Steven

需要做一次，精度會提高一點，但是不是很大，

Hi @3a532028 did you rentinaface with multi-lcasses work? I changed everything exactly followed what you mentioned,but it didn't work at all

> 采用hopenet模型对widerface数据集进行姿态预标注 你的口罩的分支怎么训练的，可以分享一下idea么？是增加一个branch专门做口罩的property输出，还是num_class改成人脸，口罩两个类型？

恩，有道理，你是吧MAFA的数据做了mosaic吧

moscaic的操作会造成人脸的distortion，这个会对性能有影响么？ On Tue, Feb 2, 2021 at 2:11 AM gengyanlei wrote: > 加1个类似口罩的分支，用于姿态估计，并且加1个mosaic操作！ > > — > You are receiving this because you commented. > Reply to this email directly,...

Yes, mosaic was performing well , especially for small face detection, but my mosaic method is slightly different, I didn't jeopardize the ratio of image, I concatenate 20-100 small single...