microsoft ProDA issues - Githubissues

microsoft / ProDA

Prototypical Pseudo Label Denoising and Target Structure Learning for Domain Adaptive Semantic Segmentation (CVPR 2021)

https://arxiv.org/abs/2101.10979

MIT License

286 stars 44 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

How can I experiment with ProDA for 13classes?

#56 juajua56 opened 1 year ago
0
Bump scipy from 1.1.0 to 1.10.0

#55 dependabot[bot] opened 1 year ago
0
about settings

#54 YYByyy opened 1 year ago
0
calc_prototype.py

#53 xylzjm opened 1 year ago
1
UMAP visualization

#52 yuheyuan opened 1 year ago
0
Which pictures of Cityscapes dataset are use to make _qualitative results of semantic segmentation_?

#51 ZHE-SAPI opened 2 years ago
1
Potential Bug issue

#50 Kaiseem opened 2 years ago
0
Expected all tensors to be on the same device, but found at least two devices, cuda:0 and cpu!

#49 yuheyuan opened 2 years ago
1
the training in Synthia - > cityscapes warmup

#48 TianX1nYu opened 2 years ago
0
Bump numpy from 1.18 to 1.22.0

#47 dependabot[bot] opened 2 years ago
0
Good job! Some questions about the file calc_prototype.py, thanks！

#46 ZHE-SAPI closed 2 years ago
0
About the mIoU metric

#45 JunXieFront opened 2 years ago
1
About the flip Image

#44 JunXieFront closed 2 years ago
1
The kl_div loss of self distillation

#43 luyvlei opened 2 years ago
1
Exact command line of warmup stage

#42 shahaf1313 opened 2 years ago
3
source-only pretrained model

#41 yuki-no-hana opened 3 years ago
0
About the function full2weak() and label_strong_T()

#40 DeveloperZWK closed 1 year ago
5
two GPUs

#39 RanXu2000 closed 3 years ago
1
pretrained/simclr/r101_1x_sk0.pth

#38 leo-hao closed 1 year ago
0
Training Stage Loss

#37 JunkunPeng17 closed 3 years ago
2
Missing key(s) when loading the pre-trained simclr model parameters in stage 3

#36 JiangXiaobai00 opened 3 years ago
7
Discriminator model

#35 saolaisa02 opened 3 years ago
0
Why the student is initialized with self-supervised weights rather than supervised weights

#34 kailigo closed 3 years ago
1
deeplabv2.py code problem

#33 engine210 closed 3 years ago
1
Have you been trained with source only?

#32 junha1125 closed 3 years ago
2
question about the argument in if statement

#31 mimiliaogo closed 3 years ago
0
Where do I get the "split.mat" file?

#30 jis3613 closed 3 years ago
1
Is warming-up a critical part in the full model performance

#29 lyxok1 closed 3 years ago
1
Question for self distillation in stage2 and stage3

#28 jiangzhengkai closed 3 years ago
1
Warm-up stage training

#27 saolaisa02 closed 3 years ago
10
Missing key(s) in state_dict

#26 y72428026 closed 3 years ago
3
the classifier is replaced with an enhanced version with SE module ?

#25 DZhaoXd closed 3 years ago
1
Model selection

#24 manhminno closed 3 years ago
17
Problem about the pretrained warmup model in "SYNTHIA -> Cityscapes"

#23 super233 closed 3 years ago
1
About the usage of your code with PyTorch>=1.6.0

#22 LinfengYuan1997 opened 3 years ago
2
Amount of epochs for both source only and warm-up stage

#21 fabriziojpiva closed 3 years ago
2
question about the loss formulation of kl_div

#20 ferric123 closed 3 years ago
4
SynchronizedBatchNorm2d or nn.BatchNorm2d?

#19 BinhuiXie closed 3 years ago
1
Results of SYNTHIA2Cityscapes in each stage

#18 super233 closed 3 years ago
2
Train with custom dataset

#17 zhou-rui1 opened 3 years ago
2
UMAP visualization

#16 DaehanKim-Korea closed 3 years ago
1
The method does not have domain alignment that works so well.

#15 jingzhengli closed 3 years ago
1
where to load the generated pseudo labels of target data?

#14 jingzhengli closed 3 years ago
2
A Problem of Using Pretrained Model

#13 flyingintoskyq closed 3 years ago
1
what is the warm-up model?

#12 liuyk97 closed 3 years ago
2
Can not reproduce in Gta5 -cityscape task.

#11 ygjwd12345 closed 3 years ago
7
Out of memory with two GPU training

#10 fabriziojpiva closed 3 years ago
4
Plotting features using UMAP

#9 fabriziojpiva opened 3 years ago
4
specific methods about the stage two.

#8 uzabon closed 3 years ago
2
this amazing performance improvments kill this area!!!!!!!!!!

#7 jingzhengli opened 3 years ago
1