Skip to content

Releases: open-mmlab/mmsegmentation

MMSegmentation v1.2.2 release

14 Dec 05:04
@xiexinch xiexinch
v1.2.2
c685fe6
This commit was created on GitHub.com and signed with GitHub’s verified signature. The key has expired.
GPG key ID: 4AEE18F83AFDEB23
Expired
Learn about vigilant mode.
Compare
Choose a tag to compare
MMSegmentation v1.2.2 release Latest
Latest

v1.2.2 (12/14/2023)

Bug Fixes

  • Fix bug in cross entropy loss (#3457)
  • Allow custom visualizer (#3455)
  • test resize with pad_shape (#3421)
  • add with-labels args to inferencer for visualization without labels (#3466)

New Contributors

Contributors

okotaku
Assets 2

MMSegmentation v1.2.1 release

17 Oct 11:51
@xiexinch xiexinch
v1.2.1
cbf9af1
This commit was created on GitHub.com and signed with GitHub’s verified signature. The key has expired.
GPG key ID: 4AEE18F83AFDEB23
Expired
Learn about vigilant mode.
Compare
Choose a tag to compare
MMSegmentation v1.2.1 release

v1.2.1 (10/17/2023)

Bug Fixes

  • Add bpe_simple_vocab_16e6.txt.gz to release (#3386)
  • Fix init api (#3388)
Assets 2

MMSegmentation v1.2.0 release

12 Oct 14:27
@xiexinch xiexinch
v1.2.0
61e0253
This commit was created on GitHub.com and signed with GitHub’s verified signature. The key has expired.
GPG key ID: 4AEE18F83AFDEB23
Expired
Learn about vigilant mode.
Compare
Choose a tag to compare
MMSegmentation v1.2.0 release

MMSegmentation v1.2.0 (10/12/2023)

From v1.1.0 to v1.2.0, we are delighted that MMSegmentation supports full-flow open-vocabulary semantic segmentation and monocular depth estimation tasks!

  • Open-vocabulary semantic segmentation: SAN and CAT-Seg:

    • Side Adapter Network for Open-Vocabulary Semantic Segmentation
      SAN
    • Cost Aggregation for Open-Vocabulary Semantic Segmentation
      CAT-Seg

  • Monocular depth estimation: VPD and AdaBins:

    • Visual Perception with Pre-trained Diffusion Models
      VPD
    • AdaBins: Depth Estimation using Adaptive Bins:
      Adabins

Features

  • Support Side Adapter Network (#3232)

Bug Fixes

  • fix wrong variables passing for set_dataset_meta (#3348)

Documentation

  • add documentation of Finetune ONNX Models (MMSegemetation) Inference for NVIDIA Jetson (#3372)
Assets 2

MMSegmentation v1.1.2 release

20 Sep 05:51
@xiexinch xiexinch
v1.1.2
1471d1e
This commit was created on GitHub.com and signed with GitHub’s verified signature. The key has expired.
GPG key ID: 4AEE18F83AFDEB23
Expired
Learn about vigilant mode.
Compare
Choose a tag to compare
MMSegmentation v1.1.2 release

v1.1.2(09/20/2023)

Features

  • Add semantic label to the segmentation visualization results (#3229)
  • Support NYU depth estimation dataset (#3269)
  • Support Kullback-Leibler divergence Loss (#3242)
  • Support depth metrics (#3297)
  • Support Remote sensing inferencer (#3131)
  • Support VPD Depth Estimator ((#3321)(#3321))
  • Support inference and visualization of VPD (#3331)
  • Support using the pytorch-grad-cam tool to visualize Class Activation Maps (CAM) (#3324)

New projects

  • Support PP-Mobilenet (#3239)
  • Support CAT-Seg (CVPR'2023) (#3098)
  • Support Adabins (#3257)
  • Add pp_mobileseg onnx inference script (#3268)

Bug Fixes

  • Fix module PascalContextDataset (#3235)
  • Fix one hot encoding for dice loss (#3237)
  • Fix confusion_matrix.py (#3291)
  • Fix inferencer visualization (#3333)

Documentation

  • Translate doc for docs/zh_cn/user_guides/5_deployment.md (#3281)

New Contributors

Contributors

Ben-Louis, XiandongWang, and 8 other contributors
Assets 2

MMSegmentation v1.1.1 release

24 Jul 07:37
@xiexinch xiexinch
v1.1.1
30a3f94
This commit was created on GitHub.com and signed with GitHub’s verified signature. The key has expired.
GPG key ID: 4AEE18F83AFDEB23
Expired
Learn about vigilant mode.
Compare
Choose a tag to compare
MMSegmentation v1.1.1 release

v1.1.1(07/24/2023)

Features

  • Add bdd100K datasets (#3158)
  • Remove batch inference assertion (#3210)

Bug Fixes

  • Fix train map path for coco-stuff164k.py (#3187)
  • Fix mim search error (#3194)
  • Fix SegTTAModel with no attribute '_gt_sem_seg' error (#3152)
  • Fix Albumentations default key mapping mismatch (#3195)

New Contributors

Contributors

OliverGrace, CastleDream, and 3 other contributors
Assets 2

MMSegmentation v1.1.0 release

04 Jul 03:20
@xiexinch xiexinch
v1.1.0
0079076
Compare
Choose a tag to compare
MMSegmentation v1.1.0 release

v1.1.0(07/04/2023)

Features

  • Support albu transform (#2943)
  • Support DDRNet (#2855)
  • Add GDAL backend and Support LEVIR-CD Dataset (#2903)
  • Support DSDL Dataset (#2925)
  • huasdorff distance loss (#2820)

New Projects

  • Support SAM inferencer (#2897)
  • Added a supported for Visual Attention Network (VAN) (#2987)
  • add GID dataset (#3038)
  • add Medical semantic seg dataset: Bactteria (#2568)
  • add Medical semantic seg dataset: Vampire (#2633)
  • add Medical semantic seg dataset: Ravir (#2635)
  • add Medical semantic seg dataset: Cranium (#2675)
  • add Medical semantic seg dataset: bccs (#2861)
  • add Medical semantic seg dataset: Gamma Task3 dataset (#2695)
  • add Medical semantic seg dataset: consep (#2724)
  • add Medical semantic seg dataset: breast_cancer_cell_seg dataset (#2726)
  • add Medical semantic seg dataset: chest_image_pneum dataset (#2727)
  • add Medical semantic seg dataset: conic2022 (#2725)
  • add Medical semantic seg dataset: dr_hagis (#2729)
  • add Medical semantic seg dataset: orvs (#2728)
  • add Medical semantic seg dataset: ISIC-2016 Task1 (#2708)
  • add Medical semantic seg dataset: ISIC-2017 Task1 (#2709)
  • add Medical semantic seg dataset: Kvasir seg (#2677)
  • add Medical semantic seg dataset: Kvasir seg aliyun (#2678)
  • add Medical semantic seg dataset: Rite (#2680)
  • add Medical semantic seg dataset: Fusc2021 (#2682)
  • add Medical semantic seg dataset: 2pm vessel (#2685)
  • add Medical semantic seg dataset: Pcam (#2684)
  • add Medical semantic seg dataset: Pannuke (#2683)
  • add Medical semantic seg dataset: Covid 19 ct cxr (#2688)
  • add Medical semantic seg dataset: Crass (#2690)
  • add Medical semantic seg dataset: Chest x ray images with pneumothorax masks (#2687)

Enhancement

  • Robust mapping from image path to seg map path (#3091)
  • Change assertion logic inference cfg.model.test_cfg (#3012)
  • Refactor dice loss (#3002)
  • Update Dockerfile libgl1-mesa-dev (#3095)
  • Prevent passed ann_file from silently failing to load (#2966)
  • Update the translation of models documentation (#2833)
  • Add docs contents at README.md (#3083)
  • Enhance swin pretrained model loading (#3097)

Bug Fixes

  • Handle case where device is neither CPU nor CUDA in HamHead (#2868)
  • Fix bugs when out_channels==1 (#2911)
  • Fix binary C=1 focal loss & dataset fileio (#2935)
  • Fix isaid dataset pre-processing tool (#3010)
  • Fix bug cannot use both '--tta' and '--out' while testing (#3067)
  • Fix inferencer ut (#3117)
  • Fix document (#2863, #2896, #2919, #2951, #2970, #2961, #3042, )
  • Fix squeeze error when N=1 and C=1 (#2933)

New Contributors

Contributors

haruishi43, wangruohui, and 17 other contributors
Assets 2

MMSegmentation v1.0.0 release

06 Apr 09:34
@xiexinch xiexinch
v1.0.0
b600f7c
Compare
Choose a tag to compare
MMSegmentation v1.0.0 release

v1.0.0(04/06/2023)

Highlights

We are excited to announce the release of MMSegmentation v1.0.0 as a part of the OpenMMLab 2.0 project! MMSegmentation v1.0.0 introduces an updated framework structure for the core package and a new section called "Projects". This section showcases a range of engaging and versatile applications built upon the MMSegmentation foundation.

mmseg_release drawio

In this latest release, we have significantly refactored the core package's code to make it clearer, more comprehensible, and disentangled. This has resulted in improved performance for several existing algorithms, ensuring that they now outperform their previous versions. Additionally, we have incorporated some cutting-edge algorithms, such as PIDNet and SegNeXt, to further enhance the capabilities of MMSegmentation and provide users with a more comprehensive and powerful toolkit.
The new "Projects" section serves as an essential addition to MMSegmentation, created to foster innovation and collaboration among users.

Exciting Features

Inferencer

In this release, we introduce the MMSegInferencer, a versatile API for inference that accommodates multiple input types. The API enables users to easily specify and customize semantic segmentation models, streamlining the process of performing semantic segmentation with MMSegmentation.

Usage:

python demo/image_demo_with_inferencer.py ${IMAGE} ${MODEL} --show --device ${DEVICE}

Optimizations

In addition to new features, MMSegmentation v1.0.0 delivers key optimizations for an enhanced user experience.

PyTorch 2.0 Compatibility

MMSegmentation v1.0.0 is now compatible with PyTorch 2.0, ensuring that users can leverage the latest features and performance improvements offered by the PyTorch 2.0 framework when using MMSegmentation. With the integration of inductor, users can expect faster model speeds. The table below shows several example models:

Model Training Speed
pspnet_r50-d8 34.0% ⬆️ (0.3474 -> 0.2293)
segformer_mit-b2 7.12% ⬆️ (0.1798 -> 0.1670)

New Features

New features from v1.0.0rc6 to v1.0.0 include:

  • Add Mapillary Vistas Datasets support to MMSegmentation Core Package (#2576)
  • Support PIDNet (#2609)
  • Support SegNeXt (#2654)
  • Support calculating FLOPs of segmentors (#2706)
  • Support multi-band image for Mosaic (#2748)
  • Support dump segment prediction (#2712)

Bug fix

  • Fix format_result and fix prefix param in cityscape metric, and rename CitysMetric to CityscapesMetric (#2660)
  • Support input gt seg map is not 2D (#2739)
  • Fix accepting an unexpected argument local-rank in PyTorch 2.0 (#2812)

Documentation

New Contributors

Contributors

KKIEEK, liuruiqiang, and 3 other contributors
Assets 2

MMSegmentation v1.0.0rc6 Release

03 Mar 09:29
@xiexinch xiexinch
v1.0.0rc6
c5a4121
Compare
Choose a tag to compare
MMSegmentation v1.0.0rc6 Release Pre-release
Pre-release

v1.0.0rc6(03/03/2023)

Highlights

  • Support MMSegInferencer (#2413, #2658), which is a easy-to-use hight level api for model inference.
>>> from mmseg.apis import MMSegInferencer
>>> # Initialize an inference
>>> inferencer = MMSegInferencer(model='deeplabv3plus_r18-d8_4xb2-80k_cityscapes-512x1024')
>>> # Inference
>>> inferencer('demo/demo.png', show=True)
>>> # Get all models in MMSegmentation
>>> models = MMSegInferencer.list_models('mmseg')
  • Support REFUGE dataset (#2554)

Features

  • Add browse_dataset.py script in mmsegmentation/tools/ (#2649)
  • Support auto import modules from registry (#2481)
  • Replace numpy ascontiguousarray with torch contiguous to speed-up (#2604)

Bug fix

  • Rename and Fix bug of projects HieraSeg (#2565)
  • Add out_channels in CascadeEncoderDecoder and update OCRNet and MobileNet v2 results (#2656)

Documentation

  • Add dataflow documentation of Chinese version (#2652)
  • Add customized runtime documentation of English version (#2533)
  • Add documentation for visualizing feature map using wandb backend (#2557)
  • Add documentation for benchmark results on NPU (HUAWEI Ascend) (#2569, #2596, #2610)
  • Fix API name error in the migration doc (#2601)
  • Refine projects documentation (#2586)
  • Refine MMSegmentation documentation (#2668, #2659)

New Contributors

Contributors

liuruiqiang, zccjjj, and 2 other contributors
Assets 2

MMSegmentation v1.0.0rc5 Release

01 Feb 13:53
@xiexinch xiexinch
v1.0.0rc5
a35e1c4
Compare
Choose a tag to compare
MMSegmentation v1.0.0rc5 Release Pre-release
Pre-release

v1.0.0rc5(02/01/2023)

Bug fix

  • Fix MaskFormer and Mask2Former when install mmdet from source (#2532)
  • Support new fileio interface in MMCV>=2.0.0rc4 (#2543)
  • Fix ERFNet URL in dev-1.x branch (#2537)
  • Fix misleading List[Tensor] types (#2546)
  • Rename typing.py to typing_utils.py (#2548)

New Contributors

Contributors

zccjjj
Assets 2

MMSegmentation v1.0.0rc4 Release

30 Jan 10:22
@xiexinch xiexinch
v1.0.0rc4
1c836be
Compare
Choose a tag to compare
MMSegmentation v1.0.0rc4 Release Pre-release
Pre-release

v1.0.0rc4(30/01/2023)

Highlights

  • Support ISNet (ICCV'2021) in projects (#2400)
  • Support HSSN (CVPR'2022) in projects (#2444)

Features

  • Add Gaussian Noise and Blur for biomedical data (#2373)
  • Add BioMedicalRandomGamma (#2406)
  • Add BioMedical3DPad (#2383)
  • Add BioMedical3DRandomFlip (#2404)
  • Add gt_edge_map field to SegDataSample (#2466)
  • Support synapse dataset (#2432, #2465)
  • Support Mapillary Vistas Dataset in projects (#2484)
  • Switch order of reduce_zero_label and applying label_map (#2517)

Documentation

  • Add ZN Customized_runtime Doc (#2502)
  • Add EN datasets.md (#2464)
  • Fix minor typo in migration package.md (#2518)

Bug fix

  • Fix incorrect img_shape value assignment in RandomCrop (#2469)
  • Fix inference api and support setting palette to SegLocalVisualizer (#2475)
  • Unfinished label conversion from -1 to 255 (#2516)

New Contributors

Contributors

siddancha, blueyo0, and 7 other contributors
Assets 2