paper-with-me

Papers

Deep Neural Compression Via Concurrent Pruning and Self-Distillation

2021-09-30 · James O' Neill, Sourav Dutta, Haytham Assem

Pruning aims to reduce the number of parameters while maintaining performance close to the original network. This work proposes a novel \emph{self-distillation} based pruning strategy, whereby the representational similarity between the pruned and unpruned versions of the same network is maximized. Unlike previous approaches that treat distillation and pruning separately, we use distillation to inform the pruning criteria, without requiring a separate student network as in knowledge distillation. We show that the proposed {\em cross-correlation objective for self-distilled pruning} implicitly encourages sparse solutions, naturally complementing magnitude-based pruning criteria. Experiments on the GLUE and XGLUE benchmarks show that self-distilled pruning increases mono- and cross-lingual language model performance. Self-distilled pruned models also outperform smaller Transformers with an equal number of parameters and are competitive against (6 times) larger distilled networks. We also observe that self-distillation (1) maximizes class separability, (2) increases the signal-to-noise ratio, and (3) converges faster after pruning steps, providing further insights into why self-distilled pruning improves generalization.

📄 PDF Abstract BibTeX arXiv:2109.15014

Code (0)

등록된 구현이 없습니다.

Tasks

Knowledge DistillationLanguage ModelingLanguage Modelling

Methods 이 논문이 사용한 방법론

Pruning 설명 없음

Similar Papers 제목 키워드 기반

EPSD: Early Pruning with Self-Distillation for Efficient Model Compression

2024-01-31 · Dong Chen, Ning Liu, Yichen Zhu, Zhengping Che 외

Neural network compression techniques, such as knowledge distillation (KD) and network pruning, have received increasing attention. Recent work `Prune, then Distill' reveals that a pruned student-friendly teacher network…

Knowledge DistillationModel CompressionNetwork PruningNeural Network Compression

Constrained Edge AI Deployment: Fine-Tuning vs Distillation for LLM Compression

2025-05-13 · Jacob Sander, David Moe, Achraf Cohen, Brent Venable 외

Modern foundational models are often compressed via a combination of structured pruning and re-training to meet the strict compute, memory, and connectivity constraints of edge deployments. While state-of-the-art pruning…

DPHuBERT: Joint Distillation and Pruning of Self-Supervised Speech Models

2023-05-28 · Yifan Peng, Yui Sudo, Shakeel Muhammad, Shinji Watanabe

Self-supervised learning (SSL) has achieved notable success in many speech processing tasks, but the large model size and heavy computational cost hinder the deployment. Knowledge distillation trains a small student mode…

Knowledge DistillationSelf-Supervised Learning

Post-Training Probability Manifold Correction via Structured SVD Pruning and Self-Referential Distillation

2026-01-30 · Aaron R. Flouro, Shawn P. Chadwick arxiv

Large language models are expensive to deploy. We introduce Sparse Knowledge Distillation (SparseKD), a post-training method that compresses transformer models by combining structured SVD pruning with self-referential kn…

Knowledge Distillation

Is Smaller Always Faster? Tradeoffs in Compressing Self-Supervised Speech Transformers

2022-11-17 · Tzu-Quan Lin, Tsung-Huan Yang, Chun-Yao Chang, Kuang-Ming Chen 외

Transformer-based self-supervised models have achieved remarkable success in speech processing, but their large size and high inference cost present significant challenges for real-world deployment. While numerous compre…

Knowledge DistillationModel CompressionSelf-Supervised Learning