Closed do-oo closed 5 years ago
发个结果过来
顺祝好。
On Aug 21, 2018, at 22:39, Will-Hui notifications@github.com wrote:
这里使用你写的valPCK 验证文件执行的结果,这个结果为什么这么高?是不是那里出现了问题了呢?
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub, or mute the thread.
结果就在截的那张图里面。那个是PCKh0.5
我在截取一部分吧 [] Processing Result in normalized distance of 0.01 [] r_anckle Accuracy : 0.991927346115 [] r_knee Accuracy : 0.991927346115 [] r_hip Accuracy : 0.991927346115 [] l_hip Accuracy : 0.991927346115 [] l_knee Accuracy : 0.991927346115 [] l_anckle Accuracy : 0.991927346115 [] pelvis Accuracy : 0.991927346115 [] thorax Accuracy : 0.991927346115 [] neck Accuracy : 0.991927346115 [] head Accuracy : 0.991927346115 [] r_wrist Accuracy : 0.991927346115 [] r_elbow Accuracy : 0.991927346115 [] r_shoulder Accuracy : 0.991927346115 [] l_shoulder Accuracy : 0.991927346115 [] l_elbow Accuracy : 0.991927346115 [] l_wrist Accuracy : 0.991927346115 [] TOTAL Accuracy : 0.991927346115 [] Processing Result in normalized distance of 0.0149494949495 [] r_anckle Accuracy : 0.991927346115 [] r_knee Accuracy : 0.991927346115 [] r_hip Accuracy : 0.991927346115 [] l_hip Accuracy : 0.991927346115 [] l_knee Accuracy : 0.991927346115 [] l_anckle Accuracy : 0.991927346115 [] pelvis Accuracy : 0.991927346115 [] thorax Accuracy : 0.991927346115 [] neck Accuracy : 0.991927346115 [] head Accuracy : 0.991927346115 [] r_wrist Accuracy : 0.991927346115 [] r_elbow Accuracy : 0.991927346115 [] r_shoulder Accuracy : 0.991927346115 [] l_shoulder Accuracy : 0.991927346115 [] l_elbow Accuracy : 0.991927346115 [] l_wrist Accuracy : 0.991927346115 [] TOTAL Accuracy : 0.991927346115
这里代码的确有问题,但是最近事情比较多,稍等我把我自己本地的代码整理下再commit。 你也可以自己debug试试。
在 2018年8月22日,上午11:19,Will-Hui <notifications@github.com mailto:notifications@github.com> 写道:
esult in normalized distance of 0.0149494949495
可以留个联系方式吗,我也是工大的哈哈
是嘛hh,好可以。
在 2018年8月22日,下午12:20,Will-Hui notifications@github.com 写道:
可以留个联系方式吗,我也是工大的哈哈
— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/mpskex/Convolutional-Pose-Machine-tf/issues/6#issuecomment-414905928, or mute the thread https://github.com/notifications/unsubscribe-auth/AIEKAli4iTjUeEyK6xM5yoOHNTRaWn6-ks5uTNwpgaJpZM4WF7J8.
这里使用你写的valPCK 验证文件执行的结果,这个结果为什么这么高?是不是那里出现了问题了呢?