GitLab

* added note

* quantize = True higlighted in the note.

* Keep "Large" in docstring
Co-authored-by: Nicolas Hug <contact@nicolas-hug.com>

* Create resnet_quant.rst

* add resnet quant

* refactor docs

* Minor fix

* Nit
Co-authored-by: Nicolas Hug <contact@nicolas-hug.com>

* [DOC] Add quantized inception updated documentation.

* Add missing file.

* Apply suggestions from code review

[ENH] Improve doc thanks to various code review comments.
Co-authored-by: Aditya Oke <47158509+oke-aditya@users.noreply.github.com>
Co-authored-by: Nicolas Hug <contact@nicolas-hug.com>
Co-authored-by: Aditya Oke <47158509+oke-aditya@users.noreply.github.com>
Co-authored-by: Nicolas Hug <contact@nicolas-hug.com>

* Adding code examples for image classification + quant

* Adding code example detection

* Adding code example segmentation

* Adding code example for video classification

* Adding information on how to use the new API.

* Putting back the comma.

* Apply suggestions from code review
Co-authored-by: Nicolas Hug <contact@nicolas-hug.com>

* Remove output to avoid staleness from flakiness.

* Minor fixes.
Co-authored-by: Nicolas Hug <contact@nicolas-hug.com>

* Adding `__repr__` in presets

* Adds `describe()` methods to all presets.

* Adding transform descriptions in the documentation.

* Change "preprocessing" to "inference"

* Remove old "minimum input size" from docstrings.

* Remove "currently only XYZ weights available"

* Fix description of wide_resnet101_2

* Make display license URLs as links.

* Clarify the order of dims of min_size.

* Remove lengthy keypoint_names from meta-table.

* added quantized mobilenetv2 docs

* remove quotes
Co-authored-by: Nicolas Hug <contact@nicolas-hug.com>

* Reinstate and deprecate `model_urls` and `quant_model_urls`

* Apply suggestions from code review
Co-authored-by: Nicolas Hug <contact@nicolas-hug.com>

* Move todo location

* Add alexnet

* Add densenet

* Add efficientnet

* Add googlenet.

* Add inception.

* Add mobilenetv3

* Add regnet

* Add resnet

* Add shufflenetv2

* Fix linter

* Add squeezenet

* Add vgg

* Add vit

* Add quantized googlenet

* Add quantized inceptionv3

* Add quantized mobilenet_v3

* Add quantized resnet

* Add quantized shufflenetv2

* Fix incorrect imports

* Add faster_rcnn

* Add fcos

* Add keypoint rcnn

* Add mask rcnn

* Add retinanet

* Add ssd

* Add ssdlite

* Add deeplabv3

* Add fcn

* Add lraspp.

* Add video resnet

* Removing weights for shufflenetv2_x1.5 and shufflenetv2_x2.0

* Update the comments
Co-authored-by: Nicolas Hug <contact@nicolas-hug.com>

Co-authored-by: Nicolas Hug <contact@nicolas-hug.com>

* rely on patched datasets home rather than passing it around

* add comment

* Clarify TypeError message

* typo

* Wrap tye() in repr()

so that the type-check stfu

* Even better message

incidentally speeds up the thing, since all() possibly checked all elements and we don't.
Co-authored-by: Vasilis Vryniotis <datumbox@users.noreply.github.com>

* POC

* Update torchvision/models/resnet.py
Co-authored-by: Vasilis Vryniotis <datumbox@users.noreply.github.com>

* Apply suggestions from code review
Co-authored-by: Vasilis Vryniotis <datumbox@users.noreply.github.com>

* Fix tests

* ufmt

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

* Adding ciou and diou support in `_box_loss()`

* Fix linter

* Addressing comments for nits

* remove explicit install of Pillow throughout CI

* explicitly install zlib in windows conda env

* check pillow on Windows

* move check into conda env

* fix

* more debug output

* more debug

* add debug also to linux

* retry w/ and w/o packaging third party library

* tag debug

* more debug

* pin setuptools < 61

* install setuptools from pip

* cleanup

* try remove one more
Co-authored-by: Vasilis Vryniotis <datumbox@users.noreply.github.com>

* Add Docs for QuantizedGoogleNet

* ufmt

* Use explicit _QuantizedWeights or _Weights suffix

* [FEAT] Add distance IoU and distance IoU loss + some tests (WIP for tests).

* [FIX] Remove URL from docstring + remove assert since it causes a big performance drop.

* [FIX] eps isn't None.

* [TEST] Update existing box dIoU test + add dIoU loss tests (inspired from cIoU ones).

* [ENH] Some pre-commit fixes + remove print + mypy.

* [ENH] Pass the device in the assertion for the dIoU loss test.

* [FIX] Remove type hints from the dIoU box test.

* [ENH] Refactor box and loss for dIoU functions + fix half tests.

* [FIX] Precommits fix.

* [ENH] Some improvement for the distance IoU tests thanks to code review.

* [ENH] Upcast in distance boxes computation to avoid overflow.

* [ENH] Revert the refactor of distance IoU loss back since it introduced a bug and can be slow.

* Precommit fix.

* [FIX] Few changes introduced by merge conflict.

* Add code reference

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

* Fixing AlexNet references

* Fix doc rendering

* [proto] Added `center_crop_bounding_box` functional op

* Fixed mypy issue

* Added one more test case

* More test cases

* Added functional `perspective_bounding_box`/`perspective_segmentation_mask` ops

* Added more comments and added a code to assert denom != 0

* Put larger r/a tolerence when matching bboxes

* Update transforms for PIL deprecation

* Changes agreed at pytorch/vision#5898

* black, sort constants, version check

* Format tests

* Square brackets

* Update torchvision/transforms/_pil_constants.py
Co-authored-by: Nicolas Hug <contact@nicolas-hug.com>
Co-authored-by: Philip Meier <github.pmeier@posteo.de>
Co-authored-by: Nicolas Hug <contact@nicolas-hug.com>
Co-authored-by: Vasilis Vryniotis <datumbox@users.noreply.github.com>

* Add resnext101_64x4d model definition

* Add test for resnext101_64x4d

* Add resnext101_64x4d weight

* Update checkpoint to use EMA weigth

* Add quantization model signature for resnext101_64x4d

* Fix class name and update accuracy using 1 gpu and batch_size=1

* Apply ufmt

* Update the quantized weight and accuracy that we still keep the training log

* Add quantized expect file

* Update docs and fix acc1

* Add recipe for quantized to PR

* Update models.rst

* [ROCm] Upgrade to rocm5.1.1

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

* Add mask rcnn to instance segmentation

* Add docs a use nicolas suggestion

* remov param

* remov docsstring, edit docs

* Add one section level
Co-authored-by: Nicolas Hug <contact@nicolas-hug.com>

* Document Keypoint RCNN separately

* Move Keypoint detection into its own section

* ufmt

* Remove references to non-existing weights

* Update torchvision/models/mnasnet.py
Co-authored-by: Nicolas Hug <contact@nicolas-hug.com>

* Update torchvision/models/mnasnet.py
Co-authored-by: Nicolas Hug <contact@nicolas-hug.com>

* apply ufmt
Co-authored-by: Nicolas Hug <contact@nicolas-hug.com>

* docs: Added SSD300 to the new doc

To resolve #5897

* Minor fixes

* ufmt
Co-authored-by: Nicolas Hug <contact@nicolas-hug.com>

* Change code to reduce variance in eval

* Remove unnecessary new line

* Fix missing import warnings

* Fix the warning on video_classification

* Fix bug to get len of UniformClipSampler