1

Google colab notebook comes packaged with PyTorch and GPU support -- see this related SO answer, and then you should go change notebook runtime to GPU from CPU (from the top Colab menubar).

Yet trying to train a model from GPyTorch always fails on

/usr/local/lib/python3.6/dist-packages/gpytorch/utils/cholesky.py in psd_safe_cholesky(A, upper, out, jitter)
     19     """
     20     try:
---> 21         L = torch.cholesky(A, upper=upper, out=out)
     22         # TODO: Remove once fixed in pytorch (#16780)
     23         if A.dim() > 2 and A.is_cuda:

RuntimeError: CUDA error: invalid device function

Why?

Relevant info:

import torch
torch.cuda.is_available()

True

print(torch.__version__)
print(gpytorch.__version__)
print(torch.backends.cudnn.version())

1.3.0+cu100
0.3.6
7603
talonmies
  • 70,661
  • 34
  • 192
  • 269
BoltzmannBrain
  • 5,082
  • 11
  • 46
  • 79
  • The version of torch you are trying to use doesn't have support for the GPU you are using – talonmies Oct 20 '19 at 16:26
  • @talonmies how can you tell? And how can I make setup the versions s.t. they work together? – BoltzmannBrain Oct 20 '19 at 17:42
  • Because I know what that CUDA runtime error means. Either the whole of the torch build doesn't any support for the colab GPU or that function doesn't. I can't help you with how to fix it – talonmies Oct 20 '19 at 18:00

0 Answers0