A Better Alternative to Error Feedback for Communication-Efficient Distributed Learning

Authors: Samuel Horváth, Peter Richtarik

ICLR 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We perform several numerical experiments which validate our theoretical findings. Finally, we provide an experimental evaluation on an array of classification tasks with CIFAR10 dataset corroborating our theoretical findings.
Researcher Affiliation Academia Samuel Horváth and Peter Richtárik KAUST Thuwal, Saudi Arabia {samuel.horvath, peter.richtarik}@kaust.edu.sa
Pseudocode Yes Algorithm 1 DCSGD; Algorithm 2 DCSGD with Error Feedback
Open Source Code No The paper does not provide an explicit statement about releasing the source code for the described methodology or a direct link to a code repository.
Open Datasets Yes We do an evaluation on CIFAR10 dataset.
Dataset Splits Yes A validation accuracy is computed based on 10 % randomly selected training data.
Hardware Specification No The paper does not provide specific hardware details (e.g., GPU/CPU models, memory amounts, or detailed computer specifications) used for running its experiments.
Software Dependencies No Our experimental results are based on a Python implementation of all the methods running in Py Torch. (No version numbers provided for Python or PyTorch).
Experiment Setup Yes We used a local batch size of 32. For every experiment, we randomly distributed the training dataset among 8 workers; each worker computes its local gradient-based on its own dataset. We consider VGG11 (Simonyan & Zisserman, 2015) and Res Net18 (He et al., 2016) models and step-sizes 0.1, 0.05 and 0.01.