GitLab

* [Fix] fix config type inconsistency

* [Fix] Fix unit test

* [fix] fix checkpoint local files detect

* [Fix] add support for path like '~/xx/file'

* [Fix] fix some details

* [Fix] fix unittest

* [Fix] Fix the path of ceph containing cluster name

* fix corner cases

* fix format

* add note

* detect detect_anomalous_params

* fix default value

* merge two case

* fix none case

* add unitest

* fix typo

* change level to error

* fix type

* add more details in docstr

* fix tensorrt test failed with pytorch 1.8+

* add comment

* [Docs] Fix the format of docstring

* fix format

* [Docs] Improve the style of pr.md

* fix symlink

* improve style

* fix format

* [Docs] Fix the format of return

* replace List with list

* format the documentation of optimizer

* Update ops docstring (#2

)

* update ops docstring

* fix typos
Co-authored-by: ChaimZhu <zhuchenming@pjlab.org.cn>
Co-authored-by: ChaimZhu <zhuchenming@pjlab.org.cn>

* Add the link of mmhuman3d

* fix typo

* add link of mim

* add link of mim in docs

* improve the table format

* [Impl] cancel previous runs that are not completed

* [Improve] use official method

* [Useless] just for test

* improve: add module dict

* Update __init__.py

* docs: add module dict zh

* docs: add module dict

* Update test_basemodule.py

* Update __init__.py

* lint

* lint

* docs

* docs

* Update base_module.py

* lint

* update title

* fix deploy docs in for readthedocs

* reformat table

* resolve comments

* [Fix] Fix compiled error on windows

* fix typo

* Use the official markdownlint hook

* upgrade the version of ruby

* use circleci to test the unit tests of CUDA

* fix conflict

* change the gpu resources from to small

* Use the official markdownlint hook

* upgrade the version of ruby

* fix conflict

* [Docs] Skip CI when some specific files were changed

* remove a unnecessary environment in lint.yml

* Use paths-ignore to ignore paths instead of paths

* fix typo

* Fix build with cuda9.0

* add note about why do we use torch/types.h

* [Fix] Fix the compiled error on windows due to the c++14 flag

* fix lint

* Add device registry for pytorch ops

* add declaration of CheckDeviceConsistency

* fix for torch130

* assert with torch check

* Refactor ops with dispatch

* update rest ops

* faster install

* update compatibility

* update compatibility, rename parameter

* move cpu implement to pytorch/cpu

* update ops/csrc/README.md

* fix rocm support

* update cn document

* update docs

* list instead of map

* [Docs] Add a script to verify the installation of mmcv-full

* polish documentation

* polish documentation

* print the running and compiling environments

* rename scripts to .dev_scripts

* polish readme

* Update mmcls models in `model_zoo/mmcls.json`

* Update `_process_mmcls_checkpoint` function.

* Add a comment

* [Fix] fix train example

* [Fix] fix a detail in train example and add warning in MMDP

* [Fix] fix docstring

* [Fix] fix docstring

* [Docs] Add the link and introduction of mmflow in README

* remove a blank line

* fix link

* add world_size in if clause

* add explanation

* remove LOCAL_RANK check

* [Fix] Fix torch.load error

* [Fix] Fix torch.load error

* rename _save to _save_ckpt

* add load_url to handle imcompatibility of PyTorch versions

* add unittest for load_url

* fix typo

* print a friendly information when error occurred

* bump version to v1.3.17

* add a document to record the previous versions of PyTorch built for mmcv-full

* polish readme