GitLab

* return features instead of vanilla tensors from prototype datasets

* fix tests

* remove inplace

* add explanation for __init_subclass__

* fix label for test split

* relax test

* remove pixels

* refactoring methods from MultiScaleRoIAlign

* Add new weights for MobileNetV3 Large.

* Update weights for ResNet101.

* Update weights for ResNet152.

* Update numbers of ResNet101 and ResNet152 with batch size 1.

* Adding new weights for ResNeXt101_32x8d

* Batch size 1 stats for ResNeXt101_32x8d

* Update weights for MobileNetV3 Large.

* Update weights for ResNeXt50.

* Fix merge issues.

* Updating the URLs to indicate the recipe config.

* Disable WeightEntry to pass-through `Weights.verify()`

* Rename `Weights.state_dict()` to `Weights.get_state_dict()`

* Add TODO for missing doc.

* Moving warning messages for googlenet.

* Upper-case global `_COMMON_META` var

* Replace argument with parameter in all warnings.

Closes #4905
Co-authored-by: Vasilis Vryniotis <datumbox@users.noreply.github.com>

* Add new weights for wide_resnet101_2 model

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

* add initial chunk of prototype transforms

* fix tests

* add error message

* fix more imports

* add explicit no-ops

* add test for no-ops

* cleanup

This reverts commit 85b78580.

* Start writing gallery example

* Remove the child image fix implementation add code

* add docs

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

* address review update thumbnail
Co-authored-by: Philip Meier <github.pmeier@posteo.de>
Co-authored-by: Vasilis Vryniotis <datumbox@users.noreply.github.com>

Description:
- Applied `contiguous` on decoded output tensor in decode_jpeg and decode_png ops
- Updated tests and docs

Related to #4880

* fix

* Outline Keypoints API

* Add utility

* make it work :)

* Fix optional type

* Add connectivity, fmassa's advice 😃



* Minor code improvement

* small fix

* fix implementation

* Add tests

* Fix tests

* Update colors

* Fix bug and test more robustly

* Add a comment, merge stuff

* Fix fmt

* Support single str for merging

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

Reverted folder.py back to using complete path to file for make_dataset and is_valid_file rather than just the filename (#4885)
Co-authored-by: Alexander Unnervik <alex.unnervik@idiap.ch>

* Adding multiweight support to SSDlite.

* Fixing lint and mypy issues.

* Adding size in the meta.

* Adding interpolation in meta.

* add test split for imagenet

* add infinite buffer size to shuffler

* remove unused imports

* update lsun

* Adding multiweight support to Keypoint RCNN.

* Fix linter.

* Updated wide_resnet50_2 model weights

* Update Accs with batch-size 1.

* Update checkpoint and accs.
Co-authored-by: Vasilis Vryniotis <datumbox@users.noreply.github.com>

Co-authored-by: Francisco Massa <fvsmassa@gmail.com>

* Adding multiweight suport on Quant MobileNetV2 and MobileNetV3.

* Fixing enum name.

* Fixing lint.

* Clean up unnecessary quant builders and add quant weights for 0.5

* Fixing mypy.

* add prototype features

* add some JIT tests

* refactor input data handling

* refactor tests

* cleanup tests

* add BoundingBox feature

* mypy

* xfail torchscript tests for now

* cleanup

* fix imports

* add tests for builtin prototype datasets

* fix caltech101

* fix emnist

* fix mnist and variants

* add iopath as test requirement

* fix MNIST warning

* fix qmnist data generation

* fix cifar data generation

* add tests for imagenet

* cleanup

* Moving builder to the bottom to use proper typing.

* Renaming weights.

* Adding quantizated inception builder.

* Correct meta info.

* Fix linter.

* Removing init_weights to avoid exposing it on the class.