GitLab

* 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

* 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 prototype flag on reference scripts.

* Import prototype instead of models/transforms.

* Correcting exception type.

* fixing none referencing

* Adding CNBlock and skeleton architecture

* Completed implementation.

* Adding model in prototypes.

* Add test and minor refactor for JIT.

* Fix mypy.

* Fixing naming conventions.

* Fixing tests.

* Fix stochastic depth percentages.

* Adding stochastic depth to tiny variant.

* Minor refactoring and adding comments.

* Adding weights.

* Update default weights.

* Fix transforms issue

* Move convnext to prototype.

* linter fix

* fix docs

* Addressing code review comments.

* Adding pretrained ViT weights

* Adding recipe as part of meta

* update checkpoints using best ema results

* Fix handle_legacy_interface and update recipe url

* Update README

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

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

As titled.

* support random seed used to shuffle the sampler

* fix bug for RA samper
Co-authored-by: Vasilis Vryniotis <datumbox@users.noreply.github.com>

* Adding repaeted data-augument sampler

* rebase on top of latest main

* fix formatting

* rename file

* adding coode source

* Change the `default` weights mechanism to sue Enum aliases.

* Change `get_weights` to work with full Enum names and make it public.

* Applying improvements from code review.

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

* Add multi-type support on get_weight.

* Fix bug on method call.

* Adding logging suffix for QAT.

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

* fix bug in amp

* fix bug in training by amp

* support use gradient clipping when amp is enabled
Co-authored-by: Vasilis Vryniotis <datumbox@users.noreply.github.com>
Co-authored-by: Prabhat Roy <prabhatroy@fb.com>

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

* Fixing mypy.

* Reordering the builders to use proper typing.

* Adding additional meta-data on existing quantized models.

* Fixing meta on unquantized model.

* Adding quantized googlenet builder.

* undo inception move.

* Adding recipe information.

* [references] Adding gradient clipping

* ufmt formatting

* remove apex code

* resolve naming issue

* Adding multi-weight support to Quantized ResNet.

* Update references script to support testing quantized models with the new API.

* Handle quantized models correctly in ref script.

* Fixing references for quantization.

* Update training references from legacy models.

* Refactoring to share common parts.

* Moving original builder at the bottom of the page to use proper typing.

* Adding multiweight support to inception.

* Update doc.

Co-authored-by: Nicolas Hug <nicolashug@fb.com>

* Update model checkpoint for resnet50.

* Add get_weight method to retrieve weights from name.

* Update the references to support prototype weights.

* Fixing mypy typing.

* Switching to a python 3.6 supported equivalent.

* Add unit-test.

* Add optional num_classes.

* Add type to default argument

To resolve issue #4694

* Resolve issue #4694

Add missing types on argument parser

* Update with ufmt

formatted with ufmt

* Updated with review

Updated with review

* Update type of arguments

Add train.py from video_classification, similarity and train_quantization.py
Co-authored-by: Vasilis Vryniotis <datumbox@users.noreply.github.com>

* Update EMA every X iters.

* Adding AdamW optimizer.

* Adjusting EMA decay scheme.

* Support custom weight decay for Normalization layers.

* Fix identation bug.

* Change EMA adjustment.

* Quality of life changes to faciliate testing

* ufmt format

* Fixing imports.

* Adding FixRes improvement.

* Support EMA in store_model_weights.

* Adding interpolation values.

* Change train_crop_size.

* Add interpolation option.

* Removing hardcoded interpolation and sizes from the scripts.

* Fixing linter.

* Incorporating feedback from code review.

* Updated classification reference script to use torch.cuda.amp

* Assigned scaler to None if amp is False

* Fixed linter errors

* add ufmt as code formatter

* cleanup

* quote ufmt requirement

* split imports into more groups

* regenerate circleci config

* fix CI

* clarify local testing utils section

* use ufmt pre-commit hook

* split relative imports into local category

* Revert "split relative imports into local category"

This reverts commit f2e224cde2008c56c9347c1f69746d39065cdd51.

* pin black and usort dependencies

* fix local test utils detection

* fix ufmt rev

* add reference utils to local category

* fix usort config

* remove custom categories sorting

* Run pre-commit without fixing flake8

* got a double import in merge
Co-authored-by: Nicolas Hug <nicolashug@fb.com>

* initial code

* add SqueezeExcitation

* initial code

* add SqueezeExcitation

* add SqueezeExcitation

* regnet blocks, stems and model definition

* nit

* add fc layer

* use Callable instead of Enum for block, stem and activation

* add regnet_x and regnet_y model build functions, add docs

* remove unused depth

* use BN/activation constructor and ConvBNActivation

* add expected test pkl files

* allow custom activation in SqueezeExcitation

* use ReLU as the default activation

* initial code

* add SqueezeExcitation

* initial code

* add SqueezeExcitation

* add SqueezeExcitation

* regnet blocks, stems and model definition

* nit

* add fc layer

* use Callable instead of Enum for block, stem and activation

* add regnet_x and regnet_y model build functions, add docs

* remove unused depth

* use BN/activation constructor and ConvBNActivation

* reuse SqueezeExcitation from efficientnet

* refactor RegNetParams into BlockParams

* use nn.init, replace np with torch

* update README

* construct model with stem, block, classifier instances

* Revert "construct model with stem, block, classifier instances"

This reverts commit 850f5f3ed01a2a9b36fcbf8405afd6e41d2e58ef.

* remove unused blocks

* support scaled model

* fuse into ConvBNActivation

* make reset_parameters private

* fix type errors

* fix for unit test

* add pretrained weights for 6 variant models, update docs

* Replaced ToTensor with a combination of PILToTensor and ConvertImageDtype

* Pass dtype