paper-with-me

Papers

Galvatron: Efficient Transformer Training over Multiple GPUs Using Automatic Parallelism

2022-11-25 · Xupeng Miao, Yujie Wang, Youhe Jiang, Chunan Shi, Xiaonan Nie, Hailin Zhang, Bin Cui

Transformer models have achieved state-of-the-art performance on various domains of applications and gradually becomes the foundations of the advanced large deep learning (DL) models. However, how to train these models over multiple GPUs efficiently is still challenging due to a large number of parallelism choices. Existing DL systems either rely on manual efforts to make distributed training plans or apply parallelism combinations within a very limited search space. In this approach, we propose Galvatron, a new system framework that incorporates multiple popular parallelism dimensions and automatically finds the most efficient hybrid parallelism strategy. To better explore such a rarely huge search space, we 1) involve a decision tree to make decomposition and pruning based on some reasonable intuitions, and then 2) design a dynamic programming search algorithm to generate the optimal plan. Evaluations on four representative Transformer workloads show that Galvatron could perform automatically distributed training with different GPU memory budgets. Among all evluated scenarios, Galvatron always achieves superior system throughput compared to previous work with limited parallelism.

📄 PDF Abstract BibTeX arXiv:2211.13878

Code (3)

pku-dair/hetu 공식 구현 pytorch
hugozhl/hetu pytorch
pku-dair/hetu-galvatron pytorch

Tasks

GPU

Methods 이 논문이 사용한 방법론

Multi-Head Attention 설명 없음
Attention 설명 없음
Pruning 설명 없음
Layer Normalization Unlike batch normalization, Layer Normalization directly estimates the normalization statistics from the summed inputs…
Adam 설명 없음
Softmax The Softmax output function transforms a previous layer's output into a vector of probabilities. It is commonly used for multiclass classification. Given an input vector $x$…
Dropout Dropout is a regularization technique for neural networks that drops a unit (along with connections) at training time with a specified probability $p$ (a common value is…
BPE Byte Pair Encoding, or BPE, is a subword segmentation algorithm that encodes rare and unknown words as sequences of subword units. The intuition is that various word…

Similar Papers 제목 키워드 기반

Improving Automatic Parallel Training via Balanced Memory Workload Optimization

2023-07-05 · Yujie Wang, Youhe Jiang, Xupeng Miao, Fangcheng Fu 외

Transformer models have emerged as the leading approach for achieving state-of-the-art performance across various application domains, serving as the foundation for advanced large-scale deep learning (DL) models. However…

GPUNavigate

Galvatron: An Automatic Distributed System for Efficient Foundation Model Training

2025-04-30 · Xinyi Liu, Yujie Wang, Shenhan Zhu, Fangcheng Fu 외

Galvatron is a distributed system for efficiently training large-scale Foundation Models. It overcomes the complexities of selecting optimal parallelism strategies by automatically identifying the most efficient hybrid s…

Benchmarking

mLoRA: Fine-Tuning LoRA Adapters via Highly-Efficient Pipeline Parallelism in Multiple GPUs

2023-12-05 · Zhengmao Ye, Dengchun Li, Zetao Hu, Tingfeng Lan 외

Transformer-based, pre-trained large language models (LLMs) have demonstrated outstanding performance across diverse domains, particularly in the emerging {\em pretrain-then-finetune} paradigm. Low-Rank Adaptation (LoRA)…

GPULarge Language Modelparameter-efficient fine-tuningScheduling

Harnessing Manycore Processors with Distributed Memory for Accelerated Training of Sparse and Recurrent Models

2023-11-07 · Jan Finkbeiner, Thomas Gmeinder, Mark Pupilli, Alexander Titterton 외

Current AI training infrastructure is dominated by single instruction multiple data (SIMD) and systolic array architectures, such as Graphics Processing Units (GPUs) and Tensor Processing Units (TPUs), that excel at acce…

Efficient Neural Network

Beyond 2:4: exploring V:N:M sparsity for efficient transformer inference on GPUs

2024-10-21 · Kang Zhao, Tao Yuan, Han Bao, Zhenfeng Su 외

To date, 2:4 sparsity has stood as the only sparse pattern that can be accelerated using sparse tensor cores on GPUs. In practice, 2:4 sparsity often possesses low actual speedups ($\leq 1.3$) and requires fixed sparse r…