How to use my GPU

[paulkp (Paul Pallaghy)]

I’ve got a

   product: GM204GL [Tesla M60]
   vendor: NVIDIA Corporation

and I want to use it in a PyTorch OpenNMT installation.

I’m running training and getting a ‘Only using CPU message’.

I’ve tried running:

CUDA_VISIBLE_DEVICES=1,3 -world_size 2 -gpu_ranks 0 1

but of course I’ve got no idea what to set for these parameters.

Help!

[francoishernandez (François Hernandez)]

Specify GPU id in OpenNMT-py

Hello I’d like to know how to specify in OpenNMT-py the use of an specific GPU. I’ve done this: set CUDA_VISIBLE_DEVICES=1 because I want to use GPU with id 1. then I ran this command: python train.py -data sonar_data/sonar-prepro --train_from sonar-model_step_30000.pt -save_model sonar-model1 -save_checkpoint_steps 3000 -gpu_ranks 1 however it always goes to the GPU 0 which is already busy with other task and leave GPU 1 free. Can you please help me with this? Thanks Claudia

Also, you can check if CUDA is available in your setup with

import torch
print(torch.cuda.is_available())

or something like that.