Low Precision Decentralized Distributed Training over IID and non-IID Data

TL;DR

This paper introduces a low precision decentralized training method that reduces computational and communication costs significantly while maintaining high accuracy, even with non-IID data, by proposing novel normalization and combining with compression techniques.

Contribution

It presents a novel low precision training approach for decentralized learning, including a new normalization layer and synergy with compression methods, addressing non-IID data challenges.

Findings

8-bit training maintains accuracy with minimal loss.

Combining low precision with sparsification causes 1-2% accuracy drop.

Training reduces complexity, memory, and energy by 4x and 20x respectively.

Abstract

Decentralized distributed learning is the key to enabling large-scale machine learning (training) on edge devices utilizing private user-generated local data, without relying on the cloud. However, the practical realization of such on-device training is limited by the communication and compute bottleneck. In this paper, we propose and show the convergence of low precision decentralized training that aims to reduce the computational complexity and communication cost of decentralized training. Many feedback-based compression techniques have been proposed in the literature to reduce communication costs. To the best of our knowledge, there is no work that applies and shows compute efficient training techniques such as quantization, pruning, etc., for peer-to-peer decentralized learning setups. Since real-world applications have a significant skew in the data distribution, we design "Range-EvoNorm" as the normalization activation layer which is better suited for low precision training over non-IID data. Moreover, we show that the proposed low precision training can be used in synergy with other communication compression methods decreasing the communication cost further. Our experiments indicate that 8-bit decentralized training has minimal accuracy loss compared to its full precision counterpart even with non-IID data. However, when low precision training is accompanied by communication compression through sparsification we observe a 1-2% drop in accuracy. The proposed low precision decentralized training decreases computational complexity, memory usage, and communication cost by 4x and compute energy by a factor of ~20x, while trading off less than a $1\%$ accuracy for both IID and non-IID data. In particular, with higher skew values, we observe an increase in accuracy (by ~ 0.5%) with low precision training, indicating the regularization effect of the quantization.