- Notifications
You must be signed in to change notification settings - Fork50
IJCAI2020 & IJCV2021 🌇 Unsupervised Scene Adaptation with Memory Regularization in vivo
License
layumi/Seg-Uncertainty
Folders and files
| Name | Name | Last commit message | Last commit date | |
|---|---|---|---|---|
Repository files navigation
In this repo, we provide the code for the two papers, i.e.,
MRNet：Unsupervised Scene Adaptation with Memory Regularization in vivo, IJCAI (2020)
MRNet+Rectifying:Rectifying Pseudo Label Learning via Uncertainty Estimation for Domain Adaptive Semantic Segmentation, IJCV (2021)[中文介绍][Poster]
The original DeepLab link of ucmerced is failed. Please use the following link.
[Google Drive]https://drive.google.com/file/d/1BMTTMCNkV98pjZh_rU0Pp47zeVqF3MEc/view?usp=share_link
[One Drive]https://1drv.ms/u/s!Avx-MJllNj5b3SqR7yurCxTgIUOK?e=A1dq3m
or use
pip install gdownpip install --upgrade gdowngdown 1BMTTMCNkV98pjZh_rU0Pp47zeVqF3MEc[19 Jan 2024] We further apply the uncertainty to compositional image retrieval. The paper is accepted by ICLR'24[code].
[27 Jan 2023] You are welcomed to check our new transformer-based workPiPa, which achieves75.6 mIoU on GTA5->Cityscapes.
[5 Sep 2021] Zheng etal. apply the Uncertainty to domain adaptive reid, and also achieve good performance. "Exploiting Sample Uncertainty for Domain Adaptive Person Re-Identification" Kecheng Zheng, Cuiling Lan, Wenjun Zeng, Zhizheng Zhang, and Zheng-Jun Zha. AAAI 2021
[13 Aug 2021] We release one new method by Adaptive Boosting (AdaBoost) for Domain Adaptation. You may check the project athttps://github.com/layumi/AdaBoost_Seg
- Why KLDivergence is always non-negative (>=0)?
Please check the wikipedia at (https://en.wikipedia.org/wiki/Kullback–Leibler_divergence#Properties) . It provides one good demonstration.
- Why both log_sm and sm are used?
You may check the pytorch doc athttps://pytorch.org/docs/stable/generated/torch.nn.KLDivLoss.html?highlight=nn%20kldivloss#torch.nn.KLDivLoss.I follow the discussion athttps://discuss.pytorch.org/t/kl-divergence-loss/65393
Core code is relatively simple, and could be directly applied to other works.
Memory in vivo:https://github.com/layumi/Seg_Uncertainty/blob/master/trainer_ms.py#L232
Recitfying Pseudo label:https://github.com/layumi/Seg_Uncertainty/blob/master/trainer_ms_variance.py#L166
- Python 3.6
- GPU Memory >= 11G (e.g., GTX2080Ti or GTX1080Ti)
- Pytorch orPaddlepaddle
Download [GTA5] and [Cityscapes] to run the basic code.Alternatively, you could download extra two datasets from [SYNTHIA] and [OxfordRobotCar].
DownloadThe GTA5 Dataset
DownloadThe SYNTHIA Dataset SYNTHIA-RAND-CITYSCAPES (CVPR16)
DownloadThe Cityscapes Dataset
DownloadThe Oxford RobotCar Dataset
The data folder is structured as follows:
├── data/│ ├── Cityscapes/ | | ├── data/| | ├── gtFine/| | ├── leftImg8bit/│ ├── GTA5/| | ├── images/| | ├── labels/| | ├── ...│ ├── synthia/ | | ├── RGB/| | ├── GT/| | ├── Depth/| | ├── ...│ └── Oxford_Robot_ICCV19| | ├── train/| | ├── ...Stage-I:
python train_ms.py --snapshot-dir ./snapshots/SE_GN_batchsize2_1024x512_pp_ms_me0_classbalance7_kl0.1_lr2_drop0.1_seg0.5 --drop 0.1 --warm-up 5000 --batch-size 2 --learning-rate 2e-4 --crop-size 1024,512 --lambda-seg 0.5 --lambda-adv-target1 0.0002 --lambda-adv-target2 0.001 --lambda-me-target 0 --lambda-kl-target 0.1 --norm-style gn --class-balance --only-hard-label 80 --max-value 7 --gpu-ids 0,1 --often-balance --use-se
Generate Pseudo Label:
python generate_plabel_cityscapes.py --restore-from ./snapshots/SE_GN_batchsize2_1024x512_pp_ms_me0_classbalance7_kl0.1_lr2_drop0.1_seg0.5/GTA5_25000.pth
Stage-II (with recitfying pseudo label):
python train_ft.py --snapshot-dir ./snapshots/1280x640_restore_ft_GN_batchsize9_512x256_pp_ms_me0_classbalance7_kl0_lr1_drop0.2_seg0.5_BN_80_255_0.8_Noaug --restore-from ./snapshots/SE_GN_batchsize2_1024x512_pp_ms_me0_classbalance7_kl0.1_lr2_drop0.1_seg0.5/GTA5_25000.pth --drop 0.2 --warm-up 5000 --batch-size 9 --learning-rate 1e-4 --crop-size 512,256 --lambda-seg 0.5 --lambda-adv-target1 0 --lambda-adv-target2 0 --lambda-me-target 0 --lambda-kl-target 0 --norm-style gn --class-balance --only-hard-label 80 --max-value 7 --gpu-ids 0,1,2 --often-balance --use-se --input-size 1280,640 --train_bn --autoaug False
*** If you want to run the code without rectifying pseudo label, please change[this line] to 'from trainer_ms import AD_Trainer', which would apply the conventional pseudo label learning. ***
python evaluate_cityscapes.py --restore-from ./snapshots/1280x640_restore_ft_GN_batchsize9_512x256_pp_ms_me0_classbalance7_kl0_lr1_drop0.2_seg0.5_BN_80_255_0.8_Noaug/GTA5_25000.pth
The trained model is available athttps://drive.google.com/file/d/1smh1sbOutJwhrfK8dk-tNvonc0HLaSsw/view?usp=sharing
- The folder with
SYin name is for SYNTHIA-to-Cityscapes - The folder with
RBin name is for Cityscapes-to-Robot Car
Note that the evaluation code I provided for SYNTHIA-to-Cityscapes is still average the IoU by divide 19.Actually, you need to re-calculate the value by divide 16. There are only 16 shared classes for SYNTHIA-to-Cityscapes.In this way, the result is same as the value reported in paper.
We also would like to thank great works as follows:
- https://github.com/wasidennis/AdaptSegNet
- https://github.com/RoyalVane/CLAN
- https://github.com/yzou2/CRST
@inproceedings{zheng2020unsupervised,title={Unsupervised Scene Adaptation with Memory Regularization in vivo},author={Zheng, Zhedong and Yang, Yi},booktitle={IJCAI},year={2020}}@article{zheng2021rectifying,title={Rectifying Pseudo Label Learning via Uncertainty Estimation for Domain Adaptive Semantic Segmentation},author={Zheng, Zhedong and Yang, Yi},journal={International Journal of Computer Vision (IJCV)},doi={10.1007/s11263-020-01395-y},note={\mbox{doi}:\url{10.1007/s11263-020-01395-y}},year={2021}}