correlation_forward_cuda at correlation_cuda.cc:80).
niuhuojian said that P100-PCIE-16GB works, unfortunately I can't specify which GPU should be used .
The only reason to stack with this version becouse intrigued the ability to specify random output FPS (60fps)

[iBobbyTS]

I have same problem with Tesla V100-SXM2-16GB,

RuntimeError: CUDA call failed (correlation_forward_cuda at correlation_cuda.cc:80).

niuhuojian said that P100-PCIE-16GB works, unfortunately I can't specify which GPU should be used .

The only reason to stack with this version becouse intrigued the ability to specify random output FPS (60fps)

Did you try my suggestions a month ago?

[semel1]

I uncommented '-gencode', 'arch=compute_70,code=sm_70' in the compiler_args.py as you suggested and switched to !pip install torch==1.0.0 torchvision==0.2.1 as TaoTeCha suggested in another post #117 (comment)
The colab is working with a V100 now

[semel1]

Any chance to make Windows binary?

[WilliamJudge94]

I uncommented '-gencode', 'arch=compute_70,code=sm_70' in the compiler_args.py as you suggested and switched to !pip install torch==1.0.0 torchvision==0.2.1 as TaoTeCha suggested in another post #117 (comment)
The colab is working with a V100 now

This is in DAIN/my_package/compiler_args.py