Skip to content

Latest commit

 

History

History
254 lines (198 loc) · 17.8 KB

changelog.md

File metadata and controls

254 lines (198 loc) · 17.8 KB
Raw

Changelog of v0.x

v0.5.0 (01/20/2023)

Highlights

  • Add BaseInferencer to provide a general inference interface
  • Provide ReduceOnPlateauParamScheduler to adjust learning rate by metric
  • Deprecate support for Python3.6

New Features & Enhancements

  • Deprecate support for Python3.6 by @HAOCHENYE in open-mmlab#863
  • Support non-scalar type metric value by @mzr1996 in open-mmlab#827
  • Remove unnecessary calls and lazily import to speed import performance by @zhouzaida in open-mmlab#837
  • Support ReduceOnPlateauParamScheduler by @LEFTeyex in open-mmlab#819
  • Disable warning of subprocess launched by dataloader by @HAOCHENYE in open-mmlab#870
  • Add BaseInferencer to provide general interface by @HAOCHENYE in open-mmlab#874

Bug Fixes

  • Fix support for Ascend device by @wangjiangben-hw in open-mmlab#847
  • Fix Config cannot parse base config when there is . in tmp path, etc. tmp/a.b/c by @HAOCHENYE in open-mmlab#856
  • Fix unloaded weights will not be initialized when using PretrainedIinit by @HAOCHENYE in open-mmlab#764
  • Fix error package name defined in PKG2PROJECT by @HAOCHENYE in open-mmlab#872

Docs

Contributors

A total of 8 developers contributed to this release. Thanks @LEFTeyex, @RangeKing, @yaqi0510, @Xiangxu-0103, @wangjiangben-hw, @mzr1996, @zhouzaida, @HAOCHENYE.

v0.4.0 (12/28/2022)

Highlights

  • Registry supports importing modules automatically
  • Upgrade the documentation and provide the English documentation
  • Provide ProfileHook to profile the running process

New Features & Enhancements

  • Add conf_path in PetrelBackend by @sunyc11 in open-mmlab#774
  • Support multiple --cfg-options. by @mzr1996 in open-mmlab#759
  • Support passing arguments to OptimWrapper.update_params by @twmht in open-mmlab#796
  • Make get_torchvision_model compatible with torch 1.13 by @HAOCHENYE in open-mmlab#793
  • Support flat_decay_mult and fix bias_decay_mult of depth-wise-conv in DefaultOptimWrapperConstructor by @RangiLyu in open-mmlab#771
  • Registry supports importing modules automatically. by @RangiLyu in open-mmlab#643
  • Add profiler hook functionality by @BayMaxBHL in open-mmlab#768
  • Make TTAModel compatible with FSDP. by @HAOCHENYE in open-mmlab#611

Bug Fixes

  • hub.get_model fails on some MMCls models by @C1rN09 in open-mmlab#784
  • Fix BaseModel.to and BaseDataPreprocessor.to to make them consistent with torch.nn.Module by @C1rN09 in open-mmlab#783
  • Fix creating a new logger at PretrainedInit by @xiexinch in open-mmlab#791
  • Fix ZeroRedundancyOptimizer ambiguous error with param groups when PyTorch < 1.12.0 by @C1rN09 in open-mmlab#818
  • Fix MessageHub set resumed key repeatedly by @HAOCHENYE in open-mmlab#839
  • Add progress argument to load_from_http by @austinmw in open-mmlab#770
  • Ensure metrics is not empty when saving best checkpoint by @zhouzaida in open-mmlab#849

Docs

Contributors

A total of 16 developers contributed to this release. Thanks @BayMaxBHL, @RangeKing, @Xiangxu-0103, @xin-li-67, @twmht, @shanmo, @sunyc11, @lyviva, @austinmw, @xiexinch, @mzr1996, @RangiLyu, @MambaWong, @C1rN09, @zhouzaida, @HAOCHENYE

v0.3.2 (11/24/2022)

New Features & Enhancements

  • Send git errors to subprocess.PIPE by @austinmw in open-mmlab#717
  • Add a common TestRunnerTestCase to build a Runner instance. by @HAOCHENYE in open-mmlab#631
  • Align the log by @HAOCHENYE in open-mmlab#436
  • Log the called order of hooks during training process by @songyuc in open-mmlab#672
  • Support setting eta_min_ratio in CosineAnnealingParamScheduler by @cir7 in open-mmlab#725
  • Enhance compatibility of revert_sync_batchnorm by @HAOCHENYE in open-mmlab#695

Bug Fixes

  • Fix distributed_training.py in examples by @PingHGao in open-mmlab#700
  • Format the log of CheckpointLoader.load_checkpoint by @HAOCHENYE in open-mmlab#685
  • Fix bug of CosineAnnealingParamScheduler by @fangyixiao18 in open-mmlab#735
  • Fix add_graph is not called bug by @shenmishajing in open-mmlab#632
  • Fix .pre-commit-config-zh-cn.yaml pyupgrade-repo github->gitee by @BayMaxBHL in open-mmlab#756

Docs

v0.3.1 (11/09/2022)

Highlights

  • Fix error when saving best checkpoint in ddp-training

New Features & Enhancements

  • Replace print with print_log for those functions called by runner by @HAOCHENYE in open-mmlab#686

Bug Fixes

  • Fix error when saving best checkpoint in ddp-training by @HAOCHENYE in open-mmlab#682

Docs

v0.3.0 (11/02/2022)

New Features & Enhancements

  • Support running on Ascend chip by @wangjiangben-hw in open-mmlab#572
  • Support torch ZeroRedundancyOptimizer by @nijkah in open-mmlab#551
  • Add non-blocking feature to BaseDataPreprocessor by @shenmishajing in open-mmlab#618
  • Add documents for clip_grad, and support clip grad by value. by @HAOCHENYE in open-mmlab#513
  • Add ROCm info when collecting env by @zhouzaida in open-mmlab#633
  • Add a function to mark the deprecated function. by @HAOCHENYE in open-mmlab#609
  • Call register_all_modules in Registry.get() by @HAOCHENYE in open-mmlab#541
  • Deprecate _save_to_state_dict implemented in mmengine by @HAOCHENYE in open-mmlab#610
  • Add ignore_keys in ConcatDataset by @BIGWangYuDong in open-mmlab#556

Docs

Bug Fixes

  • Fix error calculation of eta_min in CosineRestartParamScheduler by @Z-Fran in open-mmlab#639
  • Fix BaseDataPreprocessor.cast_data could not handle string data by @HAOCHENYE in open-mmlab#602
  • Make autocast compatible with mps by @HAOCHENYE in open-mmlab#587
  • Fix error format of log message by @HAOCHENYE in open-mmlab#508
  • Fix error implementation of is_model_wrapper by @HAOCHENYE in open-mmlab#640
  • Fix VisBackend.add_config is not called by @shenmishajing in open-mmlab#613
  • Change strict_load of EMAHook to False by default by @HAOCHENYE in open-mmlab#642
  • Fix open encoding problem of Config in Windows by @sanbuphy in open-mmlab#648
  • Fix the total number of iterations in log is a float number. by @jbwang1997 in open-mmlab#604
  • Fix pip upgrade CI by @HAOCHENYE in open-mmlab#622

New Contributors

v0.2.0 (11/10/2022)

New Features & Enhancements

Docs

Bug Fixes

  • Fix LogProcessor does not smooth loss if the name of loss doesn't start with loss by @liuyanyi in open-mmlab#539
  • Fix failed to enable detect_anomalous_params in MMSeparateDistributedDataParallel by @HAOCHENYE in open-mmlab#588
  • Fix CheckpointHook behavior unexpected if given filename_tmpl argument by @C1rN09 in open-mmlab#518
  • Fix error argument sequence in FSDP by @HAOCHENYE in open-mmlab#520
  • Fix uploading image in wandb backend @okotaku in open-mmlab#510
  • Fix loading state dictionary in EMAHook by @okotaku in open-mmlab#507
  • Fix circle import in EMAHook by @HAOCHENYE in open-mmlab#523
  • Fix unit test could fail caused by MultiProcessTestCase by @HAOCHENYE in open-mmlab#535
  • Remove unnecessary "if statement" in Registry by @MambaWong in open-mmlab#536
  • Fix _save_to_state_dict by @HAOCHENYE in open-mmlab#542
  • Support comparing NumPy array dataset meta in Runner.resume by @HAOCHENYE in open-mmlab#511
  • Use get instead of pop to dump runner_type in build_runner_from_cfg by @nijkah in open-mmlab#549
  • Upgrade pre-commit hooks by @zhouzaida in open-mmlab#576
  • Delete the error comment in registry.md by @vansin in open-mmlab#514
  • Fix Some out-of-date unit tests by @C1rN09 in open-mmlab#586
  • Fix typo in MMFullyShardedDataParallel by @yhna940 in open-mmlab#569
  • Update Github Action CI and CircleCI by @zhouzaida in open-mmlab#512
  • Fix unit test in windows by @HAOCHENYE in open-mmlab#515
  • Fix merge ci & multiprocessing unit test by @HAOCHENYE in open-mmlab#529

New Contributors