Aryan Jain

@hushell my apologies for the ping but I was hoping that one of the authors could comment on this? Again, I found your paper and codebase easy to understand but...

On second thought, is this because batched inference might not be supported? I am using the PyTorch model btw.

Really sorry for the late response. Currently, here is what I am trying to do: I have a long video of length, say, (100, 224, 224, 3). I reshape it...

Sure! Here is what the stacktrace looks like (I did wrap the model with DataParallel to allow for faster multi-batch inference) ``` File "/home/aryanjain/miniconda3/envs/fvm/lib/python3.9/site-packages/torch/nn/parallel/parallel_apply.py", line 85, in _worker output =...

@cdoersch sorry for the ping! Do you have any insight into what might be going wrong here? Let me know if there is any other information I can provide.

@cdoersch bumping the bump above!