torch.cuda.is_available() returns false in colab

--------------------------------------------------
Hire the world's top talent on demand or became one of them at Toptal: https://topt.al/25cXVn
and get $2,000 discount on your first invoice
--------------------------------------------------

Take control of your privacy with Proton's trusted, Swiss-based, secure services.
Choose what you need and safeguard your digital life:
Mail: https://go.getproton.me/SH1CU
VPN: https://go.getproton.me/SH1DI
Password Manager: https://go.getproton.me/SH1DJ
Drive: https://go.getproton.me/SH1CT

Music by Eric Matyas
https://www.soundimage.org
Track title: Popsicle Puzzles

--

Chapters
00:00 Torch.Cuda.Is_available() Returns False In Colab
00:34 Answer 1 Score 0
00:50 Accepted Answer Score 1
01:08 Answer 3 Score 21
01:23 Answer 4 Score 5
01:45 Thank you

--

Full question
https://stackoverflow.com/questions/5915...

--

Content licensed under CC BY-SA
https://meta.stackexchange.com/help/lice...

--

Tags
#python #pytorch #googlecolaboratory

#avk47

ANSWER 1

Score 21

Make sure your Hardware accelerator is set to GPU.

Runtime > Change runtime type > Hardware Accelerator

ANSWER 2

Score 5

In case anyone else comes here and makes the same mistake I was making:

If you are trying to check if GPU is available and you do:

if torch.cuda.is_available:
  print('GPU available')
else:
  print('Please set GPU via Edit -> Notebook Settings.')

It will always seem that GPU is available. Note you need to use torch.cuda.is_available() not torch.cuda.is_available.

ACCEPTED ANSWER

Score 1

Worked with all the versions mentioned above and I did not have to downgrade my CUDA to 10.0. I had restarted my colab after the updates which set my running machine back to CPU and I just had to change it back to GPU.

ANSWER 4

Score 0

Temporal fix may be to try Cuda 10.0 as explained in here.

Something like this:

conda install pytorch torchvision cudatoolkit=10.0 -c pytorch

In future versions this may work.