CoinCheung pytorch-loss issues

CoinCheung / pytorch-loss

label-smooth, amsoftmax, partial-fc, focal-loss, triplet-loss, lovasz-softmax. Maybe useful

MIT License

2.17k stars 374 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

should use log_probs in affinity loss

#41 eliqbiq opened 1 month ago
1
hard triplet conergence

#40 ffredd opened 2 years ago
0
installation failed

#39 mosheliv opened 2 years ago
3
“THC/THC.h”: No such file or directory

#38 cop123123 opened 2 years ago
5
remove THC.h

#37 zyddnys closed 2 years ago
1
fatal error: THC/THC.h: No such file or directory

#36 zyddnys closed 2 years ago
2
some question in rankloss

#35 nijiazhi closed 2 years ago
5
class balanced loss?

#34 2533245542 opened 2 years ago
1
partial_fc_amsoftmax memory leak

#33 tianxingyzxq closed 2 years ago
5
Regularizing Neural Networks by Penalizing Confident Output Distribution

#32 eakirtas closed 2 years ago
4
Amazing work！

#31 anxiangsir closed 2 years ago
1
LabelSmoothSoftmaxCEV1, LabelSmoothSoftmaxCEV2, LabelSmoothSoftmaxCEV3

#30 quanweiliu closed 2 years ago
3
Question about Multi-class focal loss

#29 DingYuan0118 closed 2 years ago
4
fatal error: math.h: No such file or directory

#28 snakers4 opened 3 years ago
2
your focal loss is wrong? it seems little different with others, can you explain your code?

#27 Jack-zz-ze closed 2 years ago
8
Calculation about focal loss

#26 BinsonW closed 3 years ago
1
哪一个损失函数最好用，损失函数类中里面又不带参数？

#25 jingzhengli closed 3 years ago
1
TPU support

#24 talhaanwarch closed 3 years ago
1
use of TO_REMOVE in GIOU loss

#23 souvik3333 closed 3 years ago
1
there is some problems in the file of focal_loss.py

#22 OliviaDdy closed 3 years ago
4
LovaszSoftmax Loss test case not working

#21 omerferhatt closed 3 years ago
4
如何调用tripletloss？

#20 392398741nm closed 3 years ago
1
The probability caculation of negative class in 'Label smooth.py'

#19 kaijieshi7 opened 3 years ago
0
feature request: dice loss support multi-class classification ?

#18 feng-1985 closed 3 years ago
1
new feature request: dice loss

#17 feng-1985 closed 3 years ago
0
请问am-softmax可以用在多分类任务上么，比如网络结构最后一层输出分类是7，那amsoftmax(）中的参数填啥，第一个in_feats应该填7，第二个num_class也填7,这应该不对，不知道怎么使用，请教了

#16 wbb95 closed 3 years ago
2
Any plan to give a tutorial?

#15 seanzhuh closed 3 years ago
1
need cuda for installing on MacOs

#14 reymondzzzz closed 3 years ago
1
where is the mish_cpp?

#13 hhl0202 closed 3 years ago
3
I want to use FocalLossV1,but has the error that your loss has no attribute to backward().

#12 hanna6666 closed 3 years ago
5
use on multi-label 3D medical image segmentation task?

#11 xwjBupt closed 3 years ago
2
BCEWithLogitsLoss has combined a Sigmoid layer and the BCELoss in one single class, But why to use torch.sigmoid again

#10 sunpeng981712364 closed 3 years ago
1
关于triplet loss

#9 AnnaXiong closed 4 years ago
3
do buffered params in EMA need to be updated?

#8 DietDietDiet closed 3 years ago
1
Regarding Generalized IOU Loss

#7 samyak0210 closed 4 years ago
2
dual focal loss

#6 damhurmuller closed 4 years ago
7
AM-softmax implement details

#5 sweetTT closed 4 years ago
2
about smooth label

#4 mk123qwe closed 4 years ago
1
How do I design my loss fuction like smoothL1？

#3 kelisiya closed 3 years ago
8
What does ignore_lb = 255 do?

#2 wusaifei closed 4 years ago
1
The purpose of lb_ignore in LabelSmoothSoftmaxCE?

#1 LightToYang closed 4 years ago
2