* add GN support for flops computation

* remove useless lines

* modify the flops computation for gn

* async inference support

* implemented concurrent decorator

* fixes for Python versions < 3.7

* async methods depend on python version

* revert changes in forward method

* async_test -> async_simple_test, debug logging is done via logger.debug

* add async test

* add asynctest to test requirements

* async tests are run in Python 3.7

* check CUDA, add docs

* fix device

* run test only if CUDA is available

* fix linting

* custom operators can run on nondefault stream

* set current stream in kernel launch configuration

* example fixes

* add async/sync interface comparison benchmark

* fix linting

Fix loader for images with no proposals

* collect env info for trouble shoting

* minor fix

* update the issue template

* fix the travis building

* update setup classifiers

* Remove keep all stage code

* remove keep_all_stage in config

* update to v1.0rc1

* update maintainers

* add multi nodes distributed test support

* fix bug in htc.py when keep_all_stages turn on

* remove package imported but unused in test.py

* reformat code in test.py

* support both cpu & gpu for gathering

* reformat

* clean code, add doc

* add docstring

* reformat doc string

* fix scale factor bug

* reformat

* clean code

* fix oom when there are too many gts

* add gpu_assign_thr

* add gpu assign thr to approx max iou assigner

* upgrade yapf format

* reformat for yapf0.29

* fix conflicts of yapf and flake8

* flake 8 is prior than yapf

* Fix issues on no ignore boxes specified

* add check during loading

* change back

* change transform.py back

* fix box_filed bug & reformat

* use append rather than extend

* add FreeAnchorRetinanet

* fix std

* refactor code

* add cfg

* fix bug in cfg

* add annotation and clean

* add readme

* fix isort error

* replace clip with torch.clamp

* fix init typo

* refactoring

* clean

* update benchmark

* rename and delete some var

* minor fix

Change pad_d to pad_w at line 821 in deform_conv_cuda_kernel.cu, which is a typo.

cache pip,
use 'before script' as it done in mmcv,
see https://github.com/open-mmlab/mmdetection/issues/1658

* Clip proposal boxes to valid range in single_mask_target

* Fix the clipping value in mask_target_single

* Fix simple_test for base and fast_rcnn detector

* Add test of detector forward methods

* Revert forward_test change and add docs indicating expected inputs for forward

* fix linter

* Update GETTING_STARTED.md

Found out that this method is required. The torch.nn.Module doesn't contain this method thus you need to implement it

* update docstring 

see https://github.com/open-mmlab/mmdetection/issues/1582

* correct for flake8

* flake8: trailing space has removed

* prevent silent skip of validation #1594

see https://github.com/open-mmlab/mmdetection/issues/1594

* correct for flake8

* Add tests for NMS

* Fix linting errors

* Add DeviceGaurd to nms_cuda

Allows to use not only 0.5 probability.

* Fix bug: index out of range when num_extra < len(indice)

* Fix bug: index out of range when num_extra < len(indice)

* formatter code

Add print of area range to each metric table.
Now metrics are looking better.

Found out that this method is required. The torch.nn.Module doesn't contain this method thus you need to implement it

* also support to view ap

* change string format

* eval class_wise in coco_eval

* reformat

* class_wise API from detectron

* reformat

* change code source

* reformat, use terminaltable