RuntimeError: CUDA error: CUBLAS_STATUS_EXECUTION_FAILED when running the code

[CUHKWilliam]

Hi! Thanks for your interesting and enlightening work on point cloud reconstruction tasks, and we are trying to reproduce your work. However, we encounter an error when running your code: I am wondering how to fix this bug? Thanks for your timely response.

[funny000]

when i use uperhead_transformer net, I have also the problem, and look for the problem in the internet, other people said a lot of method, but one of all can't solve the problem, I guess the problem about the memory of cuda, when the memory is big and sufficient.

[ListIndexOutOfRange]

What's your CUDA version ? If it is < 11, you should consider upgrade it, try and see if it fixes the error.

[vaheta]

I have the same problem with CUDA 11.1 and Pytorch 1.8.1 (these are the versions mentioned in the repo) on a machine with P100. Trying to figure out the solution now.

[cyberkarim]

did you get any clues, I'm still running into the same issues using the same poco environement specified in requirements.txt

[cyberkarim]

I have the same problem with CUDA 11.1 and Pytorch 1.8.1 (these are the versions mentioned in the repo) on a machine with P100. Trying to figure out the solution now.

Did you get any clues to solve the issue

I'm using redhat linux distro. My loaded cuda packages are :

[vaheta]

did you get any clues, I'm still running into the same issues using the same poco environement specified in requirements.txt

To be honest, I don't remember - I ended up reinstalling the CUDA and Pytorch a bunch of times, and at some point, it worked.

[cyberkarim]

Issue still pending, I was able to make it work once by reinstalling all the dependencies. Once I switched to another node server, reinstalled all dependencies following the same guidlines, it got me back to this error. @aboulch there's still no clear way to solve this issue.