Skip to content

Torchvision backend not working. #181

@kiyoon

Description

@kiyoon

Hi,

I was using the Kinectics dataloader with torchvision backend.
I had a very similar issue as #148, saying that there's no operator called probe_video_from_memory.

/opt/conda/conda-bld/pytorch_1587428398394/work/torch/csrc/utils/tensor_numpy.cpp:141: UserWarning: The given NumPy array is not writeable, and PyTorch does not support non-writeable tensors. This means you can write to the underlying (supposedly non-writeable) NumPy array using the tensor. You may want to copy the array to protect its data or make it writeable before converting it to a tensor. This type of warning will be suppressed for the rest of this program.
/opt/conda/conda-bld/pytorch_1587428398394/work/torch/csrc/utils/tensor_numpy.cpp:141: UserWarning: The given NumPy array is not writeable, and PyTorch does not support non-writeable tensors. This means you can write to the underlying (supposedly non-writeable) NumPy array using the tensor. You may want to copy the array to protect its data or make it writeable before converting it to a tensor. This type of warning will be suppressed for the rest of this program.
/opt/conda/conda-bld/pytorch_1587428398394/work/torch/csrc/utils/tensor_numpy.cpp:141: UserWarning: The given NumPy array is not writeable, and PyTorch does not support non-writeable tensors. This means you can write to the underlying (supposedly non-writeable) NumPy array using the tensor. You may want to copy the array to protect its data or make it writeable before converting it to a tensor. This type of warning will be suppressed for the rest of this program.
/opt/conda/conda-bld/pytorch_1587428398394/work/torch/csrc/utils/tensor_numpy.cpp:141: UserWarning: The given NumPy array is not writeable, and PyTorch does not support non-writeable tensors. This means you can write to the underlying (supposedly non-writeable) NumPy array using the tensor. You may want to copy the array to protect its data or make it writeable before converting it to a tensor. This type of warning will be suppressed for the rest of this program.
Failed to decode by torchvision with exception: No such operator video_reader::probe_video_from_memory
Failed to decode by torchvision with exception: No such operator video_reader::probe_video_from_memory
Failed to decode by torchvision with exception: No such operator video_reader::probe_video_from_memory
Failed to decode by torchvision with exception: No such operator video_reader::probe_video_from_memory
Failed to decode the video index 0
Failed to decode the video index 4
Choosing random video of index 21667 instead
Choosing random video of index 14461 instead
Failed to decode the video index 12
Failed to decode the video index 8
Choosing random video of index 14195 instead
Choosing random video of index 15546 instead
Failed to decode by torchvision with exception: No such operator video_reader::probe_video_from_memory

...

It used to be read_video_meta_data_from_memory but apparently this repository is updated.
I tried using PyTorch 1.4.0 / TorchVision 0.5.0 and PyTorch 1.5.0 / TorchVision 0.6.0 but no luck.
It seems like it's looking for the operator on PyTorch side, not TorchVision, which I don't understand why.

Any thoughts on how to debug and what do check?
Thank you for your support.

Metadata

Metadata

Assignees

No one assigned

    Labels

    questionFurther information is requested

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions