GitLab

* add bia_init_with_prob

* add test_weight_init

* remove supports for python 2.7

* fix the unit test for python 3.5

* add python 3.8 in CI

* try ubuntu 18.04 as the environment

* add iminvert

* DC now supports dimension other than 3

* add lmdb

* revise for lmdb

* add warmup_byepoch option

* remove not needed files

* add attribute active_loader for runner, revise code

* remove not used attrs

* remove not used attrs

* add docstring, else

* to pass CI

* to pass CI
Co-authored-by: zhaoyue-zephyrus <thuzhaoyue@gmail.com>

* add base for config

* fixed format

* rm terminal width

* support multiple & recursive base

* add test case

* fix format

* add test construct

* minor fix

* add more test, rewrite merge from opt

* avoid depulicate keys

* delete imported config as module

* rename merge_from_dict

* support pathlib in imageio

* support pathlib in imageio

* add Path type for the annotation.
Co-authored-by: xurui4 <xurui4@sensetime.com>

Signed-off-by: lizz <lizz@sensetime.com>

* Update tensorboard.py

* Update tensorboard.py

* add is_scalar in PaviLoggerHook and add unittest

* add docstrings

* update pre-commit hook config

* update the contributing guide

* fix wandb hook

* add pavi logger hook

* modify default value of init_kwargs to None

* add more features to PaviLoggerHook

* rm taskid (will cause bug)

* directly upload saved pth

* fix CI

* register PAVI Hooks

* add logging utils

* install torch (cpu only) for CI

* fix unittests for runner

* remove python 2.7 from testing list

* add a corner case

* add registry and build_from_cfg

* add some corner cases

* add some corner cases

* fix the unittest for python 3.5

* minor fix

* size hint
Signed-off-by: lizz <lizz@sensetime.com>

* add fast resize
Signed-off-by: lixuanyi <lixuanyi@sensetime.com>

* fix
Signed-off-by: lixuanyi <lixuanyi@sensetime.com>

* add tests
Signed-off-by: lixuanyi <lixuanyi@sensetime.com>

* fix
Signed-off-by: lixuanyi <lixuanyi@sensetime.com>

* Update tests/test_image.py

* change format
Signed-off-by: lixuanyi <lixuanyi@sensetime.com>

* add tests
Signed-off-by: lixuanyi <lixuanyi@sensetime.com>

* add docstring
Signed-off-by: lixuanyi <lixuanyi@sensetime.com>
Co-authored-by: lizz <innerlee@users.noreply.github.com>

* add inplace normalization
Signed-off-by: lixuanyi <lixuanyi@sensetime.com>

* fix
Signed-off-by: lixuanyi <lixuanyi@sensetime.com>

* add dtype check

* modify inplace norm
Signed-off-by: lixuanyi <lixuanyi@sensetime.com>

* fix
Signed-off-by: lixuanyi <lixuanyi@sensetime.com>

* fix
Signed-off-by: lixuanyi <lixuanyi@sensetime.com>

* fix line

* add implace flip

* add tests
Signed-off-by: lixuanyi <lixuanyi@sensetime.com>

* modify tests
Signed-off-by: lixuanyi <lixuanyi@sensetime.com>

* add inplace tests
Signed-off-by: lixuanyi <lixuanyi@sensetime.com>

* add comments
Signed-off-by: lixuanyi <lixuanyi@sensetime.com>

* fix
Signed-off-by: lixuanyi <lixuanyi@sensetime.com>

* change image decoding method from cv2 to TurboJPEG
Signed-off-by: lixuanyi <lixuanyi@sensetime.com>

* add tests

* travis
Signed-off-by: lizz <lizz@sensetime.com>

* fix
Signed-off-by: lizz <lizz@sensetime.com>

* more doc
Signed-off-by: lizz <lizz@sensetime.com>

* change per request
Signed-off-by: lizz <lizz@sensetime.com>

* more doc
Signed-off-by: lizz <lizz@sensetime.com>

* baby-sitting
Signed-off-by: lizz <lizz@sensetime.com>
Co-authored-by: lizz <innerlee@users.noreply.github.com>

* Use cv2 to normalize img

* multipy faster than divide

* also do imdenormalize

* support more channels

* even faster
Signed-off-by: lizz <lizz@sensetime.com>

* add test
Signed-off-by: lizz <lizz@sensetime.com>

* even even faster
Signed-off-by: lizz <lizz@sensetime.com>

* fffaster
Signed-off-by: lizz <lizz@sensetime.com>

* okay
Signed-off-by: lizz <lizz@sensetime.com>

* guard against implace modification
Signed-off-by: lizz <lizz@sensetime.com>

* fix imdenormalize
Signed-off-by: lizz <lizz@sensetime.com>

* tidy
Signed-off-by: lizz <lizz@sensetime.com>

* clean
Signed-off-by: lizz <lizz@sensetime.com>

* Update test_image.py

* Update test_image.py

* Update test_image.py

* Update test_image.py

* Update test_image.py

* Update test_image.py

* fix an import error

* fix an import error

* log meta

* add type check

* Fix the missing numpy module for readthedocs

* fix the linting

* Speed up bbox_clip

* Update geometry.py

* Adaptive progress bar length
Signed-off-by: lizz <lizz@sensetime.com>

* it works
Signed-off-by: lizz <lizz@sensetime.com>

* format

* pass test

* :lipstick

* test

* Update test_progressbar.py

* 2.7

* sort import
Signed-off-by: lizz <lizz@sensetime.com>

* try this
Signed-off-by: lizz <lizz@sensetime.com>

Signed-off-by: lizz <lizz@sensetime.com>

* add copyright header

* change # to // for cpp files

* use load_from_state_dict to load ckpt

* reformat

* reformat

* reformat import

* pass flake8

* isort skip

* add dist utils

* rename due to a typo

* bump version to 0.2.15

* fix imports

* add pre-commit hooks

* update yapf version

Signed-off-by: han_hans <han_hans@apple.com>