flyuz / voxelmorph-tf2-min Goto Github PK

输入图像的维度和原始数据维度不一致，报错，怎么修改，是哪里出了问题呀

Hello:
Does your code refer to VoxelMorph?or just refer to other?
I have a question:why can you train your network with only two graphs and get good results?
Waiting your answer.
Thanks!

@Jenifer-xm 如果使用的是本仓库自带的数据集，只需要修改路径就可以运行，这已经过多人测试。如果使用的是您本人的数据集，请根据实际情况修改。

Originally posted by @FlyuZ in #2 (comment)

当我输入图片大小为1440*1080的时候就会报错

您好 我跑出来出现这样的错误 tensorflow.python.framework.errors_impl.UnknownError: Failed to get convolution algorithm. This is probably because cuDNN failed to initialize, so try looking to see if a warning log message was printed above. [Op:Conv2D]

我安装的是tensorflow2.0版本 望指点！

@FlyuZ

Hi, I was wondering how you implement back propagation of the transform layer for dense field transform? I tried to reproduce VoxelMorph's work but stuck on the backpropagation part.

I will appreciate if you could share some ideas.
Thanks!