When fine-tuning large neural networks, it is common to use multiple nodes and to communicate gradients at each optimization step. By contrast, we investigate completely local fine-tuning, which we refer to as lo-fi. During lo-fi, each node is fine-tuned independently without any communication. Then, the weights are averaged across nodes at the conclusion of fine-tuning. When fine-tuning DeiT-base and DeiT-large on ImageNet, this procedure matches accuracy in-distribution and improves accuracy under distribution shift compared to the baseline, which observes the same amount of data but communicates gradients at each step. We also observe that lo-fi matches the baseline's performance when fine-tuning OPT language models (up to 1.3B parameters) on Common Crawl. By removing the communication requirement, lo-fi reduces resource barriers for fine-tuning large models and enables fine-tuning in settings with prohibitive communication cost.
Cite
@article{arxiv.2210.11948,
title = {lo-fi: distributed fine-tuning without communication},
author = {Mitchell Wortsman and Suchin Gururangan and Shen Li and Ali Farhadi and Ludwig Schmidt and Michael Rabbat and Ari S. Morcos},
journal= {arXiv preprint arXiv:2210.11948},
year = {2022}
}