GitLab

* Adding TODO placeholders.

* More placeholders.

* Add MobileNetV3 small pre-trained weights.

* Remove placeholders.

* aspect ratio must be a sampling from log scale.
reference from: https://github.com/pytorch/vision/blob/8317295c1d272e0ba7b2ce31e3fd2c048235fc73/torchvision/transforms/transforms.py#L833-L836



* add random erasing unittest code

* Increased threshold for difference in sampling rate

* move outside of the loop

* log_ratio move outside of the loop. RandomResizedCrop also
Co-authored-by: Vasilis Vryniotis <datumbox@users.noreply.github.com>

* fix STL10 folds

* use np.int64 over np.long

* [android] android gradle project for ops

* Change CMakeLists to latest PyTorch

* Use mobilenet_v3 models for detection

Don't need to have two variants of the model anymore, but I'm not removing it for now

* Fix orientation when angle = 0

* [android][test_app] Fix YUV decoding

* Use smaller version of mobilenet model

* Divide inputs by 255 again

* [android] assets mobilenetv3
Co-authored-by: Francisco Massa <fvsmassa@gmail.com>

Summary:
We accept python bytes for video in existing API, such as `_read_video_timestamps_from_memory`.
We should do so to `_read_video_timestamps_from_memory` as well.

Reviewed By: fmassa

Differential Revision: D26216777

fbshipit-source-id: 5d7ecf15c5a65dc5c57f17b343e7c63b4933189e
Co-authored-by: Zhicheng Yan <zyan3@fb.com>

* Refactoring mobilenetv3 to make code reusable.

* Adding quantizable MobileNetV3 architecture.

* Fix bug on reference script.

* Moving documentation of quantized models in the right place.

* Update documentation.

* Workaround for loading correct weights of quant model.

* Update weight URL and readme.

* Adding eval.

* Check if the file is located on Google Drive

download_file() detects if the provided URL contains a link to Google
Drive and passes the request through the download_file_from_google_drive()
if it is the case.

* remove lazy re import

* add guard clause

* use urlparse

* add tests

* Fixing lint.

* make id matching more safe

* Fixing type check.

* Remove AnyStr.
Co-authored-by: Philip Meier <github.pmeier@posteo.de>
Co-authored-by: Vasilis Vryniotis <vvryniotis@fb.com>

* Renamed original method to test center crop

* Added test method, docs and added padding when imgsize < cropsize.

* BugFix - keep odd_crop_size odd

* Do not crop when image size after padding matches crop size; updated test.
Co-authored-by: Vasilis Vryniotis <datumbox@users.noreply.github.com>

* Fixing mypy errors.

* Fixing typing issue.

* adds fill paramter

* small doc edit

* Change fill to bool

* add filled

* fix the bugs

* Fixes bugs

* adds test with fill param

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

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

Disable pretrained backbone downloading if pretrained is True
Co-authored-by: Vasilis Vryniotis <datumbox@users.noreply.github.com>

* Document undodcumented parameters

* remove setup.cfg changes

* Properly pass normalize down instead of deprecating it

* Fix flake8

* Add new CI check

* Fix type spec

* Leave normalize be part of kwargs
Co-authored-by: Vasilis Vryniotis <datumbox@users.noreply.github.com>

* initial fix

* fill=0

* docstrings

* fill type check

* fill type check

* set None to zero

* unit tests

* set instead of NotImplemented

* fix W293
Co-authored-by: vfdev <vfdev.5@gmail.com>
Co-authored-by: Vasilis Vryniotis <datumbox@users.noreply.github.com>

* Moving jpegcommon inside cpu implementation

* Adding namespaces on image and moving private methods to anonymous.

* Fixing headers.

* Renaming public image methods to match the ones on python.

* Refactoring to remove the double ifs in common_jpeg.h

* Moving registration of video methods in Video.cpp and removing unnecessary includes.

* Rename files according to cpp styles.

* Adding namespaces and moving private methods to anonymous namespaces.

* Syncing method names.

* Fixing minor issues.

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

* Adding the note to the backend function

* addressing the new master change
Co-authored-by: Francisco Massa <fvsmassa@gmail.com>

* Making _segm_resnet() generic and reusable.

* Adding fcn and deeplabv3 directly on mobilenetv3 backbone.

* Adding tests for segmentation models.

* Rename is_strided with _is_cn.

* Add dilation support on MobileNetV3 for Segmentation.

* Add Lite R-ASPP with MobileNetV3 backbone.

* Add pretrained model weights.

* Removing model fcn_mobilenet_v3_large.

* Adding docs and imports.

* Fixing typo and readme.

Removes deprecation warning

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

Summary:
Upgrading service to platform009 enforces tls1.3 authentication.
Specify tls cert file to decoder through config.
Context: https://fb.workplace.com/groups/2020511478062223

Differential Revision: D26026259

fbshipit-source-id: e89f88ff7e41ec6b6a62e85d9015d78f6377aa65

The video reader suppressed code 61 (ENODATA on Linux) that occured at
the end of a stream and logged an INFO message. On macOS ENODATA has a
different code (96), which resulted in an error message being logged.

This small patch replaces the hard-coded value with the ENODATA constant
to suppress the macOS warning.
Co-authored-by: Vasilis Vryniotis <datumbox@users.noreply.github.com>

* Fixes utils

* don't use any

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

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

* Fixed warning.

UserWarning: The default behavior for interpolate/upsample with float scale_factor changed in 1.6.0 to align with other frameworks/libraries, and now uses scale_factor directly, instead of relying on the computed output size. If you wish to restore the old behavior, please set recompute_scale_factor=True. See the documentation of nn.Upsample for details.

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

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

* Tag fasterrcnn mobilenetv3 model with 320, add new inference config that makes it 2x faster sacrificing a bit of mAP.

* Add a high resolution fasterrcnn mobilenetv3 model.

* Update tests and expected values.

* Simplify code and remove used vars.

* Simplify expressions and remove used parenthesis.

* Jit fixes.

* Making check more readable.

* fixing styles

* Minor refactoring of a private method to make it reusuable.

* Adding a FasterRCNN + MobileNetV3 with & w/o FPN models.

* Reducing Resolution to 320-640 and anchor sizes to 16-256.

* Increase anchor sizes.

* Adding rpn score threshold param on the train script.

* Adding trainable_backbone_layers param on the train script.

* Adding rpn_score_thresh param directly in fasterrcnn_mobilenet_v3_large_fpn.

* Remove fasterrcnn_mobilenet_v3_large prototype and update expected file.

* Update documentation and adding weights.

* Use buildin Identity.

* Fix spelling.

* Make download_url() follow redirects

Fix bug related to the incorrect processing of redirects.
Follow the redirect chain until the destination is reached or
the number of redirects exceeds the max allowed value (by default 10).

* Parametrize value of max allowed redirect number

Make max number of hops a function argument and assign its default value to 10

* Propagate the max number of hops to download_url()

Add the maximum number of redirect hops parameter to download_url()

* check file existence before redirect

* remove print

* remove recursion

* add tests

* Reducing max_redirect_hops
Co-authored-by: Vasilis Vryniotis <datumbox@users.noreply.github.com>
Co-authored-by: Philip Meier <github.pmeier@posteo.de>
Co-authored-by: Vasilis Vryniotis <vvryniotis@fb.com>

* 'make_dataset' as staticmethod of 'DatasetFolder'

* a better fix
Co-authored-by: Francisco Massa <fvsmassa@gmail.com>

* Add MobileNetV3 Architecture in TorchVision (#3182)

* Adding implementation of network architecture

* Adding rmsprop support on the train.py

* Adding auto-augment and random-erase in the training scripts.

* Adding support for reduced tail on MobileNetV3.

* Tagging blocks with comments.

* Adding documentation, pre-trained model URL and a minor refactoring.

* Handling better untrained supported models.