GitLab

* Add libpng requirement into conda recipe

* Try to install libjpeg-turbo

* Add PNG reading capabilities

* Remove newline

* Add image extension to compilation instructions

* Include png functions as part of the main library

* Update CMakeLists

* Detect if building on conda-build

* Debug

* More debug messages

* Print globbed libreries

* Print globbed libreries

* Point to correct PNG path

* Remove libJPEG preventively

* Debug extension loading

* Link libpng explicitly

* Link with PNG

* Add PNG reading capabilities

* Add libpng requirement into conda recipe

* Try to install libjpeg-turbo

* Remove newline

* Add image extension to compilation instructions

* Include png functions as part of the main library

* Update CMakeLists

* Detect if building on conda-build

* Debug

* More debug messages

* Print globbed libreries

* Print globbed libreries

* Point to correct PNG path

* Remove libJPEG preventively

* Debug extension loading

* Link libpng explicitly

* Link with PNG

* Install libpng on conda-based wheel distributions

* Add -y flag

* Add -y flag to yum

* Locate LibPNG on windows conda

* Remove empty else

* Copy libpng16.so

* Copy dylib on Mac

* Improve check on Windows

* Try to install ninja using conda on windows

* Use libpng on Windows

* Package lib on windows wheel

* Point library to the correct place

* Include binaries as part of wheel

* Copy libpng.so on linux

* Look for png.h on Windows when using conda-build

* Do not skip png tests on Mac/Win

* Restore libjpeg-turbo

* Install jpeg-turbo on wheel distributions

* Install libjpeg-turbo from conda-forge on wheel distributions

* Do not pull av on conda-build

* Add pillow disclaimer

* Vendors libjpeg-turbo 2.0.4

* Merge JPEG work

* Remove submodules

* Regenerate circle config

* Fix style issues

* Fix C++ style issues

* More style corrections

* Add JPEG-turbo to linking libraries

* More style corrections

* More style corrections

* More style corrections

* Install libjpeg-turbo-devel

* Install libturbo-jpeg on typing pipeline

* Update Circle template

* Windows and Unix turbojpeg have the same linking name

* Install turbojpeg-devel instead of libjpeg-turbo

* Copy TurboJPEG binaries to wheel

* Move test image

* Move back test image

* Update JPEG test path

* Remove dot from extension

* Move image functions to extension

* Use stdout arg in subprocess

* Disable image extension if libpng or turbojpeg are not found

* Append libpng stdout

* Prevent list appending on lists

* Minor path correction

* Minor error correction

* Add linking flags

* Style issues correction

* Address minor review corrections

* Refactor library search

* Restore access index

* Fix JPEG tests

* Update libpng version in Travis

* Add -y flag

* Remove dot

* Update libpng using apt

* Check libpng version

* Change libturbojpeg binary

* Update import

* Change call

* Restore av in conda recipe

* Minor error correction

* Remove unused comment in travis.yml

* Update README

* Fix missing links

* Remove fixes for 16.04
Co-authored-by: Ryad ZENINE <r.zenine@gmail.com>

* Switch torchvision registrations to new operator registration API.

This is still registering everything as catchalls, so we're really just
moving deck chairs around, but payoff is coming soon.
Signed-off-by: Edward Z. Yang <ezyang@fb.com>

* Port roi_align to actually use dispatcher
Signed-off-by: Edward Z. Yang <ezyang@fb.com>

Summary:
Allow writes of >= 2^32 bytes. High-res video can cross this threshold sometimes.
LHS is `size_t`, but RHS is all `int32`, and will overflow for output tensors >2Gb.

Reviewed By: jsawruk

Differential Revision: D21255664

fbshipit-source-id: 7b4c5da989777297a89e73615aaeee8c7a13186a
Co-authored-by: Tilak Sharma <tilaksharma@fb.com>

* Fix missing include for OSX in video decoder

* clang-format

* fix the use of contiguous() in kernels

* clang-format

* add a contiguous in nms
Co-authored-by: Yuxin Wu <ppwwyyxx@users.noreply.github.com>

* Replace **.is_cuda() to just is_cuda()

* Replace type to scalar_type

* Fix lint, clang-format

* Fix lint, clang-format

* Add clang-format to CircleCI

* Fix for clang-format version

* Fix lint and remove Travis CI

* Seeing if lost commit comes back

* Fix lint

* Re-enable all tests

* Add test for large batches in DeformConv2d

* Clean-up and (try) fix DeformConv2d

* Simplifications and bugfixes

* Try fix CUDA now

* fix shape error for deform conv gpu op

recover shape of columns for next iteration in for loops, previous version will cause error when batch_sz / n_parallel_imgs > 1

* fix shape error for deform conv cpu op

recover shape of columns for next iteration in for loops, previous version will cause error when batch_sz / n_parallel_imgs > 1

* Fix C++ lint

* More fixes

* Base decoder for video. (#1747)

Summary:
Pull Request resolved: https://github.com/pytorch/vision/pull/1747

Pull Request resolved: https://github.com/pytorch/vision/pull/1746

Added the implementation of ffmpeg based decoder with functionality that can be used in VUE and TorchVision.

Reviewed By: fmassa

Differential Revision: D19358914

fbshipit-source-id: abb672f89bfaca6351dda2354f0d35cf8e47fa0f

* Integrated base decoder into VideoReader class and video_utils.py (#1766)

Summary:
Pull Request resolved: https://github.com/pytorch/vision/pull/1766

Replaced FfmpegDecoder (incompativle with VUE) by base decoder (compatible with VUE).
Modified python utilities video_utils.py for internal simplification. Public interface got preserved.

Reviewed By: fmassa

Differential Revision: D19415903

fbshipit-source-id: 4d7a0158bd77bac0a18732fe4183fdd9a57f6402

* Optimizating base decoder performance. (#1852)

Summary:
Pull Request resolved: https://github.com/pytorch/vision/pull/1852

Changed base decoder internals for a faster clip processing.

Reviewed By: stephenyan1231

Differential Revision: D19748379

fbshipit-source-id: 58a435f0a0b25545e7bd1a3edb0b1d558176a806

* Minor fix and decoder class members access.

Summary:
Found and fix a bug in cropping algorithm (simple mistyping).
Also derived classes need access to some decoder class members, like initialization parameters - make it protected.

Reviewed By: stephenyan1231, fmassa

Differential Revision: D19895076

fbshipit-source-id: 691336c8e18526b085ae5792ac3546bc387a6db9

* Added missing header for less dependencies. (#1898)

Summary:
Pull Request resolved: https://github.com/pytorch/vision/pull/1898

Include streams/samplers shouldn't depend on decoder headers. Add dependencies directly to the place where they are required.

Reviewed By: stephenyan1231

Differential Revision: D19911404

fbshipit-source-id: ef322a053708405c02cee4562b456b1602fb12fc

* Implemented VUE Asynchronous Decoder

Summary: For Mothership we have found that asynchronous decoder provides a better performance.

Differential Revision: D20026194

fbshipit-source-id: 627b91844b4e3f917002031dd32cb19c239f4ba8

* fix a bug in API read_video_from_memory (#1942)

Summary:
Pull Request resolved: https://github.com/pytorch/vision/pull/1942

In D18720474, it introduces a bug in `read_video_from_memory` API. Thank weiyaowang for reporting it.

Reviewed By: weiyaowang

Differential Revision: D20270179

fbshipit-source-id: 66348c99a5ad1f9129b90e934524ddfaad59de03

* extend decoder to support new video_max_dimension argument (#1924)

Summary:
Pull Request resolved: https://github.com/pytorch/vision/pull/1924

Extend `video reader` decoder python API in Torchvision to support a new argument `video_max_dimension`. This enables the new video decoding use cases. When setting `video_width=0`, `video_height=0`, `video_min_dimension != 0`, and `video_max_dimension != 0`, we can rescale the video clips so that its spatial resolution (height, width) becomes
 - (video_min_dimension, video_max_dimension) if original height < original width
 - (video_max_dimension, video_min_dimension) if original height >= original width

This is useful at video model testing stage, where we perform fully convolution evaluation and take entire video frames without cropping as input. Previously, for instance we can only set `video_width=0`, `video_height=0`, `video_min_dimension = 128`, which will preserve aspect ratio. In production dataset, there are a small number of videos where aspect ratio is either extremely large or small, and when the shorter edge is rescaled to 128, the longer edge is still large. This will easily cause GPU memory OOM when we sample multiple video clips, and put them in a single minibatch.

Now, we can set (for instance) `video_width=0`, `video_height=0`, `video_min_dimension = 128` and `video_max_dimension = 171` so that the rescale resolution is either (128, 171) or (171, 128) depending on whether original height is larger than original width. Thus, we are less likely to have gpu OOM because the spatial size of video clips is determined.

Reviewed By: putivsky

Differential Revision: D20182529

fbshipit-source-id: f9c40afb7590e7c45e6908946597141efa35f57c

* Fixing samplers initialization (#1967)

Summary:
Pull Request resolved: https://github.com/pytorch/vision/pull/1967



No-ops for torchvision diff, which fixes samplers.

Differential Revision: D20397218

fbshipit-source-id: 6dc4d04364f305fbda7ca4f67a25ceecd73d0f20

* Exclude C++ test files
Co-authored-by: Yuri Putivsky <yuri@fb.com>
Co-authored-by: Zhicheng Yan <zyan3@fb.com>

* Aligned flag in the interfaces

* Aligned flag in the impl, and remove unused comments

* Handling empty bin in forward

* Remove raise error in roi_width

* Aligned flag in the Testcodes

This reverts commit 28b7f8ae.

Summary:
Pull Request resolved: https://github.com/pytorch/vision/pull/1747

Pull Request resolved: https://github.com/pytorch/vision/pull/1746



Added the implementation of ffmpeg based decoder with functionality that can be used in VUE and TorchVision.

Reviewed By: fmassa

Differential Revision: D19358914

fbshipit-source-id: abb672f89bfaca6351dda2354f0d35cf8e47fa0f
Co-authored-by: Yuri Putivsky <yuri@fb.com>

* Add Deformable Convolution operation.

This adds the deformable convolution operation, as described in Deformable Convolutional Networks (https://arxiv.org/abs/1703.06211).

- The code is based on https://github.com/open-mmlab/mmdetection/blob/master/mmdet/ops/dcn/src/deform_conv_cuda.cpp ; the whole code was modified and refactored to remove redundancies and increase clarity, and to adapt it to torchvision.

- The CPU part is a direct copy of the CUDA code; it might make sense to do follow-up adjustments in the CPU code to simplify it / optimize it, or to reuse functionality between CPU and CUDA..

- We also add tests (with a non-trivial set of parameters); they can be made more robust by randomizing the parameters and executing multiple times.

* Update DeformConv to be more consistent w/ Conv2d

* rename some variables and arguments to match Conv2d;
* add optional bias;
* add weight, offset and bias as module parameters;
* remove the n_parallel_imgs parameter;
* Fix __repr__;
* etc..

Initialization of weight and bias is the same as in Conv2d, and
initialization of offsets to zero is the same as in the paper.

This also includes some other small unrelated fixes/improvements.

* Apply clang-format in DeformConv files.

* Import Optional type annotation

* Remove offset param from DeformConv2d module

- We pass the offset in the forward of DeformConv2d, instead of having
an internal parameter. This adds some complexity to creating the module
(e.g. now you have to worry about the output size, to create the
offset), but it gives more flexibility.
- We also use make_tuple for tuple creation, in an attempt to fix error
w/ older compilers.

* Replace abs by std::abs

Old gcc versions were giving wrong results here, because they would
resolve abs as int -> int, thus causing undesired truncation. Replacing
abs by std::abs should allow for correct overloading of abs as float -> float.

* Reorder declarations for clarity

* Reorder weight and offset args in deform_conv2d

We place offset arg before the weight arg, to be more
consistent with DeformConv2d.forward(input, offset)

* Replace abs by std::abs in DeformConv_cuda

* Fix inconsistent NMS implementation

* Improve tests for NMS

* Remove unnecessary using statement

* added PSRoiAlign and PSRoiPool with C++ autograd and torch ops

* fixed linter errors

* fixed linter errors 2

* fixed linter errors 3

* extend video reader to support fast video probing

* fix c++ lint

* small fix

* allow to accept input video of type torch.Tensor

* [video reader] inception commit

* add method save_metadata to class VideoClips in video_utils.py

* add load_metadata() method to VideoClips class

* add Exception to not catch unexpected events such as memory erros, interrupt

* fix bugs in video_plus.py

* [video reader]remove logging. update setup.py

* remove time measurement in test_video_reader.py

* Remove glog and try making ffmpeg finding more robust

* Add ffmpeg to conda build

* Add ffmpeg to conda build [again]

* Make library path finding more robust

* Missing import

* One more missing fix for import

* Py2 compatibility and change package to av to avoid version conflict with ffmpeg

* Fix for python2

* [video reader] support to decode one stream only (e.g. video/audio stream)

* remove argument _precomputed_metadata_filepath

* remove save_metadata method

* add get_metadata method

* expose _precomputed_metadata and frame_rate arguments in video dataset __init__ method

* remove ssize_t

* remove size_t to pass CI check on Windows

* add PyInit__video_reader function to pass CI check on Windows

* minor fix to define PyInit_video_reader symbol

* Make c++ video reader optional

* Temporarily revert changes to test_io

* Revert changes to python files

* Rename files to make it private

* Fix python lint

* Fix C++ lint

* add a functor object EnumClassHash to make Enum class instances usable as key type of std::unordered_map

* fix cpp format check

* Use Tensor.data_ptr instead of .data

* use pytorch-nightly in CI

Updated nms_cuda signature to accept detections and scores as separate tensors.
This also required updating the indexing in the NMS CUDA kernel.

Also made the iou_threshold parameter name consistent across implementations.

* Initial layout for layers with cpp extensions

* Move files around

* Fix import after move

* Add support for multiple types to ROIAlign

* Different organization

CUDA extensions work now

* Cleanups

* Reduce memory requirements for backwards

* Replace runtime_error by AT_ERROR

* Add nms test

* Add support for compilation using CPP extensions

* Change folder structure

* Add ROIPool cuda

* Cleanups

* Add roi_pool.py

* Fix lint

* Add initial structures folder for bounding boxes

* Assertion macros compatible with pytorch master (#540)

* Support for ROI Pooling (#592)

* ROI Pooling with tests. Fix for cuda context in ROI Align.

* renamed bottom and top to follow torch conventions

* remove .type().tensor() calls in favor of the new approach to tensor initialization (#626)

* Consistent naming for rois variable (#627)

* remove .type().tensor() calls in favor of the new approach to tensor initialization

* Consistent naming for rois variable in ROIPool

* ROIPool: Support for all datatypes (#632)

* Use of torch7 naming scheme for ROIAlign forward and backward

* use common cuda helpers in ROIAlign

* use .options() in favor of .type() where applicable

* Added tests for forward pass of ROIAlign, as well as more consistent naming scheme for CPU vs CUDA

* working ROIAlign cuda backwards pass

* working ROIAlign backwards pass for CPU

* added relevant headers for ROIAlign backwards

* tests for ROIAlign layer

* replace .type() with .options() for tensor initialization in ROIAlign layers

* support for Half types in ROIAlign

* gradcheck tests for ROIAlign

* updated ROIPool on CPU to work with all datatypes

* updated and cleaned tests for ROI Pooling

* Fix rebase problem

* Remove structures folder

* Improve cleanup and bugfix in test_layers

* Update C++ headers

* Add CUDAGuard to cu files

* Add more checks to layers

* Add CUDA NMS and tests

* Add multi-type support for NMS CUDA

* Avoid using THCudaMalloc

* Add clang-format and reformat c++ code

* Remove THC includes

* Rename layers to ops

* Add documentation and rename functions

* Improve the documentation a bit

* Fix some lint errors

* Fix remaining lint inssues

* Area computation doesn't add +1 in NMS

* Update CI to use PyTorch nightly

* Make NMS return indices sorted according to the score

* Address reviewer comments

* Lint fixes

* Improve doc for roi_align and roi_pool

* move to xenial

* Fix bug pointed by @lopuhin

* Fix RoIPool reference implementation in Python 2

Also fixes a bug in the clip_boxes_to_image -- this function needs a test!

* Remove change in .travis