JanuszL
commented
4 years ago Hi, Thank you for all questions. Going one by one:
Is there a limit for the number of videos to be opened by a VideoReader? Obviously 100 frames of 4000 videos cannot fit on a GPU memory, but I have imagined that each video would be loaded singularly only at the next() call of the DALIGenericIterator, so the frames would be loaded only when needed. Am I wrong?
DALI keeps all video files open. There is an OS limit how many of them can be open simultaneously - you can increase it as in this answer https://github.com/NVIDIA/DALI/issues/1350#issuecomment-539521606 @a-sansanwal - I think we need to limit the number of the video files open at once in DALI and close files above that limit
Moreover, for taking 100 frames of each video, is it right to have batch_size=1 and seq_length=100
That is correct.
RuntimeError: Critical error in pipeline: [/opt/dali/dali/operators/fused/crop_mirror_normalize.h:155] Assert on "output_layout_.is_permutation_of(input_layout_)" failed: The requested output layout is not a permutation of input layout.Is it right to have the CropMirrorNormalize working on the input[0] element? I expect that element to be the 100 frames batch tensor, with the input[1] element being the label instead. Am I guessing right? Is something wrong in my code or in the way I am using the CropMirrorNormalize operation?
It should not be any problem running it on the input[0]. @jantonguirao can you check why "output_layout_.is_permutation_of(input_layout_ the error appears (I'm also not sure if we can crop over z-axis in this case).
his time the code runs with no error, but it seems to "stop" after loading 3 videos only. The terminal stayed "freezed" for several minutes, and I had to kill the process. So, I am wondering
DALI doesn't support videos with VFR (variable frame rate) and the user may experience this kind of hangs. DALI has some heuristics to detect this kind of input and warn the user but it is not 100% accurate. You may want to update DALI to some nightly build and check if this is still the case (maybe there some other issues we have fixed already). Also if you can narrow down this problem to a particular video and share it with use we can verify what is the root cause.
I hope that my post is comprehensible and I apologize in advance for asking maybe too many non-related questions altogether, but I could not find any answer in the docs or in other issues here on GitHub.
We are happy to help.
CrohnEngineer
jantonguirao
a-sansanwal
kavita19
0Rutuja28-97
Hi everybody,
I'm opening an issue since I am encountering several problems writing a Pipeline for loading video files. I'm not sure of wheter DALI is the best tool for my task and neither if I am using it properly, so I would start by first explaining my goal. I have a very huge dataset consisting of hundreds of thousands of videos, and I would like to use DALI's VideoReader to build a PyTorch DataLoader since, according to the documentation, DALI's VideoReader uses
NVIDIA GPU’s hardware-accelerated video decoding, so I would like to speed-up and eventually parallelize the training of a CNN using the GPU for the data loading operations. I took the Video Super Resolution example (https://docs.nvidia.com/deeplearning/sdk/dali-developer-guide/docs/examples/video/superres_pytorch/README.html) and wrote my personal DALILoader as it follows:I created a
file_list.csvwhere I have written all the paths and the labels of the videos (my task is a simple binary classification), and this simple test script:I want simply to load 100 frames of each video, then crop them randomly in the height and width dimension. As a first test, I didn't want to use the whole dataset, so I used just a portion of it (we are talking about 4000/5000 videos in any case), but when I run the code I have encountered three major errors. I report them in "discovery order", as after I have encountered the first one I have simplified my code reducing the task complexity too for doing a little debugging. I have DALI 0.16.0 installed, running the code on an Ubuntu machine with an E5-2630 CPU, 128GB of RAM and a single NVIDIA Quadro P6000 GPU.
The first error appears by simply running the script above as it is:
My first question therefore is:
next()call of theDALIGenericIterator, so the frames would be loaded only when needed. Am I wrong? Moreover, for taking 100 frames of each video, is it right to havebatch_size=1andseq_length=100?As a second experiment, I reduced the number of videos to 100. This time it seems that DALI is able to load the videos, but I got another error instead:
I am probably using the crop operation wrong, so
Finally, as a last experiment I have removed the CropMirrorNormalize operation and built the pipeline using the VideoReader only. This time the code runs with no error, but it seems to "stop" after loading 3 videos only. The terminal stayed "freezed" for several minutes, and I had to kill the process. So, I am wondering
I hope that my post is comprehensible and I apologize in advance for asking maybe too many non-related questions altogether, but I could not find any answer in the docs or in other issues here on GitHub.
Thank you in advance!