issues
search
lhoyer
/
DAFormer
[CVPR22] Official Implementation of DAFormer: Improving Network Architectures and Training Strategies for Domain-Adaptive Semantic Segmentation
Other
466
stars
91
forks
source link
issues
Newest
Newest
Most commented
Recently updated
Oldest
Least commented
Least recently updated
How to tune hyper-parameters?
#41
super233
opened
2 years ago
0
The unreproducible results
#40
JunXieFront
closed
2 years ago
4
imnet_feature_dist_classes=[6, 7, 11, 12, 13, 14, 15, 16, 17, 18]
#39
creater-zq
closed
2 years ago
1
optimizer and backward
#38
creater-zq
closed
2 years ago
0
feat_loss.backward() and Thing-Class ImageNet Feature Distance (FD)
#37
creater-zq
closed
2 years ago
2
problem of ‘def get_rare_class_sample’
#36
Xz-Alan
closed
2 years ago
3
Rare Class Sampling (RCS)
#35
creater-zq
closed
2 years ago
2
target ann_dir
#34
Changryeol-Jeon
closed
2 years ago
2
The codes in the channel alignment section seem not consistent with the paper.
#33
Levantespot
closed
2 years ago
2
How to count the class statistics in Figure S1 of the supplementary material?
#32
qianyuzqy
closed
2 years ago
3
Figure 1 of "the Progress of UDA over time on GTAV→Cityscapes"
#31
qianyuzqy
closed
2 years ago
2
KeyError: Caught KeyError in DataLoader worker process 3.
#30
hpuxing
closed
2 years ago
3
where's the process of ema updating?
#29
sunrainyg
closed
2 years ago
2
What is the config to run the baseline without any extra method?
#28
xingchenzhao
closed
2 years ago
2
How to train on multi-target dataset
#27
sunrainyg
closed
2 years ago
0
Question about loss
#26
Yulu-gan
closed
2 years ago
1
Questions about the data stream
#25
Yulu-gan
closed
2 years ago
3
KeyError: 'gta\\labels\\13432_labelTrainIds.png'
#24
mkbkxdd
closed
2 years ago
6
About accuracy
#23
BUAA-LKG
closed
2 years ago
9
loss = Nan
#22
kaigelee
closed
2 years ago
2
beautiful work！
#21
kaigelee
closed
2 years ago
2
Difference to DACS
#20
wangkaihong
closed
2 years ago
2
The get_class_mask function
#19
JunXieFront
closed
2 years ago
1
Performance gap on Synthia to Cityscapes
#18
Pranav-Goyal
closed
2 years ago
2
src.loss_imnet_feat_dist is nan
#17
Haochen-Wang409
closed
2 years ago
4
pretrained on source-only
#16
wzlyyb
closed
2 years ago
3
python run_experiments.py --exp 101 does not launch the experiments
#15
dbash
closed
2 years ago
2
How many epochs have you trained for Oracle performance?
#14
usr922
closed
2 years ago
2
Performance gap between training DACS with SwinTransformer and SegFormer
#13
super233
closed
2 years ago
2
Request for pretrained model of Table 1
#12
super233
closed
2 years ago
4
'src.loss_imnet_feat_dist' is nan
#11
yuki-no-hana
closed
2 years ago
4
Checkpoint on SYNTHIA>Cityscapes
#10
kw01sg
closed
2 years ago
6
Multi GPU training
#9
phmalek
closed
2 years ago
3
Potential Bug in Crop Resampling for RCS
#8
phmalek
closed
2 years ago
3
KeyError: 'data_time'
#7
leo-hao
closed
2 years ago
7
About RandomCrop on target image
#6
BinhuiXie
closed
2 years ago
6
performance without mix augmentation
#5
valencebond
closed
2 years ago
1
GPU Out of memory errors with RTX 3080
#4
DJayalath
closed
2 years ago
4
How to use own dataset?
#3
DJayalath
closed
2 years ago
3
Invitation of contributing to MMSegmentation.
#2
MengzhangLI
closed
2 years ago
4
Crop PL
#1
valencebond
closed
2 years ago
1
Previous