Running on multiple GPUs

Part 1 (2019) Advanced (Part 1 v3)
1

Due to the big size of my dataset I face the pressing need of running my model on multiple GPUs (2-4).

I have tried:

model=torch.nn.DataParallel(model, device_ids=[0, 3])

without no success.

What should I do?

2

You can follow this tutorial to run your model on multiple GPUs https://docs.fast.ai/distributed.html

1 Like
3

Thank you!!