search

guoqiangqi / PFLD

Implementation of PFLD A Practical Facial Landmark Detector , reference to https://arxiv.org/pdf/1902.10859.pdf
626 stars 166 forks source link

loss中的attribute_weight为什么只用五个 #90

Closed lwihtittlee closed 3 years ago

lwihtittlee commented 3 years ago

attributes_w_n = tf.to_float(attribute_batch[:, 1:6]) 为什么不是[0：6]

imistyrain commented 3 years ago

第一个是姿态，也就是角度，单独用转的欧拉角给了一个权重

lwihtittlee commented 3 years ago

好的，谢谢!

------------------ Original ------------------ From:  "MistyRain";<notifications@github.com>; Send time: Wednesday, Dec 30, 2020 5:19 PM To: "guoqiangqi/PFLD"<PFLD@noreply.github.com>; Cc: "down"<3287963521@qq.com>; "Author"<author@noreply.github.com>; Subject:  Re: [guoqiangqi/PFLD] loss中的attribute_weight为什么只用五个 (#90)

第一个是姿态，也就是角度，单独用转的欧拉角给了一个权重

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe.