Giters

Confusezius / Revisiting_Deep_Metric_Learning_PyTorch

(ICML 2020) This repo contains code for our paper "Revisiting Training Strategies and Generalization Performance in Deep Metric Learning" (https://arxiv.org/abs/2002.08473) to facilitate consistent research in the field of Deep Metric Learning.

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Multi-GPU support?

deeptek012 opened this issue · comments

commented

Hi @Confusezius ,

Thanks for the great work! It's really helpful. Now I have a large dataset to train, and I wonder if more than 1 GPU is supported? I saw you mentioned use_data_parallel, and I saw some implementation in misc.py. But I didn't see how to turn it on through parameter.py and main.py. Could you help me with that?

Thanks a lot!

commented

HI @deeptek012,

thanks for the kind words! I did use data-parallel during the inception (hence the flag), but removed it for ease of use and prototyping at some point. Technically, it should be quite straightforward to re-incorporate it by wrapping dataloaders & models into the corresponding PyTorch Wrappers, but unfortunately, the current version of the code does not provide it.

commented

Got it. Thanks a lot!