paper-with-me

Papers

JointSQ: Joint Sparsification-Quantization for Distributed Learning

2024-01-01 · CVPR 2024 1 · Weiying Xie, Haowei Li, Jitao Ma, Yunsong Li, Jie Lei, Donglai Liu, Leyuan Fang

Gradient sparsification and quantization offer a promising prospect to alleviate the communication overhead problem in distributed learning. However direct combination of the two results in suboptimal solutions due to the fact that sparsification and quantization haven't been learned together. In this paper we propose Joint Sparsification-Quantization (JointSQ) inspired by the discovery that sparsification can be treated as 0-bit quantization regardless of architectures. Specifically we mathematically formulate JointSQ as a mixed-precision quantization problem expanding the solution space. It can be solved by the designed MCKP-Greedy algorithm. Theoretical analysis demonstrates the minimal compression noise of JointSQ and extensive experiments on various network architectures including CNN RNN and Transformer also validate this point. Under the introduction of computation overhead consistent with or even lower than previous methods JointSQ achieves a compression ratio of 1000xon different models while maintaining near-lossless accuracy and brings 1.4xto 2.9xspeedup over existing methods.

📄 PDF Abstract BibTeX

Code (1)

HaoweiLi778/JointSQ pytorch

Tasks

Quantization

Methods 이 논문이 사용한 방법론

Attention 설명 없음
Residual Connection 설명 없음
Gradient Sparsification Gradient Sparsification is a technique for distributed training that sparsifies stochastic gradients to reduce the communication cost, with minor increase in the number of…
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 $x$…
Layer Normalization Unlike batch normalization, Layer Normalization directly estimates the normalization statistics from the summed inputs…
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…
Label Smoothing Label Smoothing is a regularization technique that introduces noise for the labels. This accounts for the fact that datasets may have mistakes in them, so maximizing the…
Adam 설명 없음

Similar Papers 제목 키워드 기반

Theoretically Better and Numerically Faster Distributed Optimization with Smoothness-Aware Quantization Techniques

2021-06-07 · Bokun Wang, Mher Safaryan, Peter Richtárik

To address the high communication costs of distributed machine learning, a large body of work has been devoted in recent years to designing various compression strategies, such as sparsification and quantization, and opt…

BIG-bench Machine LearningDistributed OptimizationQuantization

A Linear Speedup Analysis of Distributed Deep Learning with Sparse and Quantized Communication

2018-12-01 · NeurIPS 2018 12 · Peng Jiang, Gagan Agrawal

The large communication overhead has imposed a bottleneck on the performance of distributed Stochastic Gradient Descent (SGD) for training deep neural networks. Previous works have demonstrated the potential of using gr…

Quantization

Efficient Distributed Training through Gradient Compression with Sparsification and Quantization Techniques

2024-12-07 · Shruti Singh, Shantanu Kumar

This study investigates the impact of gradient compression on distributed training performance, focusing on sparsification and quantization techniques, including top-k, DGC, and QSGD. In baseline experiments, random-k co…

Quantization

Qsparse-local-SGD: Distributed SGD with Quantization, Sparsification and Local Computations

2019-12-01 · NeurIPS 2019 12 · Debraj Basu, Deepesh Data, Can Karakus, Suhas Diggavi

Communication bottleneck has been identified as a significant issue in distributed optimization of large-scale learning models. Recently, several approaches to mitigate this problem have been proposed, including differen…

Distributed OptimizationQuantization

Qsparse-local-SGD: Distributed SGD with Quantization, Sparsification, and Local Computations

2019-06-06 · Debraj Basu, Deepesh Data, Can Karakus, Suhas Diggavi

Communication bottleneck has been identified as a significant issue in distributed optimization of large-scale learning models. Recently, several approaches to mitigate this problem have been proposed, including differen…

Distributed OptimizationQuantization