Frames of varying length

[abhiray92]

HI @HHTseng , the CRNN model is working fine for me, however, the videos in my dataset has varying frame length, I believe you already have the code for this issue as discussed in the link below . Can you please share the code for the same?

#5

[abhiray92]

Hi @HHTseng can you please share the code for varying frame length? Its very important. Thanks

[HHTseng]

yes I had it sometime before, allow me to take a few days to find it

[abhiray92]

Hi @HHTseng Did you get a chance to check for the code?
Thanks

[HHTseng]

Please check out the files in the folder ResNetCRNN_varylength and run UCF101_ResNetCRNN_varlen.py. This necessarily takes longer time to load and run. Let me know if there's any bug spotted, thanks!

[abhiray92]

@HHTseng Thanks a MILLION! I will update you as soon as I check and run the code.

[abhiray92]

I got the below error after running the code -
PS: I have made a few changes to run the model on grayscale images. The mean and std for the transforms parameter has been set to mean=[0.5], std=[0.225].

KeyError: Caught KeyError in DataLoader worker process 0.
Original Traceback (most recent call last):
  File "C:\Users\abhir\.conda\envs\torch\lib\site-packages\torch\utils\data\_utils\worker.py", line 178, in _worker_loop
    data = fetcher.fetch(index)
  File "C:\Users\abhir\.conda\envs\torch\lib\site-packages\torch\utils\data\_utils\fetch.py", line 44, in fetch
    data = [self.dataset[idx] for idx in possibly_batched_index]
  File "C:\Users\abhir\.conda\envs\torch\lib\site-packages\torch\utils\data\_utils\fetch.py", line 44, in <listcomp>
    data = [self.dataset[idx] for idx in possibly_batched_index]
  File "D:\Theses\video-classification\ResNetCRNN_varylength\functions.py", line 51, in __getitem__
    channels = len(self.transform.__dict__['transforms'][2].__dict__['mean'])  # get number of channels from Transformation
KeyError: 'mean'

[abhiray92]

@HHTseng I am constantly getting this error -

KeyError: Caught KeyError in DataLoader worker process 0.
Original Traceback (most recent call last):
  File "C:\Users\abhir\.conda\envs\torch\lib\site-packages\torch\utils\data\_utils\worker.py", line 178, in _worker_loop
    data = fetcher.fetch(index)
  File "C:\Users\abhir\.conda\envs\torch\lib\site-packages\torch\utils\data\_utils\fetch.py", line 44, in fetch
    data = [self.dataset[idx] for idx in possibly_batched_index]
  File "C:\Users\abhir\.conda\envs\torch\lib\site-packages\torch\utils\data\_utils\fetch.py", line 44, in <listcomp>
    data = [self.dataset[idx] for idx in possibly_batched_index]
  File "D:\Theses\video-classification\CRNN\functions.py", line 51, in __getitem__
    channels = len(self.transform.__dict__['transforms'][2].__dict__['mean'])  # get number of channels from Transformation
KeyError: 'mean'

[berna-ylmz]

Did you solve this problem @abhiray92 ?

[abhiray92]

@Berna-lab Yes, my images were in grayscale, so I just set the value of 'Channel = 1'. And it worked.

[berna-ylmz]

Thanks @abhiray92