* support mlu

* add ut and refine docstring

* merge context

* update unit test

* add docstring

* fix bug in AmpOptimWrapper

* add docstring for backward

* add warning and docstring for accumuate gradient

* fix docstring

* fix docstring

* add params_group method

* fix as comment

* fix as comment

* make default_value of loss_scale to dynamic

* Fix docstring

* decouple should update and should no sync

* rename attribute in OptimWrapper

* fix docstring

* fix comment

* fix comment

* fix as comment

* fix as comment and add unit test

* [Feat]:support customizing evaluator

* fix keyname of determining using default evaluator or not

* add assertion

* fix typo

* Replace auto_scale_lr_cfg to auto_scale_lr

* Update

* add base model, ddp model and unit test

* add unit test

* fix unit test

* fix docstring

* fix cpu unit test

* refine base data preprocessor

* refine base data preprocessor

* refine interface of ddp module

* remove optimizer hook

* add forward

* fix as comment

* fix unit test

* fix as comment

* fix build optimizer wrapper

* rebase main and fix unit test

* stack_batch support stacking ndim tensor, add docstring for merge dict

* fix lint

* fix test loop

* make precision_context effective to data_preprocessor

* fix as comment

* fix as comment

* refine docstring

* change collate_data output typehints

* rename to_rgb to bgr_to_rgb and rgb_to_bgr

* support build basemodel with built DataPreprocessor

* fix as comment

* fix docstring

* Add auto scale lr fucntion

* Update

* Update

* Update

* Update

* Update

* Update

* Update

* Update

* Update

* Update

Co-authored-by: wangjiabao1.vendor <wangjiabao@pjlab.org.cn>

* support custom runner

* change build_runner_from_cfg

* refine docstring

* refine docstring

* Modify val_interval and val_begin to be the attributes of TrainLoop.

* update doc

* fix lint

* type hint

* fix build optimizer wrapper without type

* refine logic

* fix as comment

* fix optim_wrapper config error in docstring and unit test

* refine docstring of build_optim_wrapper

* Support multiple optimizers

* minor refinement

* improve unit tests

* minor fix

* Update unit tests for resuming or saving ckpt for multiple optimizers

* refine docstring

* refine docstring

* fix typo

* update docstring

* refactor the logic to build multiple optimizers

* resolve comments

* ParamSchedulers spports multiple optimizers

* add optimizer_wrapper

* fix comment and docstirng

* fix unit test

* add unit test

* refine docstring

* RuntimeInfoHook supports printing multi learning rates

* resolve comments

* add optimizer_wrapper

* fix mypy

* fix lint

* fix OptimizerWrapperDict docstring and add unit test

* rename OptimizerWrapper to OptimWrapper, OptimWrapperDict inherit OptimWrapper, and fix as comment

* Fix AmpOptimizerWrapper

* rename build_optmizer_wrapper to build_optim_wrapper

* refine optimizer wrapper

* fix AmpOptimWrapper.step, docstring

* resolve confict

* rename DefaultOptimConstructor

* fix as comment

* rename clig grad auguments

* refactor optim_wrapper config

* fix docstring of DefaultOptimWrapperConstructor

fix docstring of DefaultOptimWrapperConstructor

* add get_lr method to OptimWrapper and OptimWrapperDict

* skip some amp unit test

* fix unit test

* fix get_lr, get_momentum docstring

* refactor get_lr, get_momentum, fix as comment

* fix error message

Co-authored-by: zhouzaida <zhouzaida@163.com>

* Support multiple optimizers

* minor refinement

* improve unit tests

* minor fix

* Update unit tests for resuming or saving ckpt for multiple optimizers

* refine docstring

* refine docstring

* fix typo

* update docstring

* refactor the logic to build multiple optimizers

* resolve comments

* ParamSchedulers spports multiple optimizers

* refine docstring

* RuntimeInfoHook supports printing multi learning rates

* resolve comments

* fix typo

* improve exception in call_hook

* refine unit test

* add test_call_hook

* refine

* update docstring and ut

* add the epoch/iter that begins validating

* fix lint

* add property and fix unit test

* minor changes

* fix typos and add unit test

* add unit test about begin

* fix docstring

* [Enhance] Add RuntimeInfoHook to update runtime information.

* move lr to runtime info

* docstring

* resolve comments

* update ut and doc

* refine sync random seed

* cancel seed param in batch-sampler

* Add profiling tools

* fix docstr

* fix docstr

* update

* fix bug

* update

* update

* fix error

* fix mypy

* uodate

* merge main

* fix UT

* move get_max_cuda_memory and set_multi_processing to a public function

* fix lint

* fix lint

* fix lint

* delete _set_multi_processing

* fix error

* rename

* support custom log processor

* supplementary docs

* format code

* [Fix] Fix some bugs in hooks and runner.

* fix markdown

* fix latex formula

* resolve comments

* [Feature] Support EMA and SWA.

* add ema hook

* add avg model ut

* add more unit tests

* resolve comments

* fix warmup ema

* rename

* fix comments

* add assert

* fix typehint

* add comments

* [Fix] Replace torch distributed with mmengine dist module

* minor refinement

* move all_reduce_params to dist.py

* add unit tests

* update unit tests

* fix test_logger.py

* add examples

* [Enhance] Make loop related attributes to be runner's properties.

* move iter and epoch to loop

* resolve comments

* fix runner dump cfg

* convert dict cfg to Config

* [Refactor] Change some methods to static methods

* only change build_dataloader to static method

* [Feature] Support convert epoch-based schedulers to iter-based.

* Support convert and refactor LR and Momentum to mixin.

* Add unit tests

* fix args and add runner ut

* resolve comments

* fix logger hook

* fix to record multi-level logs, log train/a/b to a/b

* fix loggerhook json path

* fix uunit test

* change runner.train_dataloader to runner.train_loop.dataloader

* clean debug code

* refie comments and docstring

* fix unit test

* [Feature] Support collect all registered module info.

* update

* update

* add unit tests

* add to runner

* resolve comments

* MMLogger can call get_current_instance without get_instance, change log path

* fix docstring

* fix docstring and update UT

* Fix runner

* fix docstring and lint

* fix ut below python3.8

* resolve circle import

* [Fix] param_scheduler can not None when training models

* update unit tests

* fix unit tests

* refactor ParamSchedulerHook

* refactor unit tests

* param_schedulers can be an empty list

* [Docs] Refine registry documentation (#186)

* [Docs] Refine registry documentation

* reslove comments

* minor refinement

* Refine Visualizer docs (#177)

* Refine Visualizer docs

* update

* update

* update featmap

* update docs

* update visualizer docs

* [Refactor] Refine LoggerHook (#155)

* rename global accessible and intergration get_sintance and create_instance

* move ManagerMixin to utils

* fix as docstring and seporate get_instance to get_instance and get_current_instance

* fix lint

* fix docstring, rename and move test_global_meta

* rename LogBuffer to HistoryBuffer, rename MessageHub methods, MessageHub support resume

* refine MMLogger timestamp, update unit test

* MMLogger add logger_name arguments

* Fix docstring

* Add LogProcessor and some unit test

* update unit test

* complete LogProcessor unit test

* refine LoggerHook

* solve circle import

* change default logger_name to mmengine

* refactor eta

* Fix docstring comment and unitt test

* Fix with runner

* fix docstring

fix docstring

* fix docstring

* Add by_epoch attribute to LoggerHook and fix docstring

* Please mypy and fix comment

* remove \ in MMLogger

* Fix lint

* roll back pre-commit-hook

* Fix hook unit test

* Fix comments

* remove \t in log and add docstring

* Fix as comment

* should not accept other arguments if corresponding instance has been created

* fix logging ddp file saving

* fix logging ddp file saving

* move log processor to logging

* move log processor to logging

* remove current datalaoder

* fix docstring

* fix unit test

* add learing rate in messagehub

* Support output training/validation/testing message after iterations/epochs

* fix docstring

* Fix IterBasedRunner log string

* Fix IterBasedRunner log string

* Support parse validation loss in log processor

* [Enhancement] Add PolyParamScheduler, PolyMomentum and PolyLR (#188)

* [Enhancement] Add PolyParamScheduler, PolyMomentum and PolyLR

* min_lr -> eta_min, refined docstr

Co-authored-by: Zaida Zhou <58739961+zhouzaida@users.noreply.github.com>
Co-authored-by: Haian Huang(深度眸) <1286304229@qq.com>
Co-authored-by: Tong Gao <gaotongxiao@gmail.com>

* rename global accessible and intergration get_sintance and create_instance

* move ManagerMixin to utils

* fix as docstring and seporate get_instance to get_instance and get_current_instance

* fix lint

* fix docstring, rename and move test_global_meta

* rename LogBuffer to HistoryBuffer, rename MessageHub methods, MessageHub support resume

* refine MMLogger timestamp, update unit test

* MMLogger add logger_name arguments

* Fix docstring

* Add LogProcessor and some unit test

* update unit test

* complete LogProcessor unit test

* refine LoggerHook

* solve circle import

* change default logger_name to mmengine

* refactor eta

* Fix docstring comment and unitt test

* Fix with runner

* fix docstring

fix docstring

* fix docstring

* Add by_epoch attribute to LoggerHook and fix docstring

* Please mypy and fix comment

* remove \ in MMLogger

* Fix lint

* roll back pre-commit-hook

* Fix hook unit test

* Fix comments

* remove \t in log and add docstring

* Fix as comment

* should not accept other arguments if corresponding instance has been created

* fix logging ddp file saving

* fix logging ddp file saving

* move log processor to logging

* move log processor to logging

* remove current datalaoder

* fix docstring

* fix unit test

* add learing rate in messagehub

* Support output training/validation/testing message after iterations/epochs

* fix docstring

* Fix IterBasedRunner log string

* Fix IterBasedRunner log string

* Support parse validation loss in log processor

* full init the dataset when building data loader

* resolve comments

* rename global accessible and intergration get_sintance and create_instance

* move ManagerMixin to utils

* fix as docstring and seporate get_instance to get_instance and get_current_instance

* fix lint

* fix docstring, rename and move test_global_meta

* rename LogBuffer to HistoryBuffer, rename MessageHub methods, MessageHub support resume

* refine MMLogger timestamp, update unit test

* MMLogger add logger_name arguments

* Fix docstring

* change default logger_name to mmengine

* Fix docstring comment and unitt test

* fix docstring

fix docstring

* fix docstring

* Fix lint

* Fix hook unit test

* Fix comments

* should not accept other arguments if corresponding instance has been created

* fix logging ddp file saving

* fix logging ddp file saving

* fix docstring

* fix unit test

* fix docstring as comment

* automatic update iter and epoch in message_hub

* add docstring

* Update comment and docstring

* Fix as comment

* Fix docstring and comment

* refine comments

* [WIP] add inline

* refactor vis module

* [Refactor] according review

* [Fix] fix comment

* fix some error

* Get sub visualizer be Visualizer.get_instance

* fix conflict

* fix lint

* fix unit test

* fix mypy

* fix comment

* fix lint

* update docstr

* update

* update instancedata

* remove replace __mro__ with issubclass

Co-authored-by: PJLAB\huanghaian <1286304229@qq.com>
Co-authored-by: HAOCHENYE <21724054@zju.edu.cn>