How to use multi-gpus?

[maxmarketit]

Thank you for the promising package.

I wonder how I can use multi-gpus.

I tinkered with num_workers and torch_num_threads but nothing seemed to be changed.

( I wish I could have a document about configuration ).

So is there any way to use multiple gpus? or set the default gpu to use other than 0?

[LMZimmer]

There are a few ways of using parallelism:

i) Multiple workers: Start Auto-PyTorch in multiple threads, where all have the same run_id and different task_id. The worker with task_id=1 will be the master. This is done in the ensemble example.

ii) For image data when a thread has more than one GPU it will automatically apply PyTorchs nn.DataParallel.

iii) Setting the number of threads for data loading in the autonet config will set the corresponding value in nn.DataLoader.

[astro-kevin]

The Ensemble example link is broken. Is there an equivalent in the current examples?