MMClassification Release V0.20.0

Tomorrow is the Chinese new year. Happy new year!

Highlights

• Support K-fold cross-validation. The tutorial will be released later.
• Support HRNet, ConvNeXt, Twins, and EfficientNet.
• Support model conversion from PyTorch to Core-ML by a tool.

New Features

• Support K-fold cross-validation. (#563)
• Support HRNet and add pre-trained models. (#660)
• Support ConvNeXt and add pre-trained models. (#670)
• Support Twins and add pre-trained models. (#642)
• Support EfficientNet and add pre-trained models.(#649)
• Support features_only option in TIMMBackbone. (#668)
• Add conversion script from pytorch to Core-ML model. (#597)

Improvements

• New-style CPU training and inference. (#674)
• Add setup multi-processing both in train and test. (#671)
• Rewrite channel split operation in ShufflenetV2. (#632)
• Deprecate the support for "python setup.py test". (#646)
• Support single-label, softmax, custom eps by asymmetric loss. (#609)
• Save class names in best checkpoint created by evaluation hook. (#641)

Bug Fixes

• Fix potential unexcepted behaviors if metric_options is not specified in multi-label evaluation. (#647)
• Fix API changes in pytorch-grad-cam>=1.3.7. (#656)
• Fix bug which breaks cal_train_time in analyze_logs.py. (#662)

Docs Update

• Update README in configs according to OpenMMLab standard. (#672)
• Update installation guide and README. (#624)

Contributors

A total of 10 developers contributed to this release.

@Ezra-Yu @mzr1996 @rlleshi @WINDSKY45 @shinya7y @Minyus @0x4f5da2 @imyhxy @dreamer121121 @xiefeifeihu

MMClassification Release V0.19.0

Highlights

• The feature extraction function has been enhanced. See #593 for more details.
• Provide the high-acc ResNet-50 training settings from ResNet strikes back.
• Reproduce the training accuracy of T2T-ViT & RegNetX, and provide self-training checkpoints.
• Support DeiT & Conformer backbone and checkpoints.
• Provide a CAM visualization tool based on pytorch-grad-cam, and detailed user guide!

New Features

• Support Precise BN. (#401)
• Add CAM visualization tool. (#577)
• Repeated Aug and Sampler Registry. (#588)
• Add DeiT backbone and checkpoints. (#576)
• Support LAMB optimizer. (#591)
• Implement the conformer backbone. (#494)
• Add the frozen function for Swin Transformer model. (#574)
• Support using checkpoint in Swin Transformer to save memory. (#557)

Improvements

• [Reproduction] Reproduce RegNetX training accuracy. (#587)
• [Reproduction] Reproduce training results of T2T-ViT. (#610)
• [Enhance] Provide high-acc training settings of ResNet. (#572)
• [Enhance] Set a random seed when the user does not set a seed. (#554)
• [Enhance] Added NumClassCheckHook and unit tests. (#559)
• [Enhance] Enhance feature extraction function. (#593)
• [Enhance] Imporve efficiency of precision, recall, f1_score and support. (#595)
• [Enhance] Improve accuracy calculation performance. (#592)
• [Refactor] Refactor analysis_log.py. (#529)
• [Refactor] Use new API of matplotlib to handle blocking input in visualization. (#568)
• [CI] Cancel previous runs that are not completed. (#583)
• [CI] Skip build CI if only configs or docs modification. (#575)

Bug Fixes

• Fix test sampler bug. (#611)
• Try to create a symbolic link, otherwise copy. (#580)
• Fix a bug for multiple output in swin transformer. (#571)

Docs Update

• Update mmcv, torch, cuda version in Dockerfile and docs. (#594)
• Add analysis&misc docs. (#525)
• Fix docs build dependency. (#584)

Contributors

A total of 6 developers contributed to this release.

@elopezz @Ezra-Yu @mzr1996 @0x4f5da2 @fangxu622 @okotaku

MMClassification Release V0.18.0

Highlights

• Support MLP-Mixer backbone and provide pre-trained checkpoints.
• Add a tool to visualize the learning rate curve of the training phase. Welcome to use with the tutorial!

New Features

• Add MLP Mixer Backbone. (#528, #539)
• Support positive weights in BCE. (#516)
• Add a tool to visualize learning rate in each iterations. (#498)

Improvements

• Use CircleCI to do unit tests. (#567)
• Focal loss for single label tasks. (#548)
• Remove useless import_modules_from_string. (#544)
• Rename config files according to the config name standard. (#508)
• Use reset_classifier to remove head of timm backbones. (#534)
• Support passing arguments to loss from head. (#523)
• Refactor Resize transform and add Pad transform. (#506)
• Update mmcv dependency version. (#509)

Bug Fixes

• Fix bug when using ClassBalancedDataset. (#555)
• Fix a bug when using iter-based runner with 'val' workflow. (#542)
• Fix interpolation method checking in Resize. (#547)
• Fix a bug when load checkpoints in mulit-GPUs environment. (#527)
• Fix an error on indexing scalar metrics in analyze_result.py. (#518)
• Fix wrong condition judgment in analyze_logs.py and prevent empty curve. (#510)

Docs Update

• Fix vit config and model broken links. (#564)
• Add abstract and image for every paper. (#546)
• Add mmflow and mim in banner and readme. (#543)
• Add schedule and runtime tutorial docs. (#499)
• Add the top-5 acc in ResNet-CIFAR README. (#531)
• Fix TOC of visualization.md and add example images. (#513)
• Use docs link of other projects and add MMCV docs. (#511)

Contributors

A total of 9 developers contributed to this release.

@Ezra-Yu @LeoXing1996 @mzr1996 @0x4f5da2 @huoshuai-dot @imyhxy @juanjompz @okotaku @xcnick

MMClassification Release V0.17.0

Highlights

• Support Tokens-to-Token ViT backbone and Res2Net backbone. Welcome to use!
• Support ImageNet21k dataset.
• Add a pipeline visualization tool. Try it with the tutorials!

New Features

• Add Tokens-to-Token ViT backbone and converted checkpoints. (#467)
• Add Res2Net backbone and converted weights. (#465)
• Support ImageNet21k dataset. (#461)
• Support seesaw loss. (#500)
• Add a pipeline visualization tool. (#406)
• Add a tool to find broken files. (#482)
• Add a tool to test TorchServe. (#468)

Improvements

• Refator Vision Transformer. (#395)
• Use context manager to reuse matplotlib figures. (#432)

Bug Fixes

• Remove DistSamplerSeedHook if use IterBasedRunner. (#501)
• Set the priority of EvalHook to "LOW" to avoid a bug when using IterBasedRunner. (#488)
• Fix a wrong parameter of get_root_logger in apis/train.py. (#486)
• Fix version check in dataset builder. (#474)

Docs Update

• Add English Colab tutorials and update Chinese Colab tutorials. (#483, #497)
• Add tutuorial for config files. (#487)
• Add model-pages in Model Zoo. (#480)
• Add code-spell pre-commit hook and fix a large mount of typos. (#470)

Contributors

A total of 6 developers contributed to this release.

@mzr1996 @Ezra-Yu @tansor @youqingxiaozhua @0x4f5da2 @okotaku

MMClassification Release V0.16.0

Highlights

• We have improved compatibility with downstream repositories like MMDetection and MMSegmentation. We will add some examples about how to use our backbones in MMDetection.
• Add RepVGG backbone and checkpoints. Welcome to use it!
• Add timm backbones wrapper, now you can simply use backbones of pytorch-image-models in MMClassification!

New Features

• Add RepVGG backbone and checkpoints. (#414)
• Add timm backbones wrapper. (#427)

Improvements

• Fix TnT compatibility and verbose warning. (#436)
• Support setting --out-items in tools/test.py. (#437)
• Add datetime info and saving model using torch<1.6 format. (#439)
• Improve downstream repositories compatibility. (#421)
• Rename the option --options to --cfg-options in some tools. (#425)
• Add PyTorch 1.9 and Python 3.9 build workflow, and remove some CI. (#422)

Bug Fixes

• Fix format error in test.py when metric returns np.ndarray. (#441)
• Fix publish_model bug if no parent of out_file. (#463)
• Fix num_classes bug in pytorch2onnx.py. (#458)
• Fix missing runtime requirement packaging. (#459)
• Fix saving simplified model bug in ONNX export tool. (#438)

Docs Update

• Update getting_started.md and install.md. And rewrite finetune.md. (#466)
• Use PyTorch style docs theme. (#457)
• Update metafile and Readme. (#435)
• Add CITATION.cff. (#428)

Contributors

A total of 8 developers contributed to this release.
@Charlyo @Ezra-Yu @mzr1996 @amirassov @RangiLyu @zhaoxin111 @uniyushu @zhangrui-wolf

MMClassification Release V0.15.0

Highlights

• Support hparams argument in AutoAugment and RandAugment to provide hyperparameters for sub-policies.
• Support custom squeeze channels in SELayer.
• Support classwise weight in losses.

New Features

• Add hparams argument in AutoAugment and RandAugment and some other improvement. (#398)
• Support classwise weight in losses (#388)
• Enhence SELayer to support custom squeeze channels. (#417)

Code Refactor

• Better result visualization (#419)
• Use post_process function to handle pred result processing. (#390)
• Update digit_version function. (#402)
• Avoid albumentations to install both opencv and opencv-headless. (#397)
• Avoid unnecessary listdir when building ImageNet. (#396)
• Use dynamic mmcv download link in TorchServe dockerfile. (#387)

Docs Improvement

• Add readme of some algorithms and update meta yml (#418)
• Add Copyright information. (#413)
• Add PR template and modify issue template (#380)

Contributors

A total of 5 developers contributed to this release.
@azad96 @Ezra-Yu @mzr1996 @mmeendez8 @sovrasov

MMClassification Release V0.14.0

Highlights

• Add transformer-in-transformer backbone and pretrain checkpoints, refers to the paper.
• Add Chinese colab tutorial.
• Provide dockerfile to build mmcls dev docker image.

New Features

• Add transformer in transformer backbone and pretrain checkpoints. (#339)
• Support mim, welcome to use mim to manage your mmcls project. (#376)
• Add Dockerfile. (#365)
• Add ResNeSt configs. (#332)

Improvements

• Use the presistent_works option if available, to accelerate training. (#349)
• Add Chinese ipynb tutorial. (#306)
• Refactor unit tests. (#321)
• Support to test mmdet inference with mmcls backbone. (#343)
• Use zero as default value of thrs in metrics. (#341)

Bug Fixes

• Fix ImageNet dataset annotation file parse bug. (#370)
• Fix docstring typo and init bug in ShuffleNetV1. (#374)
• Use local ATTENTION registry to avoid conflict with other repositories. (#376)
• Fix swin transformer config bug. (#355)
• Fix patch_cfg argument bug in SwinTransformer. (#368)
• Fix duplicate init_weights call in ViT init function. (#373)
• Fix broken _base_ link in a resnet config. (#361)
• Fix vgg-19 model link missing. (#363)

Contributors

A total of 8 developers contributed to this release.

@Ezra-Yu, @HIT-cwh, @Junjun2016, @LXXXXR, @mzr1996, @pvys, @wangruohui, @ZwwWayne

MMClassification Release V0.13.0

New Features

• Support Swin-Transformer backbone and add training configs for Swin-Transformer on ImageNet. (#271)
• Add pretrained model of RegNetX. (#269)
• Support adding custom hooks in the config file. (#305)
• Improve and add Chinese translation of CONTRIBUTING.md and all tools tutorials. (#320)
• Dump config before training. (#282)
• Add torchscript and torchserve deployment tools. (#279, #284)

Improvements

• Improve test tools and add some new tools. (#322)
• Correct MobilenetV3 backbone structure and add pretained models. (#291)
• Refactor PatchEmbed and HybridEmbed as independent components. (#330)
• Refactor mixup and cutmix as Augments to support more funtions. (#278)
• Refactor weights initialization method. (#270, #318, #319)
• Refactor LabelSmoothLoss to support multiple calculation formulas. (#285)

Bug Fixes

• Fix bug for CPU training. (#286)
• Fix missing test data when num_imgs can not be evenly divided by num_gpus. (#299)
• Fix build compatible with pytorch v1.3-1.5. (#301)
• Fix magnitude_std bug in RandAugment. (#309)
• Fix bug when samples_per_gpu is 1. (#311)

MMClassification Release V0.12.0

New Features

• Improve and add Chinese translation of data_pipeline.md and new_modules.md. (#265)
• Build Chinese translation on readthedocs. (#267)
• Add an argument efficientnet_style to RandomResizedCrop and CenterCrop. (#268)

Improvements

• Only allow directory operation when rank==0 when testing. (#258)
• Fix typo in base_head. (#274)
• Update ResNeXt checkpoints. (#283)

Bug Fixes

• Add attribute data.test in MNIST configs. (#264)
• Download CIFAR/MNIST dataset only on rank 0. (#273)
• Fix MMCV version compatibility. (#276)
• Fix CIFAR color channels bug and update checkpoints in model zoo. (#280)

MMClassification Release V0.11.1

New Features

• Add dim argument for GlobalAveragePooling. (#236)
• Add random noise to RandAugment magnitude. (#240)
• Refine new_dataset.md and add Chinese translation of finture.md, new_dataset.md. (#243)

Improvements

• Refactor arguments passing for Heads. (#239)
• Allow more flexible magnitude_range in RandAugment. (#249)
• Inherits MMCV registry so that in the future OpenMMLab repos like MMDet and MMSeg could directly use the backbones supported in MMCls. (#252)

Bug Fixes

• Fix typo in analyze_results.py. (#237)
• Fix typo in unittests. (#238)
• Check if specified tmpdir exists when testing to avoid deleting existing data. (#242; #258)
• Add missing config files in MANIFEST.in. (#250; #255)
• Use temporary directory under shared directory to collect results to avoid unavailability of temporary directory for multi-node testing. (#251)