Issues
- 0
Why do you use tgt_soft_label.detach() in Ladv?
#28 opened by Sarmadfismael - 0
GTA5 generate_gta5_label_info.py
#27 opened by zhijiejia - 2
Freeze discriminator parameters before performing adversarial loss backward
#26 opened by daofeng2007 - 2
- 1
- 1
why batch_size // 2 ?
#24 opened by desti-nation - 2
Could not reproduce results for generate_synthia_label.py and the use of its result
#23 opened by KiwiXR - 2
How to implement the CCD in pytorch?
#22 opened by tinazoe1997 - 2
Why Test Input Size of self-distillation phase is 1024*2048? instead of 512*1024 as before?
#11 opened by nbilx - 5
train custom dataset
#20 opened by zhou-rui1 - 5
VGG pre-trained model
#19 opened by tinazoe1997 - 3
why is the test size different from the train size in the self-distill phase ?
#18 opened by taintpro98 - 3
Performance on Cityscapes to Cross-City
#14 opened by BinhuiXie - 0
- 1
Why is self-distill-training better?
#16 opened by taintpro98 - 2
- 4
why don't you set one more label with name "others" for labels in cityscapes which do not exist in GTA5 or Sythia ?
#13 opened by taintpro98 - 1
About multi-scale testing?
#12 opened by happyzhouch - 1
- 1
- 1
- 1
Performance on Synthia -> Cityscapes
#8 opened by zhijiew - 1
SYNTHIA model evaluation
#7 opened by JuiChang - 1
About Cross-City?
#6 opened by BinhuiXie - 4
- 1
Pre-trained models
#3 opened by adricarda - 2
About LICENSE
#2 opened by yukkyo - 1
comparison with previous art
#1 opened by Turlan