Issues
- 4
- 0
Where does the quality_focal_loss function in gfocal_loss.py call the weight_reduce_loss function ?
#42 opened by PanffeeReal - 0
如何才能从mmdet里拿出来啊,看源码嵌套的太深了 &_&
#41 opened by Dingerscat - 7
关于训练
#5 opened by liushuai839 - 2
用于跟踪上精度很低
#28 opened by Giveupfree - 3
关于anchor free+GFocalV2的疑问?
#9 opened by sunzhe09 - 0
questions about quality_focal_loss
#40 opened by lalalafloat - 0
Welcome update to OpenMMLab 2.0
#39 opened by vansin - 0
- 0
The software version used when speeding test
#37 opened by slacklife - 7
- 0
- 0
GIoULoss in GFL?
#35 opened by iumyx2612 - 5
TTFNet上部署GFocalV2的一些困惑
#31 opened - 0
- 4
Question about correlation between real localization quality and top-1 generated distribution at training time
#17 opened by pjh4993 - 2
获得bbox分布的可视化图的方法
#33 opened by jackhu-bme - 2
training custom dataset
#30 opened by jackhu-bme - 0
regression_target分布统计问题?
#32 opened by johnson-magic - 0
关于pytorch版本
#21 opened by XavierYoungY - 1
multi-scale testing
#24 opened by jianpursuit - 0
为什么不整合到mmdetection中?
#29 opened by long2double - 0
- 2
Installation
#23 opened by sh7jacobi - 2
mmcv version
#26 opened by jackhu-bme - 1
- 0
Can't run this code
#22 opened by yarkable - 2
RuntimeError: The size of tensor a (4) must match the size of tensor b (18) at non-singleton dimension 1
#14 opened by ma3252788 - 0
关于DCN版本和施加的层的疑问
#20 opened by lywhapp - 0
Question about Quality Focal Loss
#18 opened by harrylin-hyl - 0
- 0
Multiscale testing
#13 opened by kemaloksuz - 9
关于GFocal head中的GN
#10 opened by HAOCHENYE - 2
question about 'use_sigmoid' parameter
#12 opened by menggui1993 - 0
关于训练Negative Dataset
#11 opened by yanghgai - 0
why dive stride here?
#7 opened by qianwangn - 0
Any ideas about Joint Distribution?
#8 opened by coder-james - 1
questions about the paper
#6 opened by shuuchen - 5
Quality Focal Loss become infinitive
#4 opened by black12357 - 13
How to train GFLv2 on RetinaNet?
#3 opened by Zzh-tju - 4
Can I substitute the bridge(topk,fc or 1x1conv) betwenn LQE and Regbranch with some simple operators?such as variance or some parameters which can describe DFL's shape
#2 opened by HAOCHENYE - 0
Can I substitute the bridge(topk,fc or 1下) betwenn LQE and Regbranch with some simple
#1 opened by HAOCHENYE