GitLab

* allow single-channel-images in draw_bounding_boxes

This is bug-fix for issue #4042.

https://github.com/pytorch/vision/issues/4042

* change made for passing linting requirements

#4042

Rewrite draw_segmentation_masks and update gallery example to illustrate both instance and semantic segmentation models (#3824)

* start adding tests

* add return type and doc

* adds tests

* add no fill tests

* add rgb test

* check inplace

* bug fix

* bug fix

* rewrite make grid

* add plotting demos

* rename file

* remove

* updt

* Add viz

* updt

* update readme, add links

* complte bounding boxes

* Complete the examples!

* link fix

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

* add draw segm masks

* rewrites with new api

* fix flaky colors

* fix resize bug

* resize for sanity

* cleanup

* project the image

* Minor refactor to adopt num classes

* add uint8 in docstring

* adds alpha and docstring

* move code a bit down

* Minor fix

* fix type check

* Fixing resize bug.

* Fix type of alpha.

* Remove unnecessary RGBA conversions.

* update docs to supported only rgb

* minor edits

* adds tests

* shifts masks up

* change tests and impelementation for bool

* change mode to L

* convert to float

* fixes docs
Co-authored-by: Vasilis Vryniotis <datumbox@users.noreply.github.com>
Co-authored-by: Vasilis Vryniotis <vvryniotis@fb.com>
Co-authored-by: Francisco Massa <fvsmassa@gmail.com>

* 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>

* 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>

* Fixes utils

* don't use any

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

* Fix: Improve the bounding boxes implementation

Use write_png instead of PIL in test_draw_boxes()
Initialize txt_font only once

* Remove channels permutation in test_draw_boxes
Co-authored-by: Vasilis Vryniotis <datumbox@users.noreply.github.com>
Co-authored-by: Aditya Khursale <akhursale@nvidia.com>
Co-authored-by: Vasilis Vryniotis <datumbox@users.noreply.github.com>

* initital prototype

* flake

* Adds documentation

* minimal working bboxes

* Adds label display

* adds colors :-)

* adds suggestions and fixes CI

* handles image of dim 4

* fixes image handling

* removes dev file

* adds suggested changes

* Updating the API.

* Update test.

* Implementing code review improvements.

* Further refactoring and adding test.

* Replace random to white to reduce size and change font on tests.
Co-authored-by: Vasilis Vryniotis <vvryniotis@fb.com>

* make_grid should not avoid DBZ by adding eps

While this doesn't matter in most cases where the images is quantized to [0,255] afterwards, it still is not a faithful de-normalization. It is a simple change to make it use clamp instead.

* fix max/min shadowing

* add mypy config

* fix syntax error

* fix annotations in torchvision/utils.py

* add mypy type check to CircleCI

* add mypy cache to ignore files

* try fix CI

* ignore flake8 F821 since it interferes with mypy

* add mypy type check to config generator

* explicitly set config files

* type annotations for torchvision/utils.py

* add missing annotation for make_grid

* fix annotation for save_image

* mirror PIL annotation for fp

Summary: Fix docstring formatting issues

Reviewed By: fmassa

Differential Revision: D20736644

fbshipit-source-id: 78f66045cfd4c84cb35ca84a1e1fa6aadcd50642
Co-authored-by: Patrick Labatut <plabatut@fb.com>

* format param added

* lint fixes

* lint fixes

* lint fixes

* Fix bug of changing input tensor in function

The original code will change the image tensor passed in by multiplying it by 255. The PR fixes that.

* fix inplace multiplication in save_image

* fix save_image when height or width == 1

* test_save_image_single_pixel

Fixes https://github.com/pytorch/vision/issues/397

* load densenet tensors onto cpu

* fix uneven padding in make_grid

* revert incorrectly added commit

* DOC fixed typo in make_grid argument name

* DOC document "padding" make_grid argument

* Add documentation for transforms

* document and remove unused imports in mnist.py

* document lsun, mscoco datasets

* rest of the datasets documented

* Clean up the documentation in other functions

* Add links for datasets

* Add more documentation

* pep8 fix

* add pad_value argument

* add argument to save_image

* fix space

* requested change of order

* update docs

* clipping values, and replacing transposes with permute

Without clipping, any values larger than 255 will be replaced with
int(v mod 256) by byte(), which results in high freq noise in image.

* replaced clip with clamp