search

UKPLab / gpl

Powerful unsupervised domain adaptation method for dense retrieval. Requires only unlabeled corpus and yields massive improvement: "GPL: Generative Pseudo Labeling for Unsupervised Domain Adaptation of Dense Retrieval" https://arxiv.org/abs/2112.07577
Apache License 2.0
315 stars 39 forks source link

Training on multi gpu #9

Open alsbhn opened 2 years ago

alsbhn commented 2 years ago

Is there a way to train GPL model with multi gpu? If yes can that help for training with larger batches?

kwang2049 commented 2 years ago

The easiest solution would be to go with nn.DataParallel, although it might not be the most efficient way. Definitely, this will help for training with larger batches. I will have a try and integrate this feature if possible in the near future.