MMSegmentation v0.25.0 Release

What's Changed

Highlights

• Support PyTorch backend on MLU (1515)

Bug Fixes

• Fix the error of BCE loss when batch size is 1 (1629)
• Fix bug of resize function when align_corners is True (1592)
• Fix Dockerfile to run demo script in docker container (1568)
• Correct inference_demo.ipynb path (1576)
• Fix the build_segmentor in colab demo (1551)
• Fix md2yml script (1633, 1555)
• Fix main line link in MAE README.md (1556)
• Fix fastfcn crop_size in README.md by (1597)
• Pip upgrade when testing windows platform (1610)

Improvements

• Delete DS_Store file (1549)
• Revise owners.yml (1621, 1534)

Documentation

• Rewrite the installation guidance (1630)
• Format readme (1635)
• Replace markdownlint with mdformat to avoid ruby installation (1591)
• Add explanation and usage instructions for data configuration (1548)
• Configure Myst-parser to parse anchor tag (1589)
• Update QR code and link for QQ group (1598, 1574)

Contributors

• @atinfinity made their first contribution in #1568
• @DoubleChuang made their first contribution in #1576
• @alpha-baymax made their first contribution in #1515
• @274869388 made their first contribution in #1629

Full Changelog: v0.24.1...v0.25.0

MMSegmentation v0.24.1 Release

What's Changed

• Fix LayerDecayOptimizerConstructor for MAE training (#1539, #1540)

Full Changelog: v0.24.0...v0.24.1

MMSegmentation v0.24.0 Release

What's Changed

Highlights

• Support MAE: Masked Autoencoders Are Scalable Vision Learners
• Support Resnet strikes back

New Features

• Support MAE: Masked Autoencoders Are Scalable Vision Learners (1307, 1523)
• Support Resnet strikes back (1390)
• Support extra dataloader settings in configs (1435)

Bug Fixes

• Fix input previous results for the last cascade_decode_head (#1450)
• Fix validation loss logging (#1494)
• Fix the bug in binary_cross_entropy (1527)
• Support single channel prediction for Binary Cross Entropy Loss (#1454)
• Fix potential bugs in accuracy.py (1496)
• Avoid converting label ids twice by label map during evaluation (1417)
• Fix bug about label_map (1445)
• Fix image save path bug in Windows (1423)
• Fix MMSegmentation Colab demo (1501, 1452)
• Migrate azure blob for beit checkpoints (1503)
• Fix bug in tools/analyse_logs.py caused by wrong plot_iter in some cases (1428)

Improvements

• Merge BEiT and ConvNext's LR decay optimizer constructors (#1438)
• Register optimizer constructor with mmseg (#1456)
• Refactor transformer encode layer in ViT and BEiT backbone (#1481)
• Add build_pos_embed and build_layers for BEiT (1517)
• Add with_cp to mit and vit (1431)
• Fix inconsistent dtype of seg_label in stdc decode (1463)
• Delete random seed for training in dist_train.sh (1519)
• Revise high workers_per_gpus in config file (#1506)
• Add GPG keys and del mmcv version in Dockerfile (1534)
• Update checkpoint for model in deeplabv3plus (#1487)
• Add DistSamplerSeedHook to set epoch number to dataloader when runner is EpochBasedRunner (1449)
• Provide URLs of Swin Transformer pretrained models (1389)
• Updating Dockerfiles From Docker Directory and get_started.md to reach latest stable version of Python, PyTorch and MMCV (1446)

Documentation

• Add more clearly statement of CPU training/inference (1518)

New Contributors

• @jiangyitong made their first contribution in #1431
• @kahkeng made their first contribution in #1447
• @Nourollah made their first contribution in #1446
• @androbaza made their first contribution in #1452
• @Yzichen made their first contribution in #1445
• @whu-pzhang made their first contribution in #1423
• @panfeng-hover made their first contribution in #1417
• @Johnson-Wang made their first contribution in #1496
• @jere357 made their first contribution in #1460
• @mfernezir made their first contribution in #1494
• @donglixp made their first contribution in #1503
• @YuanLiuuuuuu made their first contribution in #1307
• @Dawn-bin made their first contribution in #1527

Full Changelog: v0.23.0...v0.24.0

MMSegmentation v0.23.0 Release

What's Changed

Highlights

• Support BEiT: BERT Pre-Training of Image Transformers
• Support K-Net: Towards Unified Image Segmentation
• Add avg_non_ignore of CELoss to support average loss over non-ignored elements
• Support dataset initialization with file client

New Features

• Support BEiT: BERT Pre-Training of Image Transformers (#1404)
• Support K-Net: Towards Unified Image Segmentation (#1289)
• Support dataset initialization with file client (#1402)
• Add class name function for STARE datasets (#1376)
• Support different seeds on different ranks when distributed training (#1362)
• Add nlc2nchw2nlc and nchw2nlc2nchw to simplify tensor with different dimension operation (#1249)

Improvements

• Synchronize random seed for distributed sampler (#1411)
• Add script and documentation for multi-machine distributed training (#1383)

Bug Fixes

• Add avg_non_ignore of CELoss to support average loss over non-ignored elements (#1409)
• Fix some wrong URLs of models or logs in ./configs (#1336)
• Add title and color theme arguments to plot function in tools/confusion_matrix.py (#1401)
• Fix outdated link in Colab demo (#1392)
• Fix typos (#1424, #1405, #1371, #1366, #1363)

Documentation

• Add FAQ document (#1420)
• Fix the config name style description in official docs(#1414)

New Contributors

• @kinglintianxia made their first contribution in #1371
• @CCODING04 made their first contribution in #1376
• @mob5566 made their first contribution in #1401
• @xiongnemo made their first contribution in #1392
• @Xiangxu-0103 made their first contribution in #1405

MMSegmentation v0.22.1 Release

Bug Fixes

• Fix the ZeroDivisionError that all pixels in one image is ignored. (#1336)

Improvements

• Provide URLs of STDC, Segmenter and Twins pretrained models (#1272)

MMSegmentation v0.22.0 Release

Highlights

• Support ConvNeXt: A ConvNet for the 2020s. Please use the latest MMClassification (0.21.0) to try it out.
• Support iSAID aerial Dataset.
• Officially Support inference on Windows OS.

New Features

• Support ConvNeXt: A ConvNet for the 2020s. (#1216)
• Support iSAID aerial Dataset. (#1115
• Generating and plotting confusion matrix. (#1301)

Improvements

• Refactor 4 decoder heads (ASPP, FCN, PSP, UPer): Split forward function into _forward_feature and cls_seg. (#1299)
• Add min_size arg in Resize to keep the shape after resize bigger than slide window. (#1318)
• Revise pre-commit-hooks. (#1315)
• Add win-ci. (#1296)

Bug Fixes

• Fix mlp_ratio type in Swin Transformer. (#1274)
• Fix path errors in ./demo . (#1269)
• Fix bug in conversion of potsdam. (#1279)
• Make accuracy take into account ignore_index. (#1259)
• Add Pytorch HardSwish assertion in unit test. (#1294)
• Fix wrong palette value in vaihingen. (#1292)
• Fix the bug that SETR cannot load pretrain. (#1293)
• Update correct In Collection in metafile of each configs. (#1239)
• Upload completed STDC models. (#1332)
• Fix DNLHead exports onnx inference difference type Cast error. (#1161)

Contributors

• @JiaYanhao made their first contribution in #1269
• @andife made their first contribution in #1281
• @SBCV made their first contribution in #1279
• @HJoonKwon made their first contribution in #1259
• @Tsingularity made their first contribution in #1290
• @Waterman0524 made their first contribution in #1115
• @MeowZheng made their first contribution in #1315
• @linfangjian01 made their first contribution in #1318

MMSegmentation v0.21.1 Release

Bug Fixes

• Fix repeating log by setup_multi_processes. (#1267)
• Fix typos in docs. (#1263)
• Upgrade isort in pre-commit hook. (#1270)

Improvements

• Use MMCV load_state_dict function in ViT/Swin. (#1272)
• Add exception for PointRend for support CPU-only. (#1271)

New Contributors

• @RangeKing made their first contribution in #1263

MMSegmentation v0.21.0 Release

Highlights

• Officially Support CPUs training and inference, please use the latest MMCV (1.4.4) to try it out.
• Support Segmenter: Transformer for Semantic Segmentation (ICCV'2021).
• Support ISPRS Potsdam and Vaihingen Dataset.
• Add Mosaic transform and MultiImageMixDataset class in dataset_wrappers.

New Features

• Support Segmenter: Transformer for Semantic Segmentation (ICCV'2021) (#955)
• Support ISPRS Potsdam and Vaihingen Dataset (#1097, #1171)
• Add segformer‘s benchmark on cityscapes (#1155)
• Add auto resume (#1172)
• Add Mosaic transform and MultiImageMixDataset class in dataset_wrappers (#1093, #1105)
• Add log collector (#1175)

Improvements

• New-style CPU training and inference (#1251)
• Add UNet benchmark with multiple losses supervision (#1143)

Bug Fixes

• Fix the model statistics in doc for readthedoc (#1153)
• Set random seed for palette if not given (#1152)
• Add COCOStuffDataset in class_names.py (#1222)
• Fix bug in non-distributed multi-gpu training/testing (#1247)
• Delete unnecessary lines of STDCHead (#1231)

New Contributors

• @jbwang1997 made their first contribution in #1152
• @BeaverCC made their first contribution in #1206
• @Echo-minn made their first contribution in #1214
• @rstrudel made their first contribution in #955

MMSegmentation v0.20.2 Release

What's Changed

• [Fix] Revise --option to --options in #1140.

Publish this version is to avoid BC-Breaking problem caused by v0.20.1.

Contributors:
@RockeyCoss

MMSegmentation v0.20.1 Release

Improvements

• Change options to cfg-options (#1129)

Bug Fixes

• Fix <!-- [ABSTRACT] --> in metafile. (#1127)
• Fix correct num_classes of HRNet in LoveDA dataset (#1136)

Contributors
@MengzhangLI
@RockeyCoss