yatengLG/Focal-Loss-Pytorch
全中文注释.(The loss function of retinanet based on pytorch).(You can use it on one-stage detection task or classifical task, to solve data imbalance influence).用于one-stage目标检测算法,提升检测效果.你也可以在分类任务中使用该损失函数,解决数据不平衡问题.
Jupyter Notebook
Issues
- 0
- 1
typos in your readme
#19 opened by jizhang02 - 0
当num_classes=2时(做二分类问题时),出现错误
#18 opened by zxcvbnmkj - 0
- 0
- 5
- 6
- 1
请问类别不平衡的解决思路是什么
#3 opened by qwe3208620 - 2
- 1
数值不稳定问题
#7 opened by yangwei-nlp - 1
当batch size是1的时候报错
#15 opened by lx9cat - 1
这个报错该怎么解决?
#10 opened by jiaoxiaosong - 1
- 3
- 12
原始论文中提到,应该给样本数量较少的类别较小的权重
#12 opened by wmathor - 0
NER中num_class怎么设置?
#13 opened by Smile-L-up - 19
alpha如何设置?
#2 opened by ChenXiao61 - 1
这个第41行的self.alpha,应该不用更新吧?
#14 opened by jjxyai - 1
老哥,第40行是不是应该用reshape啊?
#1 opened by RichardHWD