# Training Recipe for N:M Structured Sparsity with Decaying Pruning Mask

arXiv: https://arxiv.org/abs/2209.07617 · HTML: https://mlsanity.com/paper/2209.07617/

## Key facts

- Cited by 11 papers in the map (data through 14 Oct 2026).
- Builds on 51 of its 59 references that are in the map.
- First posted on arXiv 15 Sep 2022.

## Builds on (most cited first, 30 of 51)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer](https://mlsanity.com/paper/1910.10683/) (2019)
- [DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter](https://mlsanity.com/paper/1910.01108/) (2019)
- [OPT: Open Pre-trained Transformer Language Models](https://mlsanity.com/paper/2205.01068/) (2022)
- [Deep Compression: Compressing Deep Neural Networks with Pruning, Trained Quantization and Huffman Coding](https://mlsanity.com/paper/1510.00149/) (2015)
- [The Lottery Ticket Hypothesis: Finding Sparse, Trainable Neural Networks](https://mlsanity.com/paper/1803.03635/) (2018)
- [Learning both Weights and Connections for Efficient Neural Networks](https://mlsanity.com/paper/1506.02626/) (2015)
- [MiniLM: Deep Self-Attention Distillation for Task-Agnostic Compression of Pre-Trained Transformers](https://mlsanity.com/paper/2002.10957/) (2020)
- [TinyBERT: Distilling BERT for Natural Language Understanding](https://mlsanity.com/paper/1909.10351/) (2019)
- [Pruning Filters for Efficient ConvNets](https://mlsanity.com/paper/1608.08710/) (2016)
- [SNIP: Single-shot Network Pruning based on Connection Sensitivity](https://mlsanity.com/paper/1810.02340/) (2018)
- [Channel Pruning for Accelerating Very Deep Neural Networks](https://mlsanity.com/paper/1707.06168/) (2017)
- [To prune, or not to prune: exploring the efficacy of pruning for model compression](https://mlsanity.com/paper/1710.01878/) (2017)
- [Rethinking the Value of Network Pruning](https://mlsanity.com/paper/1810.05270/) (2018)
- [Learning Structured Sparsity in Deep Neural Networks](https://mlsanity.com/paper/1608.03665/) (2016)
- [Importance Estimation for Neural Network Pruning](https://mlsanity.com/paper/1906.10771/) (2019)
- [Rigging the Lottery: Making All Tickets Winners](https://mlsanity.com/paper/1911.11134/) (2019)
- [MobileBERT: a Compact Task-Agnostic BERT for Resource-Limited Devices](https://mlsanity.com/paper/2004.02984/) (2020)
- [The State of Sparsity in Deep Neural Networks](https://mlsanity.com/paper/1902.09574/) (2019)
- [Q-BERT: Hessian Based Ultra Low Precision Quantization of BERT](https://mlsanity.com/paper/1909.05840/) (2019)
- [Scalable Training of Artificial Neural Networks with Adaptive Sparse Connectivity inspired by Network Science](https://mlsanity.com/paper/1707.04780/) (2017)
- [Q8BERT: Quantized 8Bit BERT](https://mlsanity.com/paper/1910.06188/) (2019)
- [I-BERT: Integer-only BERT Quantization](https://mlsanity.com/paper/2101.01321/) (2021)
- [Comparing Rewinding and Fine-tuning in Neural Network Pruning](https://mlsanity.com/paper/2003.02389/) (2020)
- [Dynamic Network Surgery for Efficient DNNs](https://mlsanity.com/paper/1608.04493/) (2016)
- [Accelerating Sparse Deep Neural Networks](https://mlsanity.com/paper/2104.08378/) (2021)
- [Efficient Neural Audio Synthesis](https://mlsanity.com/paper/1802.08435/) (2018)

## Built on by (most cited first, 11 of 11)

- [MaskLLM: Learnable Semi-Structured Sparsity for Large Language Models](https://mlsanity.com/paper/2409.17481/) (2024)
- [STEP: Learning N:M Structured Sparsity Masks from Scratch with Precondition](https://mlsanity.com/paper/2302.01172/) (2023)
- [SLoPe: Double-Pruned Sparse Plus Lazy Low-Rank Adapter Pretraining of LLMs](https://mlsanity.com/paper/2405.16325/) (2024)
- [Effective Interplay between Sparsity and Quantization: From Theory to Practice](https://mlsanity.com/paper/2405.20935/) (2024)
- [Preserving Deep Representations In One-Shot Pruning: A Hessian-Free Second-Order Optimization Framework](https://mlsanity.com/paper/2411.18376/) (2024)
- [Enabling Unstructured Sparse Acceleration on Structured Sparse Accelerators](https://mlsanity.com/paper/2403.07953/) (2024)
- [Just CHOP: Embarrassingly Simple LLM Compression](https://mlsanity.com/paper/2305.14864/) (2023)
- [Pruning for Sparse Diffusion Models based on Gradient Flow](https://mlsanity.com/paper/2501.09464/) (2025)
- [PermLLM: Learnable Channel Permutation for N:M Sparse Large Language Models](https://mlsanity.com/paper/2510.10136/) (2025)
- [Compression Trinity: Exploring Sparsity, Quantization, and Low-Rank Approximations for LLM Compression](https://mlsanity.com/paper/2608.24070/) (2026)
- [GroupMask: Layer-Adaptive Group-wise Sparsity for Semi-Structured LLM Pruning](https://mlsanity.com/paper/2609.33977/) (2026)

## Abstract (arXiv)

Sparsity has become one of the promising methods to compress and accelerate Deep Neural Networks (DNNs). Among different categories of sparsity, structured sparsity has gained more attention due to its efficient execution on modern accelerators. Particularly, N:M sparsity is attractive because there are already hardware accelerator architectures that can leverage certain forms of N:M structured sparsity to yield higher compute-efficiency. In this work, we focus on N:M sparsity and extensively study and evaluate various training recipes for N:M sparsity in terms of the trade-off between model accuracy and compute cost (FLOPs). Building upon this study, we propose two new decay-based pruning methods, namely "pruning mask decay" and "sparse structure decay". Our evaluations indicate that these proposed methods consistently deliver state-of-the-art (SOTA) model accuracy, comparable to unstructured sparsity, on a Transformer-based model for a translation task. The increase in the accuracy of the sparse model using the new training recipes comes at the cost of marginal increase in the total training compute (FLOPs).