GitLab

* basic stanford cars structure added.

* added class in the init.

* how to decode the images

* fix typo in init

* works partially

* formatting

* formatting

* formatting

* formatting

* formatting

* minor changes

* minor changes

* issue resolved

* both datapipes are in sync

* checksum dict keys changed

* formatting

* "ufmt formatting"

* "ufmt formatting"

* formatting

* minor change

* minor change

* test added

* lint (end of line issue)

* lint (end of line issue)

* minor changes

* mock dataset formatting

* minor changes

* hopefully final

* formatting

* prototype test

* prototype test

* prototype test

* prototype test

* prototype test

* prototype test

* prototype test

* prototype test

* Update builtin_dataset_mocks.py

* Update stanford_cars.py

* Update torchvision/prototype/datasets/_builtin/stanford_cars.py
Co-authored-by: Philip Meier <github.pmeier@posteo.de>

* Update torchvision/prototype/datasets/_builtin/stanford_cars.py
Co-authored-by: Philip Meier <github.pmeier@posteo.de>

* Update torchvision/prototype/datasets/_builtin/stanford_cars.py
Co-authored-by: Philip Meier <github.pmeier@posteo.de>

* Update torchvision/prototype/datasets/_builtin/stanford_cars.py
Co-authored-by: Philip Meier <github.pmeier@posteo.de>

* formatting and cleaning

* formatting and cleaning

* Update torchvision/prototype/datasets/_builtin/stanford_cars.py
Co-authored-by: Philip Meier <github.pmeier@posteo.de>

* generate catefories script

* generate categories script added

* sigh

* sigh

* sigh

* minor changes

* linting
Co-authored-by: Philip Meier <github.pmeier@posteo.de>
Co-authored-by: Nicolas Hug <contact@nicolas-hug.com>

* Measuring execution times of models.

* Speed up models by avoiding re-estimation of eager output

* Fixing linter

* Reduce input size for big models

* Speed up jit check method.

* Add simple jitscript fallback check for flaky models.

* Restore pytest filtering

* Fixing linter

* support grayscale / RGB alpha conversions

* use _max_valu from stable

* remove extra copy for PIL conversion

* simplify test image generation for color spaces with alpha channel

* use common _max_value in tests

* replace dynamically created dicts with if/else

* make color space conversion more explicit

* make even more explicit

* simplify alpha image generation

* fix if / elif

* add error for unknown conversions

* rename RGBA to RGB_ALPHA

* cleanup

* GRAYSCALE to GRAY

* Removed _max_value method and added a dictionary

Related to https://github.com/pytorch/vision/issues/5502

* Addressed failing tests and restored _max_value method

* Added xfailing test to switch quicker

* Switch to if/else impl

* Remove from models and references.

* Adding most tests and docs.

* Adding transforms tests.

* Remove unnecesary ipython notebook.

* Simplify tests.

* Addressing comments.

* add eurosat

* revert formatting

* port test and make style changes

* add eurosat to __init__

* fix pathlib error

* create dataset zipfile and revert pre commit changes

* remove unecessary variable in resources

* revert auto formatter changes and modify ufmt version

* revert change to contributing guide
Co-authored-by: Dbhasin1 <drishti_b@me.iitr.c.in>
Co-authored-by: Nicolas Hug <nicolashug@fb.com>

* add country211

* remove unused import

* map val to valid and use path comparator

* remove unused import

* resolve keyerror

* map split names in dataset mock
Co-authored-by: Dbhasin1 <drishti_b@me.iitr.c.in>
Co-authored-by: Nicolas Hug <nicolashug@fb.com>

* improve StrEnum

* use StrEnum for model weights

* fix test

* migrate StrEnum to main area

* Extend the EfficientNet class to support v1 and v2.

* Refactor config/builder methods and add prototype builders

* Refactoring weight info.

* Update dropouts based on TF config ref

* Update BN eps on TF base_config

* Use Conv2dNormActivation.

* Adding pre-trained weights for EfficientNetV2-s

* Add Medium and Large weights

* Update stats with single batch run.

* Add accuracies in the docs.

* Improve test_video_reader

* Fix linter error

* expand prototype functional scriptability tests

* remove obsolete skips

* add prototype AugMix transform

* cleanup

* refactor auto augment subclasses to only trnasform a single image

* address review comments

* Replace get_image_size/num_channels with get_image_dims

* Reduce verbosity

* Fix JIT-scriptability

* Refactoring

* More refactoring

* Replace all _FP/_FT direct calls.

* Remove usages of get_image_size and get_image_num_channels from code-base.

* Fix JIT issues

* Adding missing assertion.

* Create dropblock.py

* add dropblock2d

* fix pylint

* refactor dropblock

* add dropblock

* Rename dropblock.py to drop_block.py

* fix pylint

* add dropblock

* add dropblock3d

* add drop_block3d

* add dropblock

* Update drop_block.py

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

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

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

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

* Update drop_block.py

* Update drop_block.py

* import torch.fx

* fix lint

* fix lint

* Update drop_block.py

* improve dropblock

* add dropblock

* refactor dropblock

* fix doc

* remove the limitation of block_size

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

* fix lint

* fix lint

* add dropblock

* Fix linter

* add dropblock random check

* reduce test time

* Update test_ops.py

* speed the dropblock test

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

* add prototype transforms that don't need dispatchers

* cleanup

* remove legacy_transform decorator

* remove legacy classes

* remove explicit param passing

* streamline extra_repr

* remove obsolete ._supports() method

* cleanup

* remove Query

* cleanup

* fix tests

* kernels -> functional

* move image size and num channels extraction to functional

* extend legacy function to extract image size and num channels

* implement dispatching for auto augment

* fix auto augment dispatch

* revert some naming changes

* remove ability to pass params to autoaugment

* fix legacy image size extraction

* align prototype.transforms.functional with transforms.functional

* cleanup

* fix image size and channels extraction

* fix affine and rotate

* revert image size to (width, height)

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

Also, remove reference to conda-forge, all CUDA toolchain should be
available from NVIDIA channel

Install h5py from pip on Windows and skip failing gaussian_blur tests if Win+CUDA11.3 is used

* Adding basic augmix implementation.

* Finish the implementation.

* Add tests and documentation.

* Fix tests.

* Simplify code.

* Speed optimizations.

* Per image weights instead of per batch.

* Fix tests.

* Update torchvision/transforms/autoaugment.py
Co-authored-by: vfdev <vfdev.5@gmail.com>

* Changing the default severity value to get by default the same strength as RandAugment.
Co-authored-by: vfdev <vfdev.5@gmail.com>

* add prototype transforms that use the prototype dispatchers

Conflicts:
	torchvision/prototype/transforms/__init__.py

* simplify

* add logger

* remove legacy classes

Conflicts:
	torchvision/prototype/transforms/_augment.py
	torchvision/prototype/transforms/_auto_augment.py
	torchvision/prototype/transforms/_geometry.py

* make get_params private

* remove randbool method

* remove AutoAugmentDispatcher

* add high level kernels for meta conversion

* remove transforms meta abstraction from auto augment transforms

* appease mypy

* add smoke tests for transforms

* remove Query object

* remove extra_repr helper

* fix tests

* appease mypy

* revert some changes on the kernel tests

* fix dispatcher annotations

* remove float cast for torch.rand

* add helper to query image

* fix imports

* address auto augment comments

* cleanup

* Fix resize when size == small_edge_size and max_size isn't None

* Better test name

* Removed custom ops for interp with AA

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

* revamp prototype features (#5283)

* remove decoding from prototype datasets (#5287)

* remove decoder from prototype datasets

* remove unused imports

* cleanup

* fix readme

* use OneHotLabel in SEMEION

* improve voc implementation

* revert unrelated changes

* fix semeion mock data

* fix pcam

* readd functional transforms API to prototype (#5295)

* readd functional transforms

* cleanup

* add missing imports

* remove __torch_function__ dispatch

* readd repr

* readd empty line

* add test for scriptability

* remove function copy

* change import from functional tensor transforms to just functional

* fix import

* fix test

* fix prototype features and functional transforms after review (#5377)

* fix prototype functional transforms after review

* address features review

* make mypy more strict on prototype features

* make mypy more strict for prototype transforms

* fix annotation

* fix kernel tests

* add automatic feature type dispatch to functional transforms (#5323)

* add auto dispatch

* fix missing arguments error message

* remove pil kernel for erase

* automate feature specific parameter detection

* fix typos

* cleanup dispatcher call

* remove __torch_function__ from transform dispatch

* remove auto-generation

* revert unrelated changes

* remove implements decorator

* change register parameter order

* change order of transforms for readability

* add documentation for __torch_function__

* fix mypy

* inline check for support

* refactor kernel registering process

* refactor dispatch to be a regular decorator

* split kernels and dispatchers

* remove sentinels

* replace pass with ...

* appease mypy

* make single kernel dispatchers more concise

* make dispatcher signatures more generic

* make kernel checking more strict

* revert doc changes

* address Franciscos comments

* remove inplace

* rename kernel test module

* fix inplace

* remove special casing for pil and vanilla tensors

* address comments

* update docs

* cleanup features / transforms feature branch (#5406)

* mark candidates for removal

* align signature of resize_bounding_box with corresponding image kernel

* fix documentation of Feature

* remove interpolation mode and antialias option from resize_segmentation_mask

* remove or privatize functionality in features / datasets / transforms

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

* Added ref tests for shear X/Y

* Added PIL tests and fixed tan(level) difference

* Updated tests

* Fixed reparam for shear X/Y in autoaugment

* Fixed arc_level -> level as atan is applied internally

* Fixed links

* Refactor BoxOps tests to use parameterize

* Refactor BoxOps tests to use parameterize

* Refactor BoxOps to use parameterize, addressed comments from PR#5380

* Refactor BoxOps to use parameterize, addressed minor styling comments from PR#5380

* Refactor BoxOps to use parameterize, addressed typing errorsfrom PR#5380

* Refactor BoxOps to use parameterize, addressed minor naming comments for PR#5380

* properly deprecate legacy implementation

* cleanup

* use warning over deprecation directive

* remove patch version

* fix link in Kinetics docstring

* Some more

* fix affine functional tests
Co-authored-by: Nicolas Hug <nicolashug@fb.com>

* Remove APIs that were deprecated before 0.8

* More stuff

* Some more

* oops

* Add random colors

* Update error message, pretty the code

* Update edge cases

* Change implementation to tuples

* Fix bugs

* Add tests

* Reuse palette

* small rename fix

* Update tests and code

* Simplify code

* ufmt

* fixed colors -> random colors in docstring

* Actually simplify further

* Silence mypy. Twice. lol.
Co-authored-by: Nicolas Hug <nicolashug@fb.com>

* Add is_qat support using a method getter

* Switch to an internal _fuse_modules

* Fix linter.

* Pass is_qat=False on PTQ

* Fix bug on ra_sampler flag.

* Set is_qat=True for QAT

* Updated gpu decoder test to use parametrize

* Remove list comprehension

* Refactor model builder

* Add 3 more convnext variants.

* Adding weights for convnext_small.

* Fix minor bug.

* Fix number of parameters for small model.

* Adding weights for the base variant.

* Adding weights for the large variant.

* Simplify LayerNorm2d implementation.

* Optimize speed of CNBlock.

* Repackage weights.

* Adding conv_stem support

* fix lint

* bug fix

* address comments

* fix after merge

* adding back checking lines

* fix failing tests

* fix iignore

* add unittest & address comments

* fix memory issue

* address comments

* Some Qs

* Some modifications

* don't need _loader in __init__

* list_names -> list_datasets

* Update torchvision/prototype/datasets/utils/_resource.py
Co-authored-by: Philip Meier <github.pmeier@posteo.de>

* Remove unsued import

* fix tests

* Some missing renames
Co-authored-by: Philip Meier <github.pmeier@posteo.de>