Can I use Deepspeed to launch/train a hybrid QML model using multi-GPUs

PennyLane Help
1

Hello!
I am exploring to use Deepspeed software package to integrate with a hybrid QML model. the purpose is to use Deepspeed to launch training runs for a hybrid QML model across multiple GPUs. (GitHub - microsoft/DeepSpeed: DeepSpeed is a deep learning optimization library that makes distributed training and inference easy, efficient, and effective.)

My question is: is it possible to do so, or has anyone successfully integrated QML model with Deepspeed, and run/train a QML model on multiple GPUs via Deepspeed?

I would appreciate any help and information.

2

Hi @JohnQu ,

Welcome to the Forum!
We have indeed used multiple GPUs to run circuits but I haven’t heard of anyone using Deepspeed.

Maybe you want to use Covalent instead. You can learn how to use it in these demos.

I hope this helps.

3

Hi @CatalinaAlbornoz ,
Good to hear from you. Thank you for your suggestion. I will look into it.

1 Like