paper-with-me

Papers

HPSGD: Hierarchical Parallel SGD With Stale Gradients Featuring

2020-09-06 · Yuhao Zhou, Qing Ye, Hailun Zhang, Jiancheng Lv

While distributed training significantly speeds up the training process of the deep neural network (DNN), the utilization of the cluster is relatively low due to the time-consuming data synchronizing between workers. To alleviate this problem, a novel Hierarchical Parallel SGD (HPSGD) strategy is proposed based on the observation that the data synchronization phase can be paralleled with the local training phase (i.e., Feed-forward and back-propagation). Furthermore, an improved model updating method is unitized to remedy the introduced stale gradients problem, which commits updates to the replica (i.e., a temporary model that has the same parameters as the global model) and then merges the average changes to the global model. Extensive experiments are conducted to demonstrate that the proposed HPSGD approach substantially boosts the distributed DNN training, reduces the disturbance of the stale gradients and achieves better accuracy in given fixed wall-time.

📄 PDF Abstract BibTeX arXiv:2009.02701

Code (1)

Soptq/Hierarchical_Local_SGD 공식 구현 pytorch

Methods 이 논문이 사용한 방법론

SGD Stochastic Gradient Descent is an iterative optimization technique that uses minibatches of data to form an expectation of the gradient, rather than the full gradient using…

Similar Papers 제목 키워드 기반

Nesterov Method for Asynchronous Pipeline Parallel Optimization

2025-05-02 · Thalaiyasingam Ajanthan, Sameera Ramasinghe, Yan Zuo, Gil Avraham 외

Pipeline Parallelism (PP) enables large neural network training on small, interconnected devices by splitting the model into multiple stages. To maximize pipeline utilization, asynchronous optimization is appealing as it…

DecoderLanguage Modelling

Accumulated Decoupled Learning: Mitigating Gradient Staleness in Inter-Layer Model Parallelization

2020-12-03 · Huiping Zhuang, Zhiping Lin, Kar-Ann Toh

Decoupled learning is a branch of model parallelism which parallelizes the training of a network by splitting it depth-wise into multiple modules. Techniques from decoupled learning usually lead to stale gradient effect …

General Classification

MindTheStep-AsyncPSGD: Adaptive Asynchronous Parallel Stochastic Gradient Descent

2019-11-08 · Karl Bäckström, Marina Papatriantafilou, Philippas Tsigas

Stochastic Gradient Descent (SGD) is very useful in optimization problems with high-dimensional non-convex target functions, and hence constitutes an important component of several Machine Learning and Data Analytics met…

Stochastic Gradient MCMC with Stale Gradients

2016-10-21 · NeurIPS 2016 12 · Changyou Chen, Nan Ding, Chunyuan Li, Yizhe Zhang 외

Stochastic gradient MCMC (SG-MCMC) has played an important role in large-scale Bayesian learning, with well-developed theoretical convergence properties. In such applications of SG-MCMC, it is becoming increasingly popul…

Asynchronous Stochastic Gradient MCMC with Elastic Coupling

2016-12-02 · Jost Tobias Springenberg, Aaron Klein, Stefan Falkner, Frank Hutter

We consider parallel asynchronous Markov Chain Monte Carlo (MCMC) sampling for problems where we can leverage (stochastic) gradients to define continuous dynamics which explore the target distribution. We outline a solut…