GitLab

Co-authored-by: Yuxin Wu <ppwwyyxx@users.noreply.github.com>

To resolve issue #5815
Co-authored-by: Vasilis Vryniotis <datumbox@users.noreply.github.com>

* Removed custom ops for interp with AA

* Fixed umft issues
Co-authored-by: Vasilis Vryniotis <datumbox@users.noreply.github.com>

fbshipit-source-id: 85954dbb817fac16099a3b103b24defc1a8672f8
(cherry picked from commit 9642c8e710e8911c64b9878f51a30ecde95ac90a)
Co-authored-by: Richard Barnes <rbarnes@fb.com>

* change to stable sort in nms implementations

Summary:
Pull Request resolved: https://github.com/pytorch/vision/pull/4620

Pull Request resolved: https://github.com/pytorch/nestedtensor/pull/455



Fixes missing kernel guards as identified by D30072495

Reviewed By: jingsh, xush6528

Differential Revision: D31553158

fbshipit-source-id: 80de017ba2ddc52e2a684df9b3eae5de84ed49f4
Co-authored-by: Richard Barnes <rbarnes@fb.com>

* Update gpu atomics include path

THC is being removed, and this header now lives in ATen.
See pytorch/pytorch#65470.

* Empty commit to trigger build
Co-authored-by: Vasilis Vryniotis <datumbox@users.noreply.github.com>

* WIP on fixing index overflow issue

* Fixed backward pass for large num_kernels

* Fixed clang formatting

* Fixed GET_BLOCKS int/int64_t types issue
Co-authored-by: vfdev-5 <vfdev-5@gmail.com>
Co-authored-by: Francisco Massa <fvsmassa@gmail.com>

* WIP on backward op interpolation with AA

* Removed cuda tests and reformat cpp code

* Fixed clang wrong formatting

* Added channels last test case

* Added CUDA support for backward pass, interpolation with AA

* Removed unused buffers
Co-authored-by: vfdev-5 <vfdev-5@gmail.com>

* WIP on backward op interpolation with AA

* Removed cuda tests and reformat cpp code

* Fixed clang wrong formatting

* Added channels last test case
Co-authored-by: vfdev-5 <vfdev-5@gmail.com>
Co-authored-by: Francisco Massa <fvsmassa@gmail.com>

* Added CUDA support for interpolation AA bilinear/bicubic

* Fixed code formatting error

* Fixed cuda tests on cpu-only builds

* Fixed cuda internal torch check for bicubic mode

* Replace type T with accumulator.

* Upcast tensors of box ops to avoid overflow in multiplications.

Co-authored-by: Vasilis Vryniotis <datumbox@users.noreply.github.com>

* Adding TORCH_SELECTIVE_* macros on op registration.

* Adding torchvision namespace.

Summary:
* Reduce unnecessary header inclusions in models and io.

* Move autocast to separate folder and hide autograd implementation in an anonymous namespace.

* Moving files in subfolders.

Reviewed By: fmassa

Differential Revision: D25461523

fbshipit-source-id: 756eeb6848aacaa474de4825ed4c1045d17e2cea