vandit15 Class-balanced-loss-pytorch issues

vandit15 / Class-balanced-loss-pytorch

Pytorch implementation of the paper "Class-Balanced Loss Based on Effective Number of Samples"

MIT License

781 stars 120 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

Update class_balanced_loss.py

#24 krittaprot opened 7 months ago
0
Why using F.binary_cross_entropy after softmax?

#23 Starethics opened 11 months ago
0
Make it available on GPU device

#22 JosieHong opened 1 year ago
0
想知道数据集怎么切分的

#21 liyang-good opened 2 years ago
0
focal_loss with softmax

#20 czjghost opened 2 years ago
0
Can I replace (1) to (-1) ?

#19 yeon-kk opened 2 years ago
0
fix typo

#18 bilzard closed 3 years ago
2
typo

#17 e-yi opened 3 years ago
1
why is no_of_classes needed for weights normialisation

#16 m-zheng closed 2 years ago
3
why sum weights?

#15 mmxuan18 opened 3 years ago
2
Sth about samples_per_cls

#14 Liudzz opened 4 years ago
0
fix bug in call to BCEwithlogits

#13 aalok-sathe opened 4 years ago
0
Is this same to apply pos_weight?

#12 jtlee90 opened 4 years ago
0
Object detection and Semantic segmentation

#11 abhigoku10 opened 4 years ago
1
Line 87 typo

#10 BenDrewry opened 4 years ago
1
why modulator?

#9 lai199508 opened 4 years ago
2
could you explain why you are using binary cross-entropy

#8 xysong1201 opened 4 years ago
1
base information about cbloss

#7 Adorablepet closed 4 years ago
2
Typo in #line87

#6 Anurag14 opened 4 years ago
0
What is samples_per_cls?

#5 Bonsen closed 4 years ago
3
Is this implementation correct?

#4 Hhhhhhao closed 4 years ago
4
Multi label application

#3 DecentMakeover closed 4 years ago
2
Sigmoid Loss does not use weights

#2 inspirit closed 4 years ago
0
Reduction for the Cross Entropy

#1 inspirit closed 5 years ago
1