Use GPU on pc to load the models

Beginners
1

Good evening,
I’m trying to load the distillbart-cnn-12-6 on my local machine, my GPU is NVIDIA GeForce GT 740M, and is located on “GPU 1”, when I try to load the model it’s not detected.
Any idea how to solve that?

2

What do you mean it is not detected. What is not detected, the model or your GPU? Can you post the code that you used to load the model?

3

I get this error:
File “c:\users\acer\appdata\local\programs\python\python38\lib\site-packages\torch\cuda_init_.py”, line 47, in _check_driver
raise AssertionError(“Torch not compiled with CUDA enabled”)
AssertionError: Torch not compiled with CUDA enabled

For running, I’m using https://github.com/huggingface/transformers/blob/master/examples/seq2seq/run_eval.py

4

You did not install the correct torch version. You have to install torch with CUDA support if you want to use the GPU. U install torch and then install the right version from https://pytorch.org/get-started/locally/

1 Like