Issues
- 2
- 2
多分类时参数α的意义是什么?
#25 opened by ChuanTianML - 0
采样 vs FocalLoss,哪个是处理训练数据不均问题的更好方法?
#26 opened by guotong1988 - 0
- 4
Initialization
#4 opened by GOATmessi8 - 0
- 0
- 0
how to implemented it in faster rcnn?
#19 opened by zqdeepbluesky - 0
doubts about cross entropy with softmax
#21 opened by Tveek - 5
Focal Loss in Keras
#8 opened by mkocabas - 1
Your alpha hyper parameter is wrong.
#9 opened by chinakook - 0
Does it sample fixed number of Rois for RCNN training just like the original one, or use much more Rois?
#20 opened by ChengpengChen - 1
- 0
Can focal loss is applied for RPN loss?
#17 opened by hdjsjyl - 0
train error
#16 opened by xiaoheizi123 - 3
About the focal loss layer
#6 opened by suzhenghang - 1
Is there something wrong when i !=j?
#15 opened by cciao - 5
- 7
Focal loss derivative is wrong when i!=j.
#1 opened by Zehaos - 0
Apply focal loss on sampled training examples?
#13 opened by yaoqi-zd - 7
test on pascal voc dataset?
#5 opened by zimenglan-sysu-512 - 0
focal loss layer not support multi_output.
#12 opened by iFighting - 9
How to use it
#2 opened by tzw-1998 - 1
- 2
How to use focal loss in Caffe Models?
#10 opened by RichardMrLu - 1
About the training result
#7 opened by franciszzj - 4
How does it work?
#3 opened by Johnson-yue