manmanCover

same problem...

> https://blog.csdn.net/hjimce/article/details/50187881 > https://www.zhihu.com/question/270988169/answer/407731947 > > @manmanCover Sorry, 这两个链接没看懂，能帮忙解释一下吗？

> 训练和测试的输入尺寸确实是确定的。 可以通过vgg的dense的方法，来评估分类器。这样就能够用到全图的像素，而又固定输入尺寸。 > […](#) > ---Original--- From: "manmanCover"

@cfzd hi, I think I found the problem. In file **PSANetFunc.py**, the **backward** methods for both **PSANetCollectFunction** and **PSANetDistributeFunction** ``` b1_grad_n = mask_grad.shape[0] b1_grad_c = (2 * mask_grad.shape[2] - 1)*(2...

@cfzd Maybe your samples are all squares. By the way, do you notice that the memory consumption of your implementation is unbalanced? When I ran the project on 2 GPUs,...

@cfzd Thank you for your test. I have checked that my input feature size is [32, 64, 128], how about yours? By the way, are the input images from training...

@cfzd By the way, is your implementation also use the sliding windows? It seems like not...

@cfzd yeah, adaptive pooling can also be a choice. The author of psanet said they use sliding windows with different input size (https://github.com/hszhao/PSANet/issues/11#issuecomment-457984270). Here's how they use slide windows: https://github.com/hszhao/PSANet/blob/master/evaluation/scale_process.m