Recall and Precision

[lixiaocong-svg]

Hello, I used your model for training and found that the Recall and Precision values are very low. Is this normal? The training set I use is celebaA

[Demo122]

Hello, I used your model for training and found that the Recall and Precision values are very low. Is this normal? The training set I use is celebaA

Have you solved it, I also encountered this problem, but I used my own data set

[Ghost0405]

When I'm training, Recall and Precision is always 1. I haven't modified this part of the code, do you know why?

[Demo122]

我也没有解决，随后便放弃了

---Original--- From: @.> Date: Fri, Mar 31, 2023 09:27 AM To: @.>; Cc: @.**@.>; Subject: Re: [knazeri/edge-connect] Recall and Precision (#171)

When I'm training, Recall and Precision is always 1. I haven't modified this part of the code, do you know why?

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: @.***>

[Ghost0405]

我也没有解决，随后便放弃了 … ---Original--- From: @.> Date: Fri, Mar 31, 2023 09:27 AM To: @.>; Cc: @.**@.>; Subject: Re: [knazeri/edge-connect] Recall and Precision (#171) When I'm training, Recall and Precision is always 1. I haven't modified this part of the code, do you know why? — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you commented.Message ID: @.***>

好的，谢谢你的回复

[kelvinlmc]

是不是因为在训练的时候load_edge函数下mask = None if self.training else (1 - mask / 255).astype(np.bool)只要是在训练那么mask始终等于None，也就是说mask在训练阶段不起作用，输入到generateor当中的masks是完整的边缘，没有明白为什么这么设置，不应该是在train的时候是拿破损图像的边缘进行生成吗

[kelvinlmc]

共同学习可以加个好友不296626103qq