Nesterov Method for Asynchronous Pipeline Parallel Optimization

• URL: http://arxiv.org/abs/2505.01099v1
• Date: Fri, 02 May 2025 08:23:29 GMT
• Title: Nesterov Method for Asynchronous Pipeline Parallel Optimization
• Authors: Thalaiyasingam Ajanthan, Sameera Ramasinghe, Yan Zuo, Gil Avraham, Alexander Long,
• Abstract summary: We introduce a variant of Nesterov Accelerated Gradient (NAG) for asynchronous optimization in Pipeline Parallelism.<n>Specifically, we modify the look-ahead step in NAG to effectively address the staleness in gradients.<n>We theoretically prove that our approach converges at a sublinear rate in the presence of fixed delay in gradients.
• Score: 59.79227116582264
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: 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 offers 100% pipeline utilization by construction. However, it is inherently challenging as the weights and gradients are no longer synchronized, leading to stale (or delayed) gradients. To alleviate this, we introduce a variant of Nesterov Accelerated Gradient (NAG) for asynchronous optimization in PP. Specifically, we modify the look-ahead step in NAG to effectively address the staleness in gradients. We theoretically prove that our approach converges at a sublinear rate in the presence of fixed delay in gradients. Our experiments on large-scale language modelling tasks using decoder-only architectures with up to 1B parameters, demonstrate that our approach significantly outperforms existing asynchronous methods, even surpassing the synchronous baseline.

Related papers

• Adaptive Deadline and Batch Layered Synchronized Federated Learning [66.93447103966439]
  Federated learning (FL) enables collaborative model training across distributed edge devices while preserving data privacy, and typically operates in a round-based synchronous manner.<n>We propose ADEL-FL, a novel framework that jointly optimize per-round deadlines and user-specific batch sizes for layer-wise aggregation.
  arXiv  Detail & Related papers  (2025-05-29T19:59:18Z)
• Asynchronous Stochastic Gradient Descent with Decoupled Backpropagation and Layer-Wise Updates [1.9241821314180372]
  Asynchronous gradient descent (ASGD) methods can improve training speed, but are sensitive to delays due to both communication and throughput differences.<n>PD-ASGD uses separate threads for the forward and backward passes, decoupling the updates and allowing for a higher ratio of forward to backward threads.<n>Our approach yields close to state-of-the-art results while running up to $5.95times$ faster than synchronous data parallelism in the presence of delays.
  arXiv  Detail & Related papers  (2024-10-08T12:32:36Z)
• OptEx: Expediting First-Order Optimization with Approximately Parallelized Iterations [12.696136981847438]
  We introduce first-order optimization expedited with approximately parallelized iterations (OptEx) OptEx is the first framework that enhances the efficiency of FOO by leveraging parallel computing to mitigate its iterative bottleneck. We provide theoretical guarantees for the reliability of our kernelized gradient estimation and the complexity of SGD-based OptEx.
  arXiv  Detail & Related papers  (2024-02-18T02:19:02Z)
• Stable Nonconvex-Nonconcave Training via Linear Interpolation [51.668052890249726]
  This paper presents a theoretical analysis of linearahead as a principled method for stabilizing (large-scale) neural network training. We argue that instabilities in the optimization process are often caused by the nonmonotonicity of the loss landscape and show how linear can help by leveraging the theory of nonexpansive operators.
  arXiv  Detail & Related papers  (2023-10-20T12:45:12Z)
• Asynchronous SGD Beats Minibatch SGD Under Arbitrary Delays [8.46491234455848]
  We prove much better guarantees for the same asynchronous gradient regardless of the delays in steps, depending instead just on the number of steps. For our analysis, we introduce a novel based on "virtual steps" and delay iterations, which allow us to derive state-of-the-art guarantees for both convex non-adaptive gradients.
  arXiv  Detail & Related papers  (2022-06-15T16:28:37Z)
• Joint inference and input optimization in equilibrium networks [68.63726855991052]
  deep equilibrium model is a class of models that foregoes traditional network depth and instead computes the output of a network by finding the fixed point of a single nonlinear layer. We show that there is a natural synergy between these two settings. We demonstrate this strategy on various tasks such as training generative models while optimizing over latent codes, training models for inverse problems like denoising and inpainting, adversarial training and gradient based meta-learning.
  arXiv  Detail & Related papers  (2021-11-25T19:59:33Z)
• Distributed stochastic optimization with large delays [59.95552973784946]
  One of the most widely used methods for solving large-scale optimization problems is distributed asynchronous gradient descent (DASGD) We show that DASGD converges to a global optimal implementation model under same delay assumptions.
  arXiv  Detail & Related papers  (2021-07-06T21:59:49Z)
• Stochastic Optimization with Laggard Data Pipelines [65.20044914532221]
  We show that "dataechoed" extensions of common optimization methods exhibit provable improvements over their synchronous counterparts. Specifically, we show that in convex optimization with minibatches, data echoing affords speedups on the curvature-dominated part of the convergence rate, while maintaining the optimal statistical rate.
  arXiv  Detail & Related papers  (2020-10-26T14:55:31Z)
• Channel-Directed Gradients for Optimization of Convolutional Neural Networks [50.34913837546743]
  We introduce optimization methods for convolutional neural networks that can be used to improve existing gradient-based optimization in terms of generalization error. We show that defining the gradients along the output channel direction leads to a performance boost, while other directions can be detrimental.
  arXiv  Detail & Related papers  (2020-08-25T00:44:09Z)
• Adaptive Braking for Mitigating Gradient Delay [0.8602553195689513]
  We introduce Adaptive Braking, a modification for momentum-based gradients that mitigates the effects of gradient delay. We show that applying AB on top of SGD with momentum enables training ResNets on CIFAR-10 and ImageNet-1k with delays with minimal drop in final test accuracy.
  arXiv  Detail & Related papers  (2020-07-02T21:26:27Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.