paper-with-me

Papers

Pushing the Limits of Sparsity: A Bag of Tricks for Extreme Pruning

2024-11-20 · Andy Li, Aiden Durrant, Milan Markovic, Lu Yin, Georgios Leontidis

Pruning of deep neural networks has been an effective technique for reducing model size while preserving most of the performance of dense networks, crucial for deploying models on memory and power-constrained devices. While recent sparse learning methods have shown promising performance up to moderate sparsity levels such as 95% and 98%, accuracy quickly deteriorates when pushing sparsities to extreme levels. Obtaining sparse networks at such extreme sparsity levels presents unique challenges, such as fragile gradient flow and heightened risk of layer collapse. In this work, we explore network performance beyond the commonly studied sparsities, and propose a collection of techniques that enable the continuous learning of networks without accuracy collapse even at extreme sparsities, including 99.90%, 99.95% and 99.99% on ResNet architectures. Our approach combines 1) Dynamic ReLU phasing, where DyReLU initially allows for richer parameter exploration before being gradually replaced by standard ReLU, 2) weight sharing which reuses parameters within a residual layer while maintaining the same number of learnable parameters, and 3) cyclic sparsity, where both sparsity levels and sparsity patterns evolve dynamically throughout training to better encourage parameter exploration. We evaluate our method, which we term Extreme Adaptive Sparse Training (EAST) at extreme sparsities using ResNet-34 and ResNet-50 on CIFAR-10, CIFAR-100, and ImageNet, achieving significant performance improvements over state-of-the-art methods we compared with.

📄 PDF Abstract BibTeX arXiv:2411.13545

Code (0)

등록된 구현이 없습니다.

Tasks

Sparse Learning

Methods 이 논문이 사용한 방법론

ReLU How Do I Communicate to Expedia? How Do I Communicate to Expedia? – Call ☎️ +1-(888) 829 (0881) or +1-805-330-4056 or +1-805-330-4056 for Live Support & Special Travel…
Average Pooling 설명 없음
Convolution A convolution is a type of matrix operation, consisting of a kernel, a small matrix of weights, that slides over input data performing element-wise multiplication with the…
Global Average Pooling Global Average Pooling is a pooling operation designed to replace fully connected layers in classical CNNs. The idea is to generate one feature map for each corresponding…
Kaiming Initialization 설명 없음
Max Pooling Max Pooling is a pooling operation that calculates the maximum value for patches of a feature map, and uses it to create a downsampled (pooled) feature map. It is usually…

Similar Papers 제목 키워드 기반

TRIM: Achieving Extreme Sparsity with Targeted Row-wise Iterative Metric-driven Pruning

2025-05-22 · Florentin Beck, William Rudman, Carsten Eickhoff

Large Language Models (LLMs) present significant computational and memory challenges due to their extensive size, making pruning essential for their efficient deployment. Existing one-shot pruning methods often apply uni…

Beyond One-Way Pruning: Bidirectional Pruning-Regrowth for Extreme Accuracy-Sparsity Tradeoff

2025-11-11 · Junchen Liu, Yi Sheng arxiv

As a widely adopted model compression technique, model pruning has demonstrated strong effectiveness across various architectures. However, we observe that when sparsity exceeds a certain threshold, both iterative and on…

Model Compression

The Unseen Frontier: Pushing the Limits of LLM Sparsity with Surrogate-Free ADMM

2025-10-02 · Kwanhee Lee, Hyeondo Jang, Dongyeop Lee, Dan Alistarh 외 arxiv

Neural network pruning is a promising technique to mitigate the excessive computational and memory requirements of large language models (LLMs). Despite its promise, however, progress in this area has diminished, as conv…

Network Pruning

Towards Higher Ranks via Adversarial Weight Pruning

2023-11-29 · NeurIPS 2023 11 · Yuchuan Tian, Hanting Chen, Tianyu Guo, Chao Xu 외

Convolutional Neural Networks (CNNs) are hard to deploy on edge devices due to its high computation and storage complexities. As a common practice for model compression, network pruning consists of two major categories: …

Model CompressionNetwork Pruning

FedPaI: Achieving Extreme Sparsity in Federated Learning via Pruning at Initialization

2025-04-01 · Haonan Wang, Zeli Liu, Kajimusugura Hoshino, Tuo Zhang 외

Federated Learning (FL) enables distributed training on edge devices but faces significant challenges due to resource constraints in edge environments, impacting both communication and computational efficiency. Existing …

Computational EfficiencyFederated Learning