gjd2017

Results 24 issues of gjd2017

How to test?

1 comment

Hi, How to test my own dataset? Do you have the test script? Thank you for your help.

#27 is right?(conv3x3(planes*2, planes*2, groups=num_group))

6 comment

self.conv2 = conv3x3(planes*2, planes*2, groups=num_group), add' groups=num_group'?

代码写的真的是一言难尽啊

6 comment

关于PFLD算法

1 comment

都说PFLD它是一个简单、快速、超高精度人脸特征点检测算法，但是这个算法还是存在问题的：眨眼的情况下，关键点定位不敏感。

Support Multi-GPU?

TNN，GPU部署，加载库和模型耗时太长，大该有3秒耗时，怎么解决？

1 comment

How can I use CRD_loss to face landmark detetct for model compression? There is no "opt.nce_k: number of negatives paired with each positive".

Only support batchsize=1?

Hi, in the roialignlayer, Currently only supports batchsize 1?

为啥总是提示：下载失败？

OST的数据集也是和DF2K一样的处理方式么？