lucidrains pixel-level-contrastive-learning issues

lucidrains / pixel-level-contrastive-learning

Implementation of Pixel-level Contrastive Learning, proposed in the paper "Propagate Yourself", in Pytorch

MIT License

252 stars 29 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

Could the pixel-level contrastive learning be adapted to 3D classification or segmentation task？

#20 Lazzzycoding opened 8 months ago
0
Apply PixPro to resnet50 with FPN

#19 simonebonato opened 9 months ago
0
Loss values are not reducing

#18 NamburiSrinath opened 2 years ago
1
Extending the model for Semantic Segmentation task

#17 NamburiSrinath opened 2 years ago
3
Could you report performance?

#16 jianlong-yuan closed 3 years ago
0
How to train PixelCL on multi-gpu?

#15 ALEX13679173326 opened 3 years ago
0
positive pairs

#14 jancylee opened 3 years ago
5
Same size models -> wildly different RAM usage

#13 makarr closed 3 years ago
2
Added parameter for returning positive pixels pairs

#12 ahmed-bensaad closed 3 years ago
0
Using pytorch-lightning to train PixelCL on multi-gpu

#11 ahmed-bensaad opened 3 years ago
11
Data Augumentation not same with BYOL?

#10 guanfuchen closed 3 years ago
3
use keops in pixel propagation module

#9 lucidrains closed 3 years ago
0
O(n^4) memory usage in PPM

#8 neonbjb closed 3 years ago
13
about the pos-neg pair selection

#7 1144181135 closed 3 years ago
1
about distributed

#6 datuizhuang closed 3 years ago
1
The distance_matrix and similarity_matrix may not match

#5 lygsbw closed 3 years ago
9
Alpha - Combination of PixPro loss and PixCL

#4 WilhelmT closed 3 years ago
6
Add the two loss functions instead of multiplying them

#3 cjrd closed 3 years ago
1
Does using RandomResizedCrop/HorizontalFlip in the augmentaion pipeline break the pixel alignment?

#2 cjrd closed 3 years ago
4
Help needed

#1 lucidrains closed 3 years ago
6