xamyzhao / brainstorm

Implementation of "Data augmentation using learned transforms for one-shot medical image segmentation"
MIT License
392 stars 91 forks source link

Check failed: dims == sizes.size() (5 vs. 4) #23

Closed zhao706 closed 4 years ago

zhao706 commented 4 years ago

Hello, I encountered the following problems while running the main.py. Do you know why? 2020-06-26 11:00:27.306985: F .\tensorflow/core/util/mkl_util.h:636] Check failed: dims == sizes.size() (5 vs. 4)

xamyzhao commented 4 years ago

It seems like the network is getting the wrong size as input. Can you check that the input is batch_size x h x w x d x 1?

On Thu, Jun 25, 2020, 8:26 PM zhao706 notifications@github.com wrote:

Hello, I encountered the following problems while running the main.py. Do you know why? 2020-06-26 11:00:27.306985: F .\tensorflow/core/util/mkl_util.h:636] Check failed: dims == sizes.size() (5 vs. 4)

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/xamyzhao/brainstorm/issues/23, or unsubscribe https://github.com/notifications/unsubscribe-auth/AA75VDIOTJFECTDXDPB6V73RYQIPBANCNFSM4OI6IWEQ .

zhao706 commented 4 years ago

Yes,I checked the input is (batch_size,h,w,d,1).I guess it's about the version, so what's your version of tensorflow?

xamyzhao commented 4 years ago

Can you include your input arguments here? I can try to repro the error.

On Fri, Jun 26, 2020, 1:28 AM zhao706 notifications@github.com wrote:

Yes,I checked the input is (batch_size,h,w,d,1).

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/xamyzhao/brainstorm/issues/23#issuecomment-650055437, or unsubscribe https://github.com/notifications/unsubscribe-auth/AA75VDMZD4XW7UZ2AHU2AMTRYRL3HANCNFSM4OI6IWEQ .

zhao706 commented 4 years ago

Thank you very much for your help. I solved this problem by reducing the version of tensorflow from 1.14.1 to 1.13.0.And I found that the code has a high memory requirement.

zhao706 commented 4 years ago

Hello, I have a doubt is what criteria you choose with label data?

------------------ Original ------------------ From: xamyzhao <notifications@github.com> Date: Sat,Jun 27,2020 0:15 AM To: xamyzhao/brainstorm <brainstorm@noreply.github.com> Cc: zhao706 <1043314579@qq.com>, Author <author@noreply.github.com> Subject: Re: [xamyzhao/brainstorm] Check failed: dims == sizes.size() (5 vs. 4) (#23)

xamyzhao commented 4 years ago

Are you asking how I chose the labeled example to use as our atlas? We picked an example that was close to the mean subject in the dataset (which was computed using standard atlas-building techniques).