ghimiredhikura/Awasome-Pruning

Awesome Pruning. ✅ Curated Resources for Neural Network Pruning.

178

150 commits

updated May 8, 2026

github.com/ghimiredhikura/Awasome-Pruning

See the code

README

Awesome Pruning

Visitors

A curated list of neural network pruning papers, surveys, toolboxes, and code resources across CNNs, ViTs, LLMs, VLMs/MLLMs, diffusion, and deployment-aware compression.

Maintained by Deepak Ghimire and the community

Inspired by he-y/Awesome-Pruning

[!NOTE] Pull requests for missing papers, code links, and corrections are welcome. If this repository helps your work, please cite it using the BibTeX in Citation. Last updated: 2026-05-08. Recent additions emphasize pruning for LLMs, VLMs/MLLMs, ViTs, diffusion/3DGS, structured sparsity, semi-structured sparsity, and deployment-oriented compression.

Section	Quick Links
Conference Publications	`2026` `2025` `2024` `2023` `2022` `2021` `2020` `2019` `2018` `2017`
Journal Publications	`2026` `2025` `2024` `2023` `2022` `2021` `2020`
Survey Articles	`2020~2026`
Other Publications	`2022~2026`
Pruning Software and Toolbox	`2019~2026`
Citation	`BibTeX`

Legend

Symbol	Meaning
`U`	Unstructured / weight pruning
`S`	Structured / filter / channel / neuron / head / layer pruning
`SS`	Semi-structured sparsity, e.g., N:M / 2:4
`T`	Token / patch / KV-cache pruning or token reduction
`D`	Dynamic / input-adaptive pruning
`Q`	Joint pruning and quantization
`A`	Official / author implementation
`O`	Unofficial / third-party implementation
`-`	Code not found or not clearly public

Quick Recent Trends

Trend	Representative Methods
LLM structural pruning	LLM-Pruner, Sheared LLaMA, LoRAPrune, 2SSP, PAT, D2 Prune, ARMOR
LLM unstructured / semi-structured pruning	SparseGPT, Wanda, RIA, Wanda++, SLoRB, PermLLM
VLM / MLLM visual-token pruning	ATP-LLaVA, DivPrune, PACT, TopV, HiMAP, DyCoke, SGL
General automatic structured pruning	DepGraph / Torch-Pruning, OTOv2/HESSO, GETA
ViT token pruning / merging	ToMe, Token Cropr, Zero-TPrune, Token Fusion
Hardware/deployment-aware compression	HALP, GETA, NPAS, APQ, 2:4 pruning, pruning+quantization

Conference Publications

2026

Venue	Title	Area	Type	Code	Notes
`ICLR`	ARMOR: High-Performance Semi-Structured Pruning via Block-Diagonal Masking	LLM	`SS`	-	Uses block-diagonal structure for high-performance semi-structured pruning.
`AAAI`	D2 Prune: Sparsifying Large Language Models via Dual Distribution-Aware Calibration	LLM	`U/SS`	-	Addresses activation distribution shift and long-tail activation behavior.
`EACL Industry`	Iterative Structured Pruning for Large Language Models	LLM	`S`	-	Iterative structured pruning with hybrid calibration for downstream generalization.
`CPAL`	Understanding Neural Network Pruning via Infinite Width Graph Limits	Theory	`U/S`	-	Theoretical graphon perspective on sparse networks induced by pruning.
`arXiv`	Efficient Post-Training Pruning of Large Language Models with Statistical Correction	LLM	`U/S`	-	Post-training pruning with statistical correction.
`arXiv`	From Local to Global: Revisiting Structured Pruning for Large Language Models	LLM	`S`	-	Global/task-aware view of structured pruning; arXiv version updated in 2026.
`arXiv`	GETA-3DGS: Automatic Joint Structured Pruning and Quantization for 3D Gaussian Splatting	3DGS	`S/Q`	-	Extends joint pruning-quantization ideas to 3D Gaussian Splatting.

2025

Venue	Title	Area	Type	Code	Notes
`CVPR`	Automatic Joint Structured Pruning and Quantization for Efficient Neural Network Training and Compression	General	`S/Q`	PyTorch[A]	GETA; architecture-agnostic joint structured pruning + mixed-precision QAT.
`CVPR`	ATP-LLaVA: Adaptive Token Pruning for Large Vision Language Models	LVLM	`T/D`	-	Layer-wise and instance-wise adaptive visual-token pruning.
`CVPR`	DivPrune: Diversity-based Visual Token Pruning for Large Multimodal Models	MLLM	`T`	PyTorch[A]	Training-free, calibration-free diversity-based visual-token selection.
`CVPR`	Token Cropr: Faster ViTs for Quite a Few Tasks	ViT	`T/D`	PyTorch[A]	End-to-end token pruner; reported 1.5×–4× speedups across vision tasks.
`CVPR`	PACT: Pruning and Clustering-Based Token Reduction for Faster Visual Language Models	VLM	`T`	PyTorch[A]	Combines pruning of irrelevant tokens with clustering/merging of redundant tokens.
`CVPR`	TopV: Compatible Token Pruning with Inference Time Optimization for Fast and Low-Memory Multimodal Vision Language Model	MLLM	`T/D`	-	Prefill-stage token pruning compatible with FlashAttention and KV-cache reduction.
`CVPR`	Lifting the Veil on Visual Information Flow in MLLMs: Unlocking Efficient Visual Token Pruning	MLLM	`T/D`	-	Introduces HiMAP, a modality-aware plug-and-play visual-token pruning method.
`CVPR`	Libra-Merging: Importance-Redundancy and Pruning-Merging Trade-off for Acceleration Plug-in in Large Vision-Language Models	LVLM	`T`	-	Balances pruning and merging to handle token importance/redundancy trade-off.
`CVPR`	DyCoke: Dynamic Compression of Tokens for Fast Video Large Language Models	Video-LLM	`T/D`	-	Training-free temporal compression and dynamic KV-cache reduction.
`CVPR`	A Stitch in Time Saves Nine: Small VLM is a Precise Guidance for Accelerating Large VLMs	VLM	`T/D`	-	Small VLM guidance for visual-token pruning in a larger VLM.
`ICLR`	The Unreasonable Ineffectiveness of the Deeper Layers	LLM	`S`	-	Studies layer pruning and redundancy in deeper LLM layers.
`ICLR Workshop`	2SSP: A Two-Stage Framework for Structured Pruning of LLMs	LLM	`S`	PyTorch[A]	Combines width pruning and depth pruning for LLMs.
`OpenReview`	MoreauPruner: Robust Structured Pruning of Large Language Models	LLM	`S`	-	Robustness-oriented structured pruning under weight perturbations.
`OpenReview`	FASP: Fast and Accurate Structured Pruning of Large Language Models	LLM	`S`	-	Fast structured pruning using interlinked sequential-layer structure.
`OpenReview`	HESSO: Towards Automatic Efficient and User Friendly Structured Pruning	General	`S`	-	OTO-style automatic structured pruning workflow.
`AAAI`	Toward Adaptive Large Language Models Structured Pruning	LLM	`S/D`	-	Multi-granularity adaptive structured pruning for LLMs.
`AAAI`	PAT: Pruning-Aware Tuning for Large Language Models	LLM	`S`	PyTorch[A]	Tunes LLMs while considering later pruning.
`AAAI`	Pruning Large Language Models with Semi-Structural Adaptive Sparse Training	LLM	`SS`	-	Semi-structured LLM pruning with adaptive sparse training; includes SLoRB.
`COLING`	Enhancing One-Shot Pruned Pre-trained Language Models Through Sparse-Dense-Sparse Strategy	PLM	`U`	-	SDS: Sparse-Dense-Sparse restoration strategy after one-shot pruning.
`ACL Findings`	Wanda++: Pruning Large Language Models via Regional Gradients	LLM	`U/SS`	-	Adds efficient block/regional gradient information to Wanda-like pruning.
`EMNLP`	On Pruning State-Space LLMs	SSM-LLM	`S/U`	-	Studies pruning behavior for state-space language models.
`BMVC`	Explainability-Aware Structured Pruning for Efficient Neural Networks	CNN/ViT	`S`	-	Structured pruning guided by explainability.
`arXiv`	NIRVANA: Structured Pruning Reimagined for Large Language Models	LLM	`S`	-	Differentiates attention/MLP structure in LLM pruning.
`arXiv`	Týr-the-Pruner: Structural Pruning LLMs via Global Sparsity Allocation Search	LLM	`S`	-	Searches global sparsity allocation across layers.
`arXiv`	Pruning Large Language Models by Identifying and Preserving Functional Networks	LLM	`S`	PyTorch[A]	Brain-network-inspired functional-neuron preservation.
`arXiv`	Pruning Weights but Not Truth: Safeguarding Truthfulness While Pruning LLMs	LLM Safety	`U/S`	-	Truthfulness-aware pruning using activation outlier signals.
`arXiv`	PermLLM: Learnable Channel Permutation for N:M Sparse Large Language Models	LLM	`SS`	-	Channel permutation for improved N:M sparse pruning.
`arXiv`	DSA: Discovering Sparsity Allocation for Layer-wise Pruning of Large Language Models	LLM	`U/S`	-	Automated sparsity allocation for layer-wise LLM pruning.
`arXiv`	OATS: Outlier-Aware Pruning Through Sparse and Low-Rank Decomposition	Transformer	`U/S`	-	Compresses transformer weights as sparse + low-rank components.

2024

Venue	Title	Area	Type	Code	Notes
`ICLR`	Towards Meta-Pruning via Optimal Transport	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`ICLR`	Towards Energy Efficient Spiking Neural Networks: An Unstructured Pruning Framework	SNN	`U`	PyTorch[A]	Unstructured sparsity / weight pruning method.
`ICLR`	Masks, Signs, And Learning Rate Rewinding	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`ICLR`	Scaling Laws for Sparsely-Connected Foundation Models	Sparsity Theory	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`ICLR`	Sparse Model Soups: A Recipe for Improved Pruning via Model Averaging	Sparsity Theory	`S`	-	Structured pruning for hardware-friendly compression.
`ICLR`	Adaptive Sharpness-Aware Pruning for Robust Sparse Networks	Sparsity Theory	`S`	-	Structured pruning for hardware-friendly compression.
`ICLR`	What Makes a Good Prune? Maximal Unstructured Pruning for Maximal Cosine Similarity	Sparsity Theory	`U`	PyTorch[A]	Unstructured sparsity / weight pruning method.
`ICLR`	In defense of parameter sharing for model-compression	General	`S/U`	-	-
`ICLR`	ECoFLaP: Efficient Coarse-to-Fine Layer-Wise Pruning for Vision-Language Models	LLM	`U`	-	Unstructured sparsity / weight pruning method.
`ICLR`	Data-independent Module-aware Pruning for Hierarchical Vision Transformers	ViT/Transformer	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`ICLR`	SWAP: Sparse Entropic Wasserstein Regression for Robust Network Pruning	Sparsity Theory	`S`	-	Structured pruning for hardware-friendly compression.
`ICLR`	Sparse Weight Averaging with Multiple Particles for Iterative Magnitude Pruning	Sparsity Theory	`U`	-	Unstructured sparsity / weight pruning method.
`ICLR`	Synergistic Patch Pruning for Vision Transformer: Unifying Intra- & Inter-Layer Patch Importance	ViT/Transformer	`S`	-	Token/patch reduction for transformer acceleration.
`ICLR`	FedP3: Federated Personalized and Privacy-friendly Network Pruning under Model Heterogeneity	Federated/Privacy	`S`	-	Pruning method for federated or distributed settings.
`ICLR`	The Need for Speed: Pruning Transformers with One Recipe	ViT/Transformer	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`ICLR`	SAS: Structured Activation Sparsification	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`CVPR`	OrthCaps: An Orthogonal CapsNet with Sparse Attention Routing and Pruning	Sparsity Theory	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`CVPR`	Zero-TPrune: Zero-Shot Token Pruning through Leveraging of the Attention Graph in Pre-Trained Transformers	ViT/Transformer	`S`	PyTorch[A]	Token/patch reduction for transformer acceleration.
`CVPR`	Finding Lottery Tickets in Vision Models via Data-driven Spectral Foresight Pruning	Sparsity Theory	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`CVPR`	BilevelPruning: Unified Dynamic and Static Channel Pruning for Convolutional Neural Networks	CNN	`S`	-	Input-adaptive or dynamic pruning strategy.
`CVPR`	FedMef: Towards Memory-efficient Federated Dynamic Pruning	Federated/Privacy	`S`	-	Pruning method for federated or distributed settings.
`CVPR`	Resource-Efficient Transformer Pruning for Finetuning of Large Models	ViT/Transformer	`S`	-	Structured pruning for hardware-friendly compression.
`CVPR`	Device-Wise Federated Network Pruning	Federated/Privacy	`S`	-	Pruning method for federated or distributed settings.
`CVPR`	Auto-Train-Once: Controller Network Guided Automatic Network Pruning from Scratch	General	`S`	-	Structured pruning for hardware-friendly compression.
`CVPR`	Jointly Training and Pruning CNNs via Learnable Agent Guidance and Alignment	CNN	`S`	-	Structured pruning for hardware-friendly compression.
`CVPR`	Diversity-aware Channel Pruning for StyleGAN Compression	GAN	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`CVPR`	MADTP: Multimodal Alignment-Guided Dynamic Token Pruning for Accelerating Vision-Language Transformer	VLM/MLLM	`S`	PyTorch[A]	Recent method for pruning or sparsifying large language models.
`AAAI`	Dynamic Feature Pruning and Consolidation for Occluded Person Re-Identification	General	`S`	-	Input-adaptive or dynamic pruning strategy.
`AAAI`	REPrune: Channel Pruning via Kernel Representative Selection	CNN	`S`	-	Structured pruning for hardware-friendly compression.
`AAAI`	Revisiting Gradient Pruning: A Dual Realization for Defending against Gradient Attacks	General	`S`	-	Structured pruning for hardware-friendly compression.
`AAAI`	IRPruneDet: Efficient Infrared Small Target Detection via Wavelet Structure-Regularized Soft Channel Pruning	Detection/Segmentation	`S`	-	Structured pruning for hardware-friendly compression.
`AAAI`	EPSD: Early Pruning with Self-Distillation for Efficient Model Compression	General	`S`	-	Structured pruning for hardware-friendly compression.
`WACV`	Pruning from Scratch via Shared Pruning Module and Nuclear norm-based Regularization	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`WACV`	Towards Better Structured Pruning Saliency by Reorganizing Convolution	GAN	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`WACV`	Torque based Structured Pruning for Deep Neural Network	General	`S`	-	Structured pruning for hardware-friendly compression.
`WACV`	Revisiting Token Pruning for Object Detection and Instance Segmentation	ViT/Transformer	`S`	PyTorch[A]	Token/patch reduction for transformer acceleration.
`WACV`	Token Fusion: Bridging the Gap Between Token Pruning and Token Merging	ViT/Transformer	`S`	-	Token/patch reduction for transformer acceleration.
`WACV`	PATROL: Privacy-Oriented Pruning for Collaborative Inference Against Model Inversion Attacks	Federated/Privacy	`S`	-	Structured pruning for hardware-friendly compression.
`ICLR`	Sheared LLaMA: Accelerating Language Model Pre-training via Structured Pruning	LLM	`S`	PyTorch[A]	Structured pruning + continued pretraining for smaller LLMs.
`ICLR`	Plug-and-Play: An Efficient Post-training Pruning Method for Large Language Models	LLM	`U`	-	RIA + channel permutation for post-training LLM pruning and N:M semi-structured sparsity.
`ACL Findings`	LoRAPrune: Structured Pruning Meets Low-Rank Parameter-Efficient Fine-Tuning	LLM	`S`	PyTorch[A]	LoRA-guided structured pruning with reduced memory cost.
`ECCV`	FastV: An Image is Worth 1/2 Tokens After Layer 2	LVLM	`T`	PyTorch[A]	Plug-and-play visual-token pruning for LMM inference acceleration.

2023

Venue	Title	Area	Type	Code	Notes
`NIPS`	Diff-Pruning: Structural Pruning for Diffusion Models	Diffusion	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`NIPS`	LLM-Pruner: On the Structural Pruning of Large Language Models	LLM	`S`	PyTorch[A]	Recent method for pruning or sparsifying large language models.
`ICCV`	Automatic Network Pruning via Hilbert-Schmidt Independence Criterion Lasso under Information Bottleneck Principle	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`ICCV`	Unified Data-Free Compression: Pruning and Quantization without Fine-Tuning	General	`S`	PyTorch[A]	Combines pruning with quantization/compression.
`ICCV`	Structural Alignment for Network Pruning through Partial Regularization	General	`S`	-	Structured pruning for hardware-friendly compression.
`ICCV`	Differentiable Transportation Pruning	General	`S`	-	Structured pruning for hardware-friendly compression.
`ICCV`	Dynamic Token Pruning in Plain Vision Transformers for Semantic Segmentation	ViT/Transformer	`S`	PyTorch[A]	Token/patch reduction for transformer acceleration.
`ICCV`	Towards Fairness-aware Adversarial Network Pruning	General	`S`	-	Structured pruning for hardware-friendly compression.
`ICCV`	Efficient Joint Optimization of Layer-Adaptive Weight Pruning in Deep Neural Networks	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`CVPR`	DepGraph: Towards Any Structural Pruning	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`CVPR`	X-Pruner: eXplainable Pruning for Vision Transformers	ViT/Transformer	`U/S`	-	-
`CVPR`	Joint Token Pruning and Squeezing Towards More Aggressive Compression of Vision Transformers	ViT/Transformer	`S`	PyTorch[A]	Token/patch reduction for transformer acceleration.
`CVPR`	Global Vision Transformer Pruning with Hessian-Aware Saliency	ViT/Transformer	`S`	-	Structured pruning for hardware-friendly compression.
`CVPR`	CP3: Channel Pruning Plug-in for Point-based Networks	3D/Point Cloud	`S`	-	Structured pruning for hardware-friendly compression.
`CVPR`	Training Debiased Subnetworks With Contrastive Weight Pruning	General	`U`	-	Unstructured sparsity / weight pruning method.
`CVPR`	Pruning Parameterization With Bi-Level Optimization for Efficient Semantic Segmentation on the Edge	Detection/Segmentation	`S`	-	Structured pruning for hardware-friendly compression.
`CVPR`	Structural Alignment for Network Pruning through Partial Regularization	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`ICLR`	JaxPruner: A concise library for sparsity research	Sparsity Theory	`U/S`	PyTorch[A]	-
`ICLR`	OTOv2: Automatic, Generic, User-Friendly	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`ICLR`	How I Learned to Stop Worrying and Love Retraining	General	`U`	PyTorch[A]	Unstructured sparsity / weight pruning method.
`ICLR`	Token Merging: Your ViT But Faster	ViT/Transformer	`U/S`	PyTorch[A]	Token/patch reduction for transformer acceleration.
`ICLR`	Revisiting Pruning at Initialization Through the Lens of Ramanujan Graphs	Sparsity Theory	`U`	PyTorch[A] (soon...)	Unstructured sparsity / weight pruning method.
`ICLR`	Unmasking the Lottery Ticket Hypothesis: What's Encoded in a Winning Ticket's Mask?	Sparsity Theory	`U`	-	Unstructured sparsity / weight pruning method.
`ICLR`	NTK-SAP: Improving neural network pruning by aligning training dynamics	General	`U`	-	Unstructured sparsity / weight pruning method.
`ICLR`	DFPC: Data flow driven pruning of coupled channels without data	CNN	`S`	PyTorch[A]	Data-free or calibration-light pruning method.
`ICLR`	TVSPrune - Pruning Non-discriminative filters via Total Variation separability of intermediate representations without fine tuning	CNN	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`ICLR`	Pruning Deep Neural Networks from a Sparsity Perspective	Sparsity Theory	`U`	PyTorch[A]	Unstructured sparsity / weight pruning method.
`ICLR`	A Unified Framework of Soft Threshold Pruning	General	`U`	PyTorch[A]	Unstructured sparsity / weight pruning method.
`WACV`	Calibrating Deep Neural Networks Using Explicit Regularisation and Dynamic Data Pruning	General	`S`	-	Input-adaptive or dynamic pruning strategy.
`WACV`	Attend Who Is Weak: Pruning-Assisted Medical Image Localization Under Sophisticated and Implicit Imbalances	Medical Imaging	`S`	-	Structured pruning for hardware-friendly compression.
`ICASSP`	WHC: Weighted Hybrid Criterion for Filter Pruning on Convolutional Neural Networks	CNN	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`ICML`	SparseGPT: Massive Language Models Can Be Accurately Pruned in One-Shot	LLM	`U/SS`	PyTorch[A]	First widely used one-shot pruning method for 10B–100B+ parameter GPT-family models.
`arXiv`	Wanda: A Simple and Effective Pruning Approach for Large Language Models	LLM	`U/SS`	PyTorch[A]	Weight magnitude × activation norm; no retraining or weight update.
`arXiv`	LoRAPrune: Structured Pruning Meets Low-Rank Parameter-Efficient Fine-Tuning	LLM	`S`	PyTorch[A]	Early version later appearing in ACL Findings 2024.

2022

Venue	Title	Area	Type	Code	Notes
`CVPR`	Interspace Pruning: Using Adaptive Filter Representations To Improve Training of Sparse CNNs	Sparsity Theory	`U`	-	Unstructured sparsity / weight pruning method.
`CVPR`	Revisiting Random Channel Pruning for Neural Network Compression	CNN	`S`	PyTorch[A] (soon...)	Structured pruning for hardware-friendly compression.
`CVPR`	Fire Together Wire Together: A Dynamic Pruning Approach With Self-Supervised Mask Prediction	General	`S`	PyTorch[A]	Input-adaptive or dynamic pruning strategy.
`CVPR`	When to Prune? A Policy towards Early Structural Pruning	General	`S`	-	Structured pruning for hardware-friendly compression.
`CVPR`	Dreaming to Prune Image Deraining Networks	Low-level Vision	`S`	-	Structured pruning for hardware-friendly compression.
`ICLR`	SOSP: Efficiently Capturing Global Correlations by Second-Order Structured Pruning	General	`S`	-	Structured pruning for hardware-friendly compression.
`ICLR`	Learning Pruning-Friendly Networks via Frank-Wolfe: One-Shot, Any-Sparsity, And No Retraining	Sparsity Theory	`U`	PyTorch[A]	Unstructured sparsity / weight pruning method.
`ICLR`	Revisit Kernel Pruning with Lottery Regulated Grouped Convolutions	Sparsity Theory	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`ICLR`	Dual Lottery Ticket Hypothesis	Sparsity Theory	`U`	PyTorch[A]	Unstructured sparsity / weight pruning method.
`NIPS`	SAViT: Structure-Aware Vision Transformer Pruning via Collaborative Optimization	ViT/Transformer	`S`	PyTorch[A](soon...)	Structured pruning for hardware-friendly compression.
`NIPS`	Structural Pruning via Latency-Saliency Knapsack	Deployment	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`ACCV`	Filter Pruning via Automatic Pruning Rate Search⋆	CNN	`S`	-	Structured pruning for hardware-friendly compression.
`ACCV`	Network Pruning via Feature Shift Minimization	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`ACCV`	Lightweight Alpha Matting Network Using Distillation-Based Channel Pruning	Low-level Vision	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`ACCV`	Adaptive FSP : Adaptive Architecture Search with Filter Shape Pruning	CNN	`S`	-	Structured pruning for hardware-friendly compression.
`ECCV`	Soft Masking for Cost-Constrained Channel Pruning	CNN	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`WACV`	Hessian-Aware Pruning and Optimal Neural Implant	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`WACV`	PPCD-GAN: Progressive Pruning and Class-Aware Distillation for Large-Scale Conditional GANs Compression	GAN	`S`	-	Structured pruning for hardware-friendly compression.
`WACV`	Channel Pruning via Lookahead Search Guided Reinforcement Learning	CNN	`S`	-	Structured pruning for hardware-friendly compression.
`WACV`	EZCrop: Energy-Zoned Channels for Robust Output Pruning	Deployment	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`ICIP`	One-Cycle Pruning: Pruning Convnets With Tight Training Budget	CNN	`U`	-	Unstructured sparsity / weight pruning method.
`ICIP`	RAPID: A Single Stage Pruning Framework	General	`U`	-	Unstructured sparsity / weight pruning method.
`ICIP`	The Rise of the Lottery Heroes: Why Zero-Shot Pruning is Hard	Sparsity Theory	`U`	-	Unstructured sparsity / weight pruning method.
`ICIP`	Truncated Lottery Ticket for Deep Pruning	Sparsity Theory	`U`	-	Unstructured sparsity / weight pruning method.
`ICIP`	Which Metrics For Network Pruning: Final Accuracy? or Accuracy Drop?	General	`S/U`	-	-
`ISMSI`	Structured Pruning with Automatic Pruning Rate Derivation for Image Processing Neural Networks	Low-level Vision	`S`	-	Structured pruning for hardware-friendly compression.

2021

Venue	Title	Area	Type	Code	Notes
`ICLR`	Neural Pruning via Growing Regularization	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`ICLR`	Network Pruning That Matters: A Case Study on Retraining Variants	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`ICLR`	Layer-adaptive Sparsity for the Magnitude-based Pruning	Sparsity Theory	`U`	PyTorch[A]	Unstructured sparsity / weight pruning method.
`NIPS`	Only Train Once: A One-Shot Neural Network Training And Pruning Framework	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`CVPR`	NPAS: A Compiler-Aware Framework of Unified Network Pruning and Architecture Search for Beyond Real-Time Mobile Acceleration	Deployment	`S`	-	Structured pruning for hardware-friendly compression.
`CVPR`	Network Pruning via Performance Maximization	General	`S`	-	Structured pruning for hardware-friendly compression.
`CVPR`	Convolutional Neural Network Pruning With Structural Redundancy Reduction*	CNN	`S`	-	Structured pruning for hardware-friendly compression.
`CVPR`	Manifold Regularized Dynamic Network Pruning	General	`S`	PyTorch[A]	Input-adaptive or dynamic pruning strategy.
`CVPR`	Joint-DetNAS: Upgrade Your Detector With NAS, Pruning and Dynamic Distillation	General	`S`	-	Input-adaptive or dynamic pruning strategy.
`ICCV`	ResRep: Lossless CNN Pruning via Decoupling Remembering and Forgetting	CNN	`S`	-	Structured pruning for hardware-friendly compression.
`ICCV`	Achieving On-Mobile Real-Time Super-Resolution With Neural Architecture and Pruning Search	Low-level Vision	`S`	-	Structured pruning for hardware-friendly compression.
`ICCV`	GDP: Stabilized Neural Network Pruning via Gates With Differentiable Polarization*	General	`S`	-	Structured pruning for hardware-friendly compression.
`WACV`	Holistic Filter Pruning for Efficient Deep Neural Networks	CNN	`S`	-	Structured pruning for hardware-friendly compression.
`ICML`	Accelerate CNNs from Three Dimensions: A Comprehensive Pruning Framework	CNN	`S`	-	Structured pruning for hardware-friendly compression.
`ICML`	Group Fisher Pruning for Practical Network Compression	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.

2020

Venue	Title	Area	Type	Code	Notes
`CVPR`	HRank: Filter Pruning using High-Rank Feature Map	CNN	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`CVPR`	Towards efficient model compression via learned global ranking	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`CVPR`	Learning Filter Pruning Criteria for Deep Convolutional Neural Networks Acceleration	CNN	`S`	-	Structured pruning for hardware-friendly compression.
`CVPR`	Group Sparsity: The Hinge Between Filter Pruning and Decomposition for Network Compression	Sparsity Theory	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`CVPR`	APQ: Joint Search for Network Architecture, Pruning and Quantization Policy	General	`S`	PyTorch[A]	Combines pruning with quantization/compression.
`ICLR`	Budgeted Training: Rethinking Deep Neural Network Training Under Resource Constraints	General	`U`	-	Unstructured sparsity / weight pruning method.
`MLSys`	Shrinkbench: What is the State of Neural Network Pruning?	General	-	PyTorch[A]	-
`BMBS`	Similarity Based Filter Pruning for Efficient Super-Resolution Models	Low-level Vision	`S`	-	Structured pruning for hardware-friendly compression.

2019

Venue	Title	Area	Type	Code	Notes
`CVPR`	Filter Pruning via Geometric Median for Deep Convolutional Neural Networks Acceleration	CNN	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`CVPR`	Variational Convolutional Neural Network Pruning	CNN	`S`	-	Structured pruning for hardware-friendly compression.
`CVPR`	Towards Optimal Structured CNN Pruning via Generative Adversarial Learning	CNN	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`CVPR`	Partial Order Pruning: For Best Speed/Accuracy Trade-Off in Neural Architecture Search	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`CVPR`	Importance Estimation for Neural Network Pruning	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`ICLR`	The Lottery Ticket Hypothesis: Finding Sparse, Trainable Neural Networks	Sparsity Theory	`U`	PyTorch[A]	Unstructured sparsity / weight pruning method.
`ICLR`	SNIP: Single-shot Network Pruning based on Connection Sensitivity	ViT/Transformer	`U`	Tensorflow[A]	Unstructured sparsity / weight pruning method.
`ICCV`	MetaPruning: Meta-Learning for Automatic Neural Network Channel Pruning	CNN	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`ICCV`	Accelerate CNN via Recursive Bayesian Pruning	CNN	`S`	-	Structured pruning for hardware-friendly compression.

2018

Venue	Title	Area	Type	Code	Notes
`CVPR`	PackNet: Adding Multiple Tasks to a Single Network by Iterative Pruning	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`CVPR`	NISP: Pruning Networks Using Neuron Importance Score Propagation	General	`S`	-	Structured pruning for hardware-friendly compression.
`ICIP`	Online Filter Clustering and Pruning for Efficient Convnets	CNN	`S`	-	Structured pruning for hardware-friendly compression.
`IJCAI`	Soft Filter Pruning for Accelerating Deep Convolutional Neural Networks	CNN	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.

2017

Venue	Title	Area	Type	Code	Notes
`CVPR`	Designing Energy-Efficient Convolutional Neural Networks Using Energy-Aware Pruning	Deployment	`S`	-	Structured pruning for hardware-friendly compression.
`ICLR`	Pruning Filters for Efficient ConvNets	CNN	`S`	PyTorch[O]	Structured pruning for hardware-friendly compression.
`ICCV`	Channel Pruning for Accelerating Very Deep Neural Networks	CNN	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`ICCV`	ThiNet: A Filter Level Pruning Method for Deep Neural Network Compression	CNN	`S`	Caffe[A]	Structured pruning for hardware-friendly compression.
`ICCV`	Learning Efficient Convolutional Networks Through Network Slimming	CNN	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.

Journal Publications

2026

Journal	Title	Area	Type	Code	Notes
-	-	General	-	-	No peer-reviewed journal pruning papers were confidently added for 2026 yet; most 2026 additions are conference or preprint entries above.

2025

Journal	Title	Area	Type	Code	Notes
Neural Networks	OOPS: Outlier-aware and Quadratic Programming Based Structured Pruning for Large Language Models	LLM	`S`	-	Structured LLM pruning with outlier-aware quadratic-programming formulation.

2024

Journal	Title	Area	Type	Code	Notes
Neural Networks	Efficient tensor decomposition-based filter pruning	CNN	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
IEEE Trans. NNLS	Enhanced Network Compression Through Tensor Decompositions and Pruning	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
IEEE Transactions on Artificial Intelligence	Distilled Gradual Pruning with Pruned Fine-tuning	General	`U`	PyTorch[A]	Unstructured sparsity / weight pruning method.

2023

Journal	Title	Area	Type	Code	Notes
IEEE Trans Circuits Syst Video Technol	DCFP: Distribution Calibrated Filter Pruning for Lightweight and Accurate Long-tail Semantic Segmentation	Detection/Segmentation	`S`	-	Structured pruning for hardware-friendly compression.
IEEE Internet Things J.	SNPF: Sensitiveness Based Network Pruning Framework for Efficient Edge Computing	Deployment	`S`	-	Structured pruning for hardware-friendly compression.
IEEE Trans. NNLS	Manipulating Identical Filter Redundancy for Efficient Pruning on Deep and Complicated CNN	CNN	`S`	-	Structured pruning for hardware-friendly compression.
IEEE Trans. NNLS	Block-Wise Partner Learning for Model Compression	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
IEEE Trans. NNLS	Hierarchical Threshold Pruning Based on Uniform Response Criterion	General	`S`	-	Structured pruning for hardware-friendly compression.
IEEE Trans. NNLS	CATRO: Channel Pruning via Class-Aware Trace Ratio Optimization	CNN	`S`	-	Structured pruning for hardware-friendly compression.
IEEE Trans. NNLS	Adaptive Filter Pruning via Sensitivity Feedback	ViT/Transformer	`S`	-	Structured pruning for hardware-friendly compression.
Neurocomputing	Filter pruning with uniqueness mechanism in the frequency domain for efficient neural networks	CNN	`S`	-	Structured pruning for hardware-friendly compression.
IEEE Trans. PAMI	Compact Neural Network via Stacking Hybrid Units	General	`S`	-	Structured pruning for hardware-friendly compression.
IEEE Trans. PAMI	Performance-aware Approximation of Global Channel Pruning for Multitask CNNs	CNN	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
IEEE Trans. PAMI	Adaptive Search-and-Training for Robust and Efficient Network Pruning	General	`S`	-	Structured pruning for hardware-friendly compression.
Image Vis. Comput.	Loss-aware automatic selection of structured pruning criteria for deep neural network acceleration	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
Comput. Vis. Image Underst.	Feature independent Filter Pruning by Successive Layers analysis	CNN	`S`	-	Structured pruning for hardware-friendly compression.
IEEE Access	Differentiable Neural Architecture, Mixed Precision and Accelerator Co-Search	Deployment	`S`	-	Structured pruning for hardware-friendly compression.

2022

Journal	Title	Area	Type	Code	Notes
IEEE Trans. Image Process.	Efficient Layer Compression Without Pruning	General	`S`	-	Structured pruning for hardware-friendly compression.
IEEE Trans. PAMI	Learning to Explore Distillability and Sparsability: A Joint Framework for Model Compression	General	`S`	-	Structured pruning for hardware-friendly compression.
IEEE Trans. PAMI	1xN Pattern for Pruning Convolutional Neural Networks	CNN	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
IEEE Trans. NNLS	Filter Pruning by Switching to Neighboring CNNs With Good Attribute	CNN	`S`	-	Structured pruning for hardware-friendly compression.
IEEE Trans. NNLS	Model Pruning Enables Efficient Federated Learning on Edge Devices	Federated/Privacy	`S`	-	Pruning method for federated or distributed settings.
IEEE Trans. NNLS	DAIS: Automatic Channel Pruning via Differentiable Annealing Indicator Search	CNN	`S`	-	Structured pruning for hardware-friendly compression.
IEEE Trans. NNLS	Network Pruning Using Adaptive Exemplar Filters	CNN	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
IEEE Trans. NNLS	Carrying Out CNN Channel Pruning in a White Box	CNN	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
IEEE Trans. NNLS	Pruning Networks With Cross-Layer Ranking & k-Reciprocal Nearest Filters	CNN	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
IEEE Trans. NNLS	Filter Sketch for Network Pruning	CNN	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
Neurocomputing	FPFS: Filter-level pruning via distance weight measuring filter similarity	CNN	`S`	-	Structured pruning for hardware-friendly compression.
Neurocomputing	RUFP: Reinitializing unimportant filters for soft pruning	CNN	`S`	-	Structured pruning for hardware-friendly compression.
Neural Netw	HRel: Filter pruning based on High Relevance between activation maps and class labels	CNN	`S`	PyTorch[A]*	Structured pruning for hardware-friendly compression.
Comput. Intell. Neurosci.	Differentiable Network Pruning via Polarization of Probabilistic Channelwise Soft Masks	CNN	`S`	-	Structured pruning for hardware-friendly compression.
J. Syst. Archit.	Optimizing deep neural networks on intelligent edge accelerators via flexible-rate filter pruning	Deployment	`S`	-	Structured pruning for hardware-friendly compression.
Appl. Sci.	Magnitude and Similarity Based Variable Rate Filter Pruning for Efficient Convolution Neural Networks	Sparsity Theory	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
Sensors	Filter Pruning via Measuring Feature Map Information	CNN	`S`	-	Structured pruning for hardware-friendly compression.
IEEE Access	Automated Filter Pruning Based on High-Dimensional Bayesian Optimization	CNN	`S`	-	Structured pruning for hardware-friendly compression.
IEEE Signal Process. Lett.	A Low-Complexity Modified ThiNet Algorithm for Pruning Convolutional Neural Networks	CNN	`S`	-	Structured pruning for hardware-friendly compression.

2021

Journal	Title	Area	Type	Code	Notes
IEEE Trans. PAMI	Discrimination-Aware Network Pruning for Deep Model Compression	General	`S`	PyTorch[A]~	Structured pruning for hardware-friendly compression.

2020

Journal	Title	Area	Type	Code	Notes
IEEE Trans. NNLS	EDP: An Efficient Decomposition and Pruning Scheme for Convolutional Neural Network Compression	CNN	`S`	-	Structured pruning for hardware-friendly compression.
IEEE Access	Filter Pruning Without Damaging Networks Capacity	CNN	`S`	-	Structured pruning for hardware-friendly compression.
Electronics	Pruning Convolutional Neural Networks with an Attention Mechanism for Remote Sensing Image Classification	CNN	`S`	-	Structured pruning for hardware-friendly compression.

Survey Articles

Year	Venue	Title	Area	Type	Code	Notes
`2026`	`arXiv`	Towards Efficient Multimodal Large Language Models: A Survey on Token Compression	LLM	-	-	Survey / overview resource.
`2024`	`IEEE TPAMI`	A Survey on Deep Neural Network Pruning: Taxonomy, Comparison, Analysis, and Recommendations	Survey	-	-	Survey / overview resource.
`2024`	`arXiv`	A Survey on Model Compression for Large Language Models	LLM	-	-	Survey / overview resource.
`2023`	`Artif. Intell. Rev.`	Deep neural network pruning method based on sensitive layers and reinforcement learning	General	-	-	Survey / overview resource.
`2023`	`arVix`	A Survey on Deep Neural Network Pruning: Taxonomy, Comparison, Analysis, and Recommendations	Survey	-	-	Survey / overview resource.
`2023`	`arVix`	Structured Pruning for Deep Convolutional Neural Networks: A survey	CNN	-	-	Survey / overview resource.
`2022`	`Electronics`	A Survey on Efficient Convolutional Neural Networks and Hardware Acceleration	Deployment	-	-	Survey / overview resource.
`2022`	`I-SMAC`	A Survey on Filter Pruning Techniques for Optimization of Deep Neural Networks	CNN	-	-	Survey / overview resource.
`2021`	`JMLR`	Sparsity in Deep Learning: Pruning and growth for efficient inference and training in neural networks	Sparsity Theory	-	-	Survey / overview resource.
`2021`	`Neurocomputing`	Pruning and quantization for deep neural network acceleration: A survey	Survey	-	-	Survey / overview resource.
`2020`	`IEEE Access`	Methods for Pruning Deep Neural Networks	General	-	-	Survey / overview resource.

Other Publications

Venue	Title	Area	Type	Code	Notes
`2023`	`arVix`	Why is the State of Neural Network Pruning so Confusing? On the Fairness, Comparison Setup, and Trainability in Network Pruning	General	-	PyTorch[A]
`2023`	`arVix`	Ten Lessons We Have Learned in the New "Sparseland": A Short Handbook for Sparse Neural Network Researchers	Sparsity Theory	-	-
`2022`	`ICML`	Tutorial -- Sparsity in Deep Learning: Pruning and growth for efficient inference and training	Sparsity Theory	-	-

Pruning Software and Toolbox

Year	Title	Area	Type	Code	Notes
`2026`	LLM Compressor / vLLM compression tools	LLM	`U/S/Q/SS`	Python[A]	Recent method for pruning or sparsifying large language models.
`2025`	GETA	General	`S/Q`	PyTorch[A]	-
`2025`	PACT	General	`T`	PyTorch[A]	-
`2025`	DivPrune	General	`T`	PyTorch[A]	-
`2025`	Token Cropr	ViT/Transformer	`T`	PyTorch[A]	Token/patch reduction for transformer acceleration.
`2025`	2SSP	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`2024`	LLM-Shearing / Sheared LLaMA	LLM	`S`	PyTorch[A]	Recent method for pruning or sparsifying large language models.
`2024`	LoRAPrune	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`2023`	SparseGPT	LLM	`U/SS`	PyTorch[A]	Recent method for pruning or sparsifying large language models.
`2023`	Wanda	General	`U/SS`	PyTorch[A]	-
`2023`	UPop: Unified and Progressive Pruning for Compressing Vision-Language Transformers	VLM/MLLM	`S`	PyTorch[A]	Recent method for pruning or sparsifying large language models.
`2023`	DepGraph: Towards Any Structural Pruning	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`2023`	Torch-Pruning	Toolbox	`S`	PyTorch[A]	Software resource for pruning/compression workflows.
`2023`	JaxPruner: JaxPruner: A concise library for sparsity research	Sparsity Theory	`U/S`	PyTorch[A]	-
`2022`	FasterAI: Prune and Distill your models with FastAI and PyTorch	General	`U`	PyTorch[A]	Unstructured sparsity / weight pruning method.
`2022`	Simplify: A Python library for optimizing pruned neural networks	Toolbox	-	PyTorch[A]	Software resource for pruning/compression workflows.
`2021`	PyTorchViz [A small package to create visualizations of PyTorch execution graphs]	General	-	PyTorch[A]	-
`2020`	What is the State of Neural Network Pruning?	General	`S/U`	PyTorch[A]	-
`2019`	Official PyTorch Pruning Tool	Toolbox	`S/U`	PyTorch[A]	Software resource for pruning/compression workflows.

Citation

If this repository helps your research, please cite it as:

@misc{awesome_pruning,
  title  = {Awesome Pruning},
  author = {Deepak Ghimire and others},
  year   = {2026},
  note   = {Curated list of pruning papers, surveys, software, and code across CNNs, ViTs, LLMs, VLMs/MLLMs, diffusion, and deployment-aware compression}
}

awesome

pruning

pytorch

structured-pruning

unstructured-pruning

Contributors

140 commits

5 commits

4 commits

1 commits

Similar repos

hrcheng1066/awesome-pruning

309

liyunqianggyn/Awesome-LLMs-Pruning

Awesome LLM pruning papers all-in-one repository with integrating all useful resources and…

180

horseee/Awesome-Efficient-LLM

A curated list for Efficient Large Language Models

2,043

LumosJiang/Awesome-On-Device-LLMs

[Artificial Intelligence Review 2026] Awesome list and resources for on-device LLMs: model…

daixiangzi/Awesome-Token-Compress

A paper list of some recent works about Token Compress for Vit and VLM

957

cedrickchee/awesome-ml-model-compression

Awesome machine learning model compression research papers, quantization, tools, and learning…

546

AIoT-MLSys-Lab/Efficient-LLMs-Survey

[TMLR 2024] Efficient Large Language Models: A Survey

1,257

VainF/Torch-Pruning

[CVPR 2023] DepGraph: Towards Any Structural Pruning; LLMs, Vision Foundation Models, etc.

3,357

ghimiredhikura/Awasome-Pruning

Awesome Pruning. ✅ Curated Resources for Neural Network Pruning.

178

150 commits

updated May 8, 2026

github.com/ghimiredhikura/Awasome-Pruning

See the code

README

Awesome Pruning

Visitors

A curated list of neural network pruning papers, surveys, toolboxes, and code resources across CNNs, ViTs, LLMs, VLMs/MLLMs, diffusion, and deployment-aware compression.

Maintained by Deepak Ghimire and the community

Inspired by he-y/Awesome-Pruning

[!NOTE] Pull requests for missing papers, code links, and corrections are welcome. If this repository helps your work, please cite it using the BibTeX in Citation. Last updated: 2026-05-08. Recent additions emphasize pruning for LLMs, VLMs/MLLMs, ViTs, diffusion/3DGS, structured sparsity, semi-structured sparsity, and deployment-oriented compression.

Section	Quick Links
Conference Publications	`2026` `2025` `2024` `2023` `2022` `2021` `2020` `2019` `2018` `2017`
Journal Publications	`2026` `2025` `2024` `2023` `2022` `2021` `2020`
Survey Articles	`2020~2026`
Other Publications	`2022~2026`
Pruning Software and Toolbox	`2019~2026`
Citation	`BibTeX`

Legend

Symbol	Meaning
`U`	Unstructured / weight pruning
`S`	Structured / filter / channel / neuron / head / layer pruning
`SS`	Semi-structured sparsity, e.g., N:M / 2:4
`T`	Token / patch / KV-cache pruning or token reduction
`D`	Dynamic / input-adaptive pruning
`Q`	Joint pruning and quantization
`A`	Official / author implementation
`O`	Unofficial / third-party implementation
`-`	Code not found or not clearly public

Quick Recent Trends

Trend	Representative Methods
LLM structural pruning	LLM-Pruner, Sheared LLaMA, LoRAPrune, 2SSP, PAT, D2 Prune, ARMOR
LLM unstructured / semi-structured pruning	SparseGPT, Wanda, RIA, Wanda++, SLoRB, PermLLM
VLM / MLLM visual-token pruning	ATP-LLaVA, DivPrune, PACT, TopV, HiMAP, DyCoke, SGL
General automatic structured pruning	DepGraph / Torch-Pruning, OTOv2/HESSO, GETA
ViT token pruning / merging	ToMe, Token Cropr, Zero-TPrune, Token Fusion
Hardware/deployment-aware compression	HALP, GETA, NPAS, APQ, 2:4 pruning, pruning+quantization

Conference Publications

2026

Venue	Title	Area	Type	Code	Notes
`ICLR`	ARMOR: High-Performance Semi-Structured Pruning via Block-Diagonal Masking	LLM	`SS`	-	Uses block-diagonal structure for high-performance semi-structured pruning.
`AAAI`	D2 Prune: Sparsifying Large Language Models via Dual Distribution-Aware Calibration	LLM	`U/SS`	-	Addresses activation distribution shift and long-tail activation behavior.
`EACL Industry`	Iterative Structured Pruning for Large Language Models	LLM	`S`	-	Iterative structured pruning with hybrid calibration for downstream generalization.
`CPAL`	Understanding Neural Network Pruning via Infinite Width Graph Limits	Theory	`U/S`	-	Theoretical graphon perspective on sparse networks induced by pruning.
`arXiv`	Efficient Post-Training Pruning of Large Language Models with Statistical Correction	LLM	`U/S`	-	Post-training pruning with statistical correction.
`arXiv`	From Local to Global: Revisiting Structured Pruning for Large Language Models	LLM	`S`	-	Global/task-aware view of structured pruning; arXiv version updated in 2026.
`arXiv`	GETA-3DGS: Automatic Joint Structured Pruning and Quantization for 3D Gaussian Splatting	3DGS	`S/Q`	-	Extends joint pruning-quantization ideas to 3D Gaussian Splatting.

2025

Venue	Title	Area	Type	Code	Notes
`CVPR`	Automatic Joint Structured Pruning and Quantization for Efficient Neural Network Training and Compression	General	`S/Q`	PyTorch[A]	GETA; architecture-agnostic joint structured pruning + mixed-precision QAT.
`CVPR`	ATP-LLaVA: Adaptive Token Pruning for Large Vision Language Models	LVLM	`T/D`	-	Layer-wise and instance-wise adaptive visual-token pruning.
`CVPR`	DivPrune: Diversity-based Visual Token Pruning for Large Multimodal Models	MLLM	`T`	PyTorch[A]	Training-free, calibration-free diversity-based visual-token selection.
`CVPR`	Token Cropr: Faster ViTs for Quite a Few Tasks	ViT	`T/D`	PyTorch[A]	End-to-end token pruner; reported 1.5×–4× speedups across vision tasks.
`CVPR`	PACT: Pruning and Clustering-Based Token Reduction for Faster Visual Language Models	VLM	`T`	PyTorch[A]	Combines pruning of irrelevant tokens with clustering/merging of redundant tokens.
`CVPR`	TopV: Compatible Token Pruning with Inference Time Optimization for Fast and Low-Memory Multimodal Vision Language Model	MLLM	`T/D`	-	Prefill-stage token pruning compatible with FlashAttention and KV-cache reduction.
`CVPR`	Lifting the Veil on Visual Information Flow in MLLMs: Unlocking Efficient Visual Token Pruning	MLLM	`T/D`	-	Introduces HiMAP, a modality-aware plug-and-play visual-token pruning method.
`CVPR`	Libra-Merging: Importance-Redundancy and Pruning-Merging Trade-off for Acceleration Plug-in in Large Vision-Language Models	LVLM	`T`	-	Balances pruning and merging to handle token importance/redundancy trade-off.
`CVPR`	DyCoke: Dynamic Compression of Tokens for Fast Video Large Language Models	Video-LLM	`T/D`	-	Training-free temporal compression and dynamic KV-cache reduction.
`CVPR`	A Stitch in Time Saves Nine: Small VLM is a Precise Guidance for Accelerating Large VLMs	VLM	`T/D`	-	Small VLM guidance for visual-token pruning in a larger VLM.
`ICLR`	The Unreasonable Ineffectiveness of the Deeper Layers	LLM	`S`	-	Studies layer pruning and redundancy in deeper LLM layers.
`ICLR Workshop`	2SSP: A Two-Stage Framework for Structured Pruning of LLMs	LLM	`S`	PyTorch[A]	Combines width pruning and depth pruning for LLMs.
`OpenReview`	MoreauPruner: Robust Structured Pruning of Large Language Models	LLM	`S`	-	Robustness-oriented structured pruning under weight perturbations.
`OpenReview`	FASP: Fast and Accurate Structured Pruning of Large Language Models	LLM	`S`	-	Fast structured pruning using interlinked sequential-layer structure.
`OpenReview`	HESSO: Towards Automatic Efficient and User Friendly Structured Pruning	General	`S`	-	OTO-style automatic structured pruning workflow.
`AAAI`	Toward Adaptive Large Language Models Structured Pruning	LLM	`S/D`	-	Multi-granularity adaptive structured pruning for LLMs.
`AAAI`	PAT: Pruning-Aware Tuning for Large Language Models	LLM	`S`	PyTorch[A]	Tunes LLMs while considering later pruning.
`AAAI`	Pruning Large Language Models with Semi-Structural Adaptive Sparse Training	LLM	`SS`	-	Semi-structured LLM pruning with adaptive sparse training; includes SLoRB.
`COLING`	Enhancing One-Shot Pruned Pre-trained Language Models Through Sparse-Dense-Sparse Strategy	PLM	`U`	-	SDS: Sparse-Dense-Sparse restoration strategy after one-shot pruning.
`ACL Findings`	Wanda++: Pruning Large Language Models via Regional Gradients	LLM	`U/SS`	-	Adds efficient block/regional gradient information to Wanda-like pruning.
`EMNLP`	On Pruning State-Space LLMs	SSM-LLM	`S/U`	-	Studies pruning behavior for state-space language models.
`BMVC`	Explainability-Aware Structured Pruning for Efficient Neural Networks	CNN/ViT	`S`	-	Structured pruning guided by explainability.
`arXiv`	NIRVANA: Structured Pruning Reimagined for Large Language Models	LLM	`S`	-	Differentiates attention/MLP structure in LLM pruning.
`arXiv`	Týr-the-Pruner: Structural Pruning LLMs via Global Sparsity Allocation Search	LLM	`S`	-	Searches global sparsity allocation across layers.
`arXiv`	Pruning Large Language Models by Identifying and Preserving Functional Networks	LLM	`S`	PyTorch[A]	Brain-network-inspired functional-neuron preservation.
`arXiv`	Pruning Weights but Not Truth: Safeguarding Truthfulness While Pruning LLMs	LLM Safety	`U/S`	-	Truthfulness-aware pruning using activation outlier signals.
`arXiv`	PermLLM: Learnable Channel Permutation for N:M Sparse Large Language Models	LLM	`SS`	-	Channel permutation for improved N:M sparse pruning.
`arXiv`	DSA: Discovering Sparsity Allocation for Layer-wise Pruning of Large Language Models	LLM	`U/S`	-	Automated sparsity allocation for layer-wise LLM pruning.
`arXiv`	OATS: Outlier-Aware Pruning Through Sparse and Low-Rank Decomposition	Transformer	`U/S`	-	Compresses transformer weights as sparse + low-rank components.

2024

Venue	Title	Area	Type	Code	Notes
`ICLR`	Towards Meta-Pruning via Optimal Transport	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`ICLR`	Towards Energy Efficient Spiking Neural Networks: An Unstructured Pruning Framework	SNN	`U`	PyTorch[A]	Unstructured sparsity / weight pruning method.
`ICLR`	Masks, Signs, And Learning Rate Rewinding	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`ICLR`	Scaling Laws for Sparsely-Connected Foundation Models	Sparsity Theory	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`ICLR`	Sparse Model Soups: A Recipe for Improved Pruning via Model Averaging	Sparsity Theory	`S`	-	Structured pruning for hardware-friendly compression.
`ICLR`	Adaptive Sharpness-Aware Pruning for Robust Sparse Networks	Sparsity Theory	`S`	-	Structured pruning for hardware-friendly compression.
`ICLR`	What Makes a Good Prune? Maximal Unstructured Pruning for Maximal Cosine Similarity	Sparsity Theory	`U`	PyTorch[A]	Unstructured sparsity / weight pruning method.
`ICLR`	In defense of parameter sharing for model-compression	General	`S/U`	-	-
`ICLR`	ECoFLaP: Efficient Coarse-to-Fine Layer-Wise Pruning for Vision-Language Models	LLM	`U`	-	Unstructured sparsity / weight pruning method.
`ICLR`	Data-independent Module-aware Pruning for Hierarchical Vision Transformers	ViT/Transformer	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`ICLR`	SWAP: Sparse Entropic Wasserstein Regression for Robust Network Pruning	Sparsity Theory	`S`	-	Structured pruning for hardware-friendly compression.
`ICLR`	Sparse Weight Averaging with Multiple Particles for Iterative Magnitude Pruning	Sparsity Theory	`U`	-	Unstructured sparsity / weight pruning method.
`ICLR`	Synergistic Patch Pruning for Vision Transformer: Unifying Intra- & Inter-Layer Patch Importance	ViT/Transformer	`S`	-	Token/patch reduction for transformer acceleration.
`ICLR`	FedP3: Federated Personalized and Privacy-friendly Network Pruning under Model Heterogeneity	Federated/Privacy	`S`	-	Pruning method for federated or distributed settings.
`ICLR`	The Need for Speed: Pruning Transformers with One Recipe	ViT/Transformer	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`ICLR`	SAS: Structured Activation Sparsification	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`CVPR`	OrthCaps: An Orthogonal CapsNet with Sparse Attention Routing and Pruning	Sparsity Theory	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`CVPR`	Zero-TPrune: Zero-Shot Token Pruning through Leveraging of the Attention Graph in Pre-Trained Transformers	ViT/Transformer	`S`	PyTorch[A]	Token/patch reduction for transformer acceleration.
`CVPR`	Finding Lottery Tickets in Vision Models via Data-driven Spectral Foresight Pruning	Sparsity Theory	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`CVPR`	BilevelPruning: Unified Dynamic and Static Channel Pruning for Convolutional Neural Networks	CNN	`S`	-	Input-adaptive or dynamic pruning strategy.
`CVPR`	FedMef: Towards Memory-efficient Federated Dynamic Pruning	Federated/Privacy	`S`	-	Pruning method for federated or distributed settings.
`CVPR`	Resource-Efficient Transformer Pruning for Finetuning of Large Models	ViT/Transformer	`S`	-	Structured pruning for hardware-friendly compression.
`CVPR`	Device-Wise Federated Network Pruning	Federated/Privacy	`S`	-	Pruning method for federated or distributed settings.
`CVPR`	Auto-Train-Once: Controller Network Guided Automatic Network Pruning from Scratch	General	`S`	-	Structured pruning for hardware-friendly compression.
`CVPR`	Jointly Training and Pruning CNNs via Learnable Agent Guidance and Alignment	CNN	`S`	-	Structured pruning for hardware-friendly compression.
`CVPR`	Diversity-aware Channel Pruning for StyleGAN Compression	GAN	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`CVPR`	MADTP: Multimodal Alignment-Guided Dynamic Token Pruning for Accelerating Vision-Language Transformer	VLM/MLLM	`S`	PyTorch[A]	Recent method for pruning or sparsifying large language models.
`AAAI`	Dynamic Feature Pruning and Consolidation for Occluded Person Re-Identification	General	`S`	-	Input-adaptive or dynamic pruning strategy.
`AAAI`	REPrune: Channel Pruning via Kernel Representative Selection	CNN	`S`	-	Structured pruning for hardware-friendly compression.
`AAAI`	Revisiting Gradient Pruning: A Dual Realization for Defending against Gradient Attacks	General	`S`	-	Structured pruning for hardware-friendly compression.
`AAAI`	IRPruneDet: Efficient Infrared Small Target Detection via Wavelet Structure-Regularized Soft Channel Pruning	Detection/Segmentation	`S`	-	Structured pruning for hardware-friendly compression.
`AAAI`	EPSD: Early Pruning with Self-Distillation for Efficient Model Compression	General	`S`	-	Structured pruning for hardware-friendly compression.
`WACV`	Pruning from Scratch via Shared Pruning Module and Nuclear norm-based Regularization	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`WACV`	Towards Better Structured Pruning Saliency by Reorganizing Convolution	GAN	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`WACV`	Torque based Structured Pruning for Deep Neural Network	General	`S`	-	Structured pruning for hardware-friendly compression.
`WACV`	Revisiting Token Pruning for Object Detection and Instance Segmentation	ViT/Transformer	`S`	PyTorch[A]	Token/patch reduction for transformer acceleration.
`WACV`	Token Fusion: Bridging the Gap Between Token Pruning and Token Merging	ViT/Transformer	`S`	-	Token/patch reduction for transformer acceleration.
`WACV`	PATROL: Privacy-Oriented Pruning for Collaborative Inference Against Model Inversion Attacks	Federated/Privacy	`S`	-	Structured pruning for hardware-friendly compression.
`ICLR`	Sheared LLaMA: Accelerating Language Model Pre-training via Structured Pruning	LLM	`S`	PyTorch[A]	Structured pruning + continued pretraining for smaller LLMs.
`ICLR`	Plug-and-Play: An Efficient Post-training Pruning Method for Large Language Models	LLM	`U`	-	RIA + channel permutation for post-training LLM pruning and N:M semi-structured sparsity.
`ACL Findings`	LoRAPrune: Structured Pruning Meets Low-Rank Parameter-Efficient Fine-Tuning	LLM	`S`	PyTorch[A]	LoRA-guided structured pruning with reduced memory cost.
`ECCV`	FastV: An Image is Worth 1/2 Tokens After Layer 2	LVLM	`T`	PyTorch[A]	Plug-and-play visual-token pruning for LMM inference acceleration.

2023

Venue	Title	Area	Type	Code	Notes
`NIPS`	Diff-Pruning: Structural Pruning for Diffusion Models	Diffusion	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`NIPS`	LLM-Pruner: On the Structural Pruning of Large Language Models	LLM	`S`	PyTorch[A]	Recent method for pruning or sparsifying large language models.
`ICCV`	Automatic Network Pruning via Hilbert-Schmidt Independence Criterion Lasso under Information Bottleneck Principle	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`ICCV`	Unified Data-Free Compression: Pruning and Quantization without Fine-Tuning	General	`S`	PyTorch[A]	Combines pruning with quantization/compression.
`ICCV`	Structural Alignment for Network Pruning through Partial Regularization	General	`S`	-	Structured pruning for hardware-friendly compression.
`ICCV`	Differentiable Transportation Pruning	General	`S`	-	Structured pruning for hardware-friendly compression.
`ICCV`	Dynamic Token Pruning in Plain Vision Transformers for Semantic Segmentation	ViT/Transformer	`S`	PyTorch[A]	Token/patch reduction for transformer acceleration.
`ICCV`	Towards Fairness-aware Adversarial Network Pruning	General	`S`	-	Structured pruning for hardware-friendly compression.
`ICCV`	Efficient Joint Optimization of Layer-Adaptive Weight Pruning in Deep Neural Networks	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`CVPR`	DepGraph: Towards Any Structural Pruning	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`CVPR`	X-Pruner: eXplainable Pruning for Vision Transformers	ViT/Transformer	`U/S`	-	-
`CVPR`	Joint Token Pruning and Squeezing Towards More Aggressive Compression of Vision Transformers	ViT/Transformer	`S`	PyTorch[A]	Token/patch reduction for transformer acceleration.
`CVPR`	Global Vision Transformer Pruning with Hessian-Aware Saliency	ViT/Transformer	`S`	-	Structured pruning for hardware-friendly compression.
`CVPR`	CP3: Channel Pruning Plug-in for Point-based Networks	3D/Point Cloud	`S`	-	Structured pruning for hardware-friendly compression.
`CVPR`	Training Debiased Subnetworks With Contrastive Weight Pruning	General	`U`	-	Unstructured sparsity / weight pruning method.
`CVPR`	Pruning Parameterization With Bi-Level Optimization for Efficient Semantic Segmentation on the Edge	Detection/Segmentation	`S`	-	Structured pruning for hardware-friendly compression.
`CVPR`	Structural Alignment for Network Pruning through Partial Regularization	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`ICLR`	JaxPruner: A concise library for sparsity research	Sparsity Theory	`U/S`	PyTorch[A]	-
`ICLR`	OTOv2: Automatic, Generic, User-Friendly	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`ICLR`	How I Learned to Stop Worrying and Love Retraining	General	`U`	PyTorch[A]	Unstructured sparsity / weight pruning method.
`ICLR`	Token Merging: Your ViT But Faster	ViT/Transformer	`U/S`	PyTorch[A]	Token/patch reduction for transformer acceleration.
`ICLR`	Revisiting Pruning at Initialization Through the Lens of Ramanujan Graphs	Sparsity Theory	`U`	PyTorch[A] (soon...)	Unstructured sparsity / weight pruning method.
`ICLR`	Unmasking the Lottery Ticket Hypothesis: What's Encoded in a Winning Ticket's Mask?	Sparsity Theory	`U`	-	Unstructured sparsity / weight pruning method.
`ICLR`	NTK-SAP: Improving neural network pruning by aligning training dynamics	General	`U`	-	Unstructured sparsity / weight pruning method.
`ICLR`	DFPC: Data flow driven pruning of coupled channels without data	CNN	`S`	PyTorch[A]	Data-free or calibration-light pruning method.
`ICLR`	TVSPrune - Pruning Non-discriminative filters via Total Variation separability of intermediate representations without fine tuning	CNN	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`ICLR`	Pruning Deep Neural Networks from a Sparsity Perspective	Sparsity Theory	`U`	PyTorch[A]	Unstructured sparsity / weight pruning method.
`ICLR`	A Unified Framework of Soft Threshold Pruning	General	`U`	PyTorch[A]	Unstructured sparsity / weight pruning method.
`WACV`	Calibrating Deep Neural Networks Using Explicit Regularisation and Dynamic Data Pruning	General	`S`	-	Input-adaptive or dynamic pruning strategy.
`WACV`	Attend Who Is Weak: Pruning-Assisted Medical Image Localization Under Sophisticated and Implicit Imbalances	Medical Imaging	`S`	-	Structured pruning for hardware-friendly compression.
`ICASSP`	WHC: Weighted Hybrid Criterion for Filter Pruning on Convolutional Neural Networks	CNN	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`ICML`	SparseGPT: Massive Language Models Can Be Accurately Pruned in One-Shot	LLM	`U/SS`	PyTorch[A]	First widely used one-shot pruning method for 10B–100B+ parameter GPT-family models.
`arXiv`	Wanda: A Simple and Effective Pruning Approach for Large Language Models	LLM	`U/SS`	PyTorch[A]	Weight magnitude × activation norm; no retraining or weight update.
`arXiv`	LoRAPrune: Structured Pruning Meets Low-Rank Parameter-Efficient Fine-Tuning	LLM	`S`	PyTorch[A]	Early version later appearing in ACL Findings 2024.

2022

Venue	Title	Area	Type	Code	Notes
`CVPR`	Interspace Pruning: Using Adaptive Filter Representations To Improve Training of Sparse CNNs	Sparsity Theory	`U`	-	Unstructured sparsity / weight pruning method.
`CVPR`	Revisiting Random Channel Pruning for Neural Network Compression	CNN	`S`	PyTorch[A] (soon...)	Structured pruning for hardware-friendly compression.
`CVPR`	Fire Together Wire Together: A Dynamic Pruning Approach With Self-Supervised Mask Prediction	General	`S`	PyTorch[A]	Input-adaptive or dynamic pruning strategy.
`CVPR`	When to Prune? A Policy towards Early Structural Pruning	General	`S`	-	Structured pruning for hardware-friendly compression.
`CVPR`	Dreaming to Prune Image Deraining Networks	Low-level Vision	`S`	-	Structured pruning for hardware-friendly compression.
`ICLR`	SOSP: Efficiently Capturing Global Correlations by Second-Order Structured Pruning	General	`S`	-	Structured pruning for hardware-friendly compression.
`ICLR`	Learning Pruning-Friendly Networks via Frank-Wolfe: One-Shot, Any-Sparsity, And No Retraining	Sparsity Theory	`U`	PyTorch[A]	Unstructured sparsity / weight pruning method.
`ICLR`	Revisit Kernel Pruning with Lottery Regulated Grouped Convolutions	Sparsity Theory	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`ICLR`	Dual Lottery Ticket Hypothesis	Sparsity Theory	`U`	PyTorch[A]	Unstructured sparsity / weight pruning method.
`NIPS`	SAViT: Structure-Aware Vision Transformer Pruning via Collaborative Optimization	ViT/Transformer	`S`	PyTorch[A](soon...)	Structured pruning for hardware-friendly compression.
`NIPS`	Structural Pruning via Latency-Saliency Knapsack	Deployment	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`ACCV`	Filter Pruning via Automatic Pruning Rate Search⋆	CNN	`S`	-	Structured pruning for hardware-friendly compression.
`ACCV`	Network Pruning via Feature Shift Minimization	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`ACCV`	Lightweight Alpha Matting Network Using Distillation-Based Channel Pruning	Low-level Vision	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`ACCV`	Adaptive FSP : Adaptive Architecture Search with Filter Shape Pruning	CNN	`S`	-	Structured pruning for hardware-friendly compression.
`ECCV`	Soft Masking for Cost-Constrained Channel Pruning	CNN	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`WACV`	Hessian-Aware Pruning and Optimal Neural Implant	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`WACV`	PPCD-GAN: Progressive Pruning and Class-Aware Distillation for Large-Scale Conditional GANs Compression	GAN	`S`	-	Structured pruning for hardware-friendly compression.
`WACV`	Channel Pruning via Lookahead Search Guided Reinforcement Learning	CNN	`S`	-	Structured pruning for hardware-friendly compression.
`WACV`	EZCrop: Energy-Zoned Channels for Robust Output Pruning	Deployment	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`ICIP`	One-Cycle Pruning: Pruning Convnets With Tight Training Budget	CNN	`U`	-	Unstructured sparsity / weight pruning method.
`ICIP`	RAPID: A Single Stage Pruning Framework	General	`U`	-	Unstructured sparsity / weight pruning method.
`ICIP`	The Rise of the Lottery Heroes: Why Zero-Shot Pruning is Hard	Sparsity Theory	`U`	-	Unstructured sparsity / weight pruning method.
`ICIP`	Truncated Lottery Ticket for Deep Pruning	Sparsity Theory	`U`	-	Unstructured sparsity / weight pruning method.
`ICIP`	Which Metrics For Network Pruning: Final Accuracy? or Accuracy Drop?	General	`S/U`	-	-
`ISMSI`	Structured Pruning with Automatic Pruning Rate Derivation for Image Processing Neural Networks	Low-level Vision	`S`	-	Structured pruning for hardware-friendly compression.

2021

Venue	Title	Area	Type	Code	Notes
`ICLR`	Neural Pruning via Growing Regularization	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`ICLR`	Network Pruning That Matters: A Case Study on Retraining Variants	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`ICLR`	Layer-adaptive Sparsity for the Magnitude-based Pruning	Sparsity Theory	`U`	PyTorch[A]	Unstructured sparsity / weight pruning method.
`NIPS`	Only Train Once: A One-Shot Neural Network Training And Pruning Framework	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`CVPR`	NPAS: A Compiler-Aware Framework of Unified Network Pruning and Architecture Search for Beyond Real-Time Mobile Acceleration	Deployment	`S`	-	Structured pruning for hardware-friendly compression.
`CVPR`	Network Pruning via Performance Maximization	General	`S`	-	Structured pruning for hardware-friendly compression.
`CVPR`	Convolutional Neural Network Pruning With Structural Redundancy Reduction*	CNN	`S`	-	Structured pruning for hardware-friendly compression.
`CVPR`	Manifold Regularized Dynamic Network Pruning	General	`S`	PyTorch[A]	Input-adaptive or dynamic pruning strategy.
`CVPR`	Joint-DetNAS: Upgrade Your Detector With NAS, Pruning and Dynamic Distillation	General	`S`	-	Input-adaptive or dynamic pruning strategy.
`ICCV`	ResRep: Lossless CNN Pruning via Decoupling Remembering and Forgetting	CNN	`S`	-	Structured pruning for hardware-friendly compression.
`ICCV`	Achieving On-Mobile Real-Time Super-Resolution With Neural Architecture and Pruning Search	Low-level Vision	`S`	-	Structured pruning for hardware-friendly compression.
`ICCV`	GDP: Stabilized Neural Network Pruning via Gates With Differentiable Polarization*	General	`S`	-	Structured pruning for hardware-friendly compression.
`WACV`	Holistic Filter Pruning for Efficient Deep Neural Networks	CNN	`S`	-	Structured pruning for hardware-friendly compression.
`ICML`	Accelerate CNNs from Three Dimensions: A Comprehensive Pruning Framework	CNN	`S`	-	Structured pruning for hardware-friendly compression.
`ICML`	Group Fisher Pruning for Practical Network Compression	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.

2020

Venue	Title	Area	Type	Code	Notes
`CVPR`	HRank: Filter Pruning using High-Rank Feature Map	CNN	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`CVPR`	Towards efficient model compression via learned global ranking	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`CVPR`	Learning Filter Pruning Criteria for Deep Convolutional Neural Networks Acceleration	CNN	`S`	-	Structured pruning for hardware-friendly compression.
`CVPR`	Group Sparsity: The Hinge Between Filter Pruning and Decomposition for Network Compression	Sparsity Theory	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`CVPR`	APQ: Joint Search for Network Architecture, Pruning and Quantization Policy	General	`S`	PyTorch[A]	Combines pruning with quantization/compression.
`ICLR`	Budgeted Training: Rethinking Deep Neural Network Training Under Resource Constraints	General	`U`	-	Unstructured sparsity / weight pruning method.
`MLSys`	Shrinkbench: What is the State of Neural Network Pruning?	General	-	PyTorch[A]	-
`BMBS`	Similarity Based Filter Pruning for Efficient Super-Resolution Models	Low-level Vision	`S`	-	Structured pruning for hardware-friendly compression.

2019

Venue	Title	Area	Type	Code	Notes
`CVPR`	Filter Pruning via Geometric Median for Deep Convolutional Neural Networks Acceleration	CNN	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`CVPR`	Variational Convolutional Neural Network Pruning	CNN	`S`	-	Structured pruning for hardware-friendly compression.
`CVPR`	Towards Optimal Structured CNN Pruning via Generative Adversarial Learning	CNN	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`CVPR`	Partial Order Pruning: For Best Speed/Accuracy Trade-Off in Neural Architecture Search	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`CVPR`	Importance Estimation for Neural Network Pruning	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`ICLR`	The Lottery Ticket Hypothesis: Finding Sparse, Trainable Neural Networks	Sparsity Theory	`U`	PyTorch[A]	Unstructured sparsity / weight pruning method.
`ICLR`	SNIP: Single-shot Network Pruning based on Connection Sensitivity	ViT/Transformer	`U`	Tensorflow[A]	Unstructured sparsity / weight pruning method.
`ICCV`	MetaPruning: Meta-Learning for Automatic Neural Network Channel Pruning	CNN	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`ICCV`	Accelerate CNN via Recursive Bayesian Pruning	CNN	`S`	-	Structured pruning for hardware-friendly compression.

2018

Venue	Title	Area	Type	Code	Notes
`CVPR`	PackNet: Adding Multiple Tasks to a Single Network by Iterative Pruning	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`CVPR`	NISP: Pruning Networks Using Neuron Importance Score Propagation	General	`S`	-	Structured pruning for hardware-friendly compression.
`ICIP`	Online Filter Clustering and Pruning for Efficient Convnets	CNN	`S`	-	Structured pruning for hardware-friendly compression.
`IJCAI`	Soft Filter Pruning for Accelerating Deep Convolutional Neural Networks	CNN	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.

2017

Venue	Title	Area	Type	Code	Notes
`CVPR`	Designing Energy-Efficient Convolutional Neural Networks Using Energy-Aware Pruning	Deployment	`S`	-	Structured pruning for hardware-friendly compression.
`ICLR`	Pruning Filters for Efficient ConvNets	CNN	`S`	PyTorch[O]	Structured pruning for hardware-friendly compression.
`ICCV`	Channel Pruning for Accelerating Very Deep Neural Networks	CNN	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`ICCV`	ThiNet: A Filter Level Pruning Method for Deep Neural Network Compression	CNN	`S`	Caffe[A]	Structured pruning for hardware-friendly compression.
`ICCV`	Learning Efficient Convolutional Networks Through Network Slimming	CNN	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.

Journal Publications

2026

Journal	Title	Area	Type	Code	Notes
-	-	General	-	-	No peer-reviewed journal pruning papers were confidently added for 2026 yet; most 2026 additions are conference or preprint entries above.

2025

Journal	Title	Area	Type	Code	Notes
Neural Networks	OOPS: Outlier-aware and Quadratic Programming Based Structured Pruning for Large Language Models	LLM	`S`	-	Structured LLM pruning with outlier-aware quadratic-programming formulation.

2024

Journal	Title	Area	Type	Code	Notes
Neural Networks	Efficient tensor decomposition-based filter pruning	CNN	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
IEEE Trans. NNLS	Enhanced Network Compression Through Tensor Decompositions and Pruning	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
IEEE Transactions on Artificial Intelligence	Distilled Gradual Pruning with Pruned Fine-tuning	General	`U`	PyTorch[A]	Unstructured sparsity / weight pruning method.

2023

Journal	Title	Area	Type	Code	Notes
IEEE Trans Circuits Syst Video Technol	DCFP: Distribution Calibrated Filter Pruning for Lightweight and Accurate Long-tail Semantic Segmentation	Detection/Segmentation	`S`	-	Structured pruning for hardware-friendly compression.
IEEE Internet Things J.	SNPF: Sensitiveness Based Network Pruning Framework for Efficient Edge Computing	Deployment	`S`	-	Structured pruning for hardware-friendly compression.
IEEE Trans. NNLS	Manipulating Identical Filter Redundancy for Efficient Pruning on Deep and Complicated CNN	CNN	`S`	-	Structured pruning for hardware-friendly compression.
IEEE Trans. NNLS	Block-Wise Partner Learning for Model Compression	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
IEEE Trans. NNLS	Hierarchical Threshold Pruning Based on Uniform Response Criterion	General	`S`	-	Structured pruning for hardware-friendly compression.
IEEE Trans. NNLS	CATRO: Channel Pruning via Class-Aware Trace Ratio Optimization	CNN	`S`	-	Structured pruning for hardware-friendly compression.
IEEE Trans. NNLS	Adaptive Filter Pruning via Sensitivity Feedback	ViT/Transformer	`S`	-	Structured pruning for hardware-friendly compression.
Neurocomputing	Filter pruning with uniqueness mechanism in the frequency domain for efficient neural networks	CNN	`S`	-	Structured pruning for hardware-friendly compression.
IEEE Trans. PAMI	Compact Neural Network via Stacking Hybrid Units	General	`S`	-	Structured pruning for hardware-friendly compression.
IEEE Trans. PAMI	Performance-aware Approximation of Global Channel Pruning for Multitask CNNs	CNN	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
IEEE Trans. PAMI	Adaptive Search-and-Training for Robust and Efficient Network Pruning	General	`S`	-	Structured pruning for hardware-friendly compression.
Image Vis. Comput.	Loss-aware automatic selection of structured pruning criteria for deep neural network acceleration	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
Comput. Vis. Image Underst.	Feature independent Filter Pruning by Successive Layers analysis	CNN	`S`	-	Structured pruning for hardware-friendly compression.
IEEE Access	Differentiable Neural Architecture, Mixed Precision and Accelerator Co-Search	Deployment	`S`	-	Structured pruning for hardware-friendly compression.

2022

Journal	Title	Area	Type	Code	Notes
IEEE Trans. Image Process.	Efficient Layer Compression Without Pruning	General	`S`	-	Structured pruning for hardware-friendly compression.
IEEE Trans. PAMI	Learning to Explore Distillability and Sparsability: A Joint Framework for Model Compression	General	`S`	-	Structured pruning for hardware-friendly compression.
IEEE Trans. PAMI	1xN Pattern for Pruning Convolutional Neural Networks	CNN	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
IEEE Trans. NNLS	Filter Pruning by Switching to Neighboring CNNs With Good Attribute	CNN	`S`	-	Structured pruning for hardware-friendly compression.
IEEE Trans. NNLS	Model Pruning Enables Efficient Federated Learning on Edge Devices	Federated/Privacy	`S`	-	Pruning method for federated or distributed settings.
IEEE Trans. NNLS	DAIS: Automatic Channel Pruning via Differentiable Annealing Indicator Search	CNN	`S`	-	Structured pruning for hardware-friendly compression.
IEEE Trans. NNLS	Network Pruning Using Adaptive Exemplar Filters	CNN	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
IEEE Trans. NNLS	Carrying Out CNN Channel Pruning in a White Box	CNN	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
IEEE Trans. NNLS	Pruning Networks With Cross-Layer Ranking & k-Reciprocal Nearest Filters	CNN	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
IEEE Trans. NNLS	Filter Sketch for Network Pruning	CNN	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
Neurocomputing	FPFS: Filter-level pruning via distance weight measuring filter similarity	CNN	`S`	-	Structured pruning for hardware-friendly compression.
Neurocomputing	RUFP: Reinitializing unimportant filters for soft pruning	CNN	`S`	-	Structured pruning for hardware-friendly compression.
Neural Netw	HRel: Filter pruning based on High Relevance between activation maps and class labels	CNN	`S`	PyTorch[A]*	Structured pruning for hardware-friendly compression.
Comput. Intell. Neurosci.	Differentiable Network Pruning via Polarization of Probabilistic Channelwise Soft Masks	CNN	`S`	-	Structured pruning for hardware-friendly compression.
J. Syst. Archit.	Optimizing deep neural networks on intelligent edge accelerators via flexible-rate filter pruning	Deployment	`S`	-	Structured pruning for hardware-friendly compression.
Appl. Sci.	Magnitude and Similarity Based Variable Rate Filter Pruning for Efficient Convolution Neural Networks	Sparsity Theory	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
Sensors	Filter Pruning via Measuring Feature Map Information	CNN	`S`	-	Structured pruning for hardware-friendly compression.
IEEE Access	Automated Filter Pruning Based on High-Dimensional Bayesian Optimization	CNN	`S`	-	Structured pruning for hardware-friendly compression.
IEEE Signal Process. Lett.	A Low-Complexity Modified ThiNet Algorithm for Pruning Convolutional Neural Networks	CNN	`S`	-	Structured pruning for hardware-friendly compression.

2021

Journal	Title	Area	Type	Code	Notes
IEEE Trans. PAMI	Discrimination-Aware Network Pruning for Deep Model Compression	General	`S`	PyTorch[A]~	Structured pruning for hardware-friendly compression.

2020

Journal	Title	Area	Type	Code	Notes
IEEE Trans. NNLS	EDP: An Efficient Decomposition and Pruning Scheme for Convolutional Neural Network Compression	CNN	`S`	-	Structured pruning for hardware-friendly compression.
IEEE Access	Filter Pruning Without Damaging Networks Capacity	CNN	`S`	-	Structured pruning for hardware-friendly compression.
Electronics	Pruning Convolutional Neural Networks with an Attention Mechanism for Remote Sensing Image Classification	CNN	`S`	-	Structured pruning for hardware-friendly compression.

Survey Articles

Year	Venue	Title	Area	Type	Code	Notes
`2026`	`arXiv`	Towards Efficient Multimodal Large Language Models: A Survey on Token Compression	LLM	-	-	Survey / overview resource.
`2024`	`IEEE TPAMI`	A Survey on Deep Neural Network Pruning: Taxonomy, Comparison, Analysis, and Recommendations	Survey	-	-	Survey / overview resource.
`2024`	`arXiv`	A Survey on Model Compression for Large Language Models	LLM	-	-	Survey / overview resource.
`2023`	`Artif. Intell. Rev.`	Deep neural network pruning method based on sensitive layers and reinforcement learning	General	-	-	Survey / overview resource.
`2023`	`arVix`	A Survey on Deep Neural Network Pruning: Taxonomy, Comparison, Analysis, and Recommendations	Survey	-	-	Survey / overview resource.
`2023`	`arVix`	Structured Pruning for Deep Convolutional Neural Networks: A survey	CNN	-	-	Survey / overview resource.
`2022`	`Electronics`	A Survey on Efficient Convolutional Neural Networks and Hardware Acceleration	Deployment	-	-	Survey / overview resource.
`2022`	`I-SMAC`	A Survey on Filter Pruning Techniques for Optimization of Deep Neural Networks	CNN	-	-	Survey / overview resource.
`2021`	`JMLR`	Sparsity in Deep Learning: Pruning and growth for efficient inference and training in neural networks	Sparsity Theory	-	-	Survey / overview resource.
`2021`	`Neurocomputing`	Pruning and quantization for deep neural network acceleration: A survey	Survey	-	-	Survey / overview resource.
`2020`	`IEEE Access`	Methods for Pruning Deep Neural Networks	General	-	-	Survey / overview resource.

Other Publications

Venue	Title	Area	Type	Code	Notes
`2023`	`arVix`	Why is the State of Neural Network Pruning so Confusing? On the Fairness, Comparison Setup, and Trainability in Network Pruning	General	-	PyTorch[A]
`2023`	`arVix`	Ten Lessons We Have Learned in the New "Sparseland": A Short Handbook for Sparse Neural Network Researchers	Sparsity Theory	-	-
`2022`	`ICML`	Tutorial -- Sparsity in Deep Learning: Pruning and growth for efficient inference and training	Sparsity Theory	-	-

Pruning Software and Toolbox

Year	Title	Area	Type	Code	Notes
`2026`	LLM Compressor / vLLM compression tools	LLM	`U/S/Q/SS`	Python[A]	Recent method for pruning or sparsifying large language models.
`2025`	GETA	General	`S/Q`	PyTorch[A]	-
`2025`	PACT	General	`T`	PyTorch[A]	-
`2025`	DivPrune	General	`T`	PyTorch[A]	-
`2025`	Token Cropr	ViT/Transformer	`T`	PyTorch[A]	Token/patch reduction for transformer acceleration.
`2025`	2SSP	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`2024`	LLM-Shearing / Sheared LLaMA	LLM	`S`	PyTorch[A]	Recent method for pruning or sparsifying large language models.
`2024`	LoRAPrune	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`2023`	SparseGPT	LLM	`U/SS`	PyTorch[A]	Recent method for pruning or sparsifying large language models.
`2023`	Wanda	General	`U/SS`	PyTorch[A]	-
`2023`	UPop: Unified and Progressive Pruning for Compressing Vision-Language Transformers	VLM/MLLM	`S`	PyTorch[A]	Recent method for pruning or sparsifying large language models.
`2023`	DepGraph: Towards Any Structural Pruning	General	`S`	PyTorch[A]	Structured pruning for hardware-friendly compression.
`2023`	Torch-Pruning	Toolbox	`S`	PyTorch[A]	Software resource for pruning/compression workflows.
`2023`	JaxPruner: JaxPruner: A concise library for sparsity research	Sparsity Theory	`U/S`	PyTorch[A]	-
`2022`	FasterAI: Prune and Distill your models with FastAI and PyTorch	General	`U`	PyTorch[A]	Unstructured sparsity / weight pruning method.
`2022`	Simplify: A Python library for optimizing pruned neural networks	Toolbox	-	PyTorch[A]	Software resource for pruning/compression workflows.
`2021`	PyTorchViz [A small package to create visualizations of PyTorch execution graphs]	General	-	PyTorch[A]	-
`2020`	What is the State of Neural Network Pruning?	General	`S/U`	PyTorch[A]	-
`2019`	Official PyTorch Pruning Tool	Toolbox	`S/U`	PyTorch[A]	Software resource for pruning/compression workflows.

Citation

If this repository helps your research, please cite it as:

@misc{awesome_pruning,
  title  = {Awesome Pruning},
  author = {Deepak Ghimire and others},
  year   = {2026},
  note   = {Curated list of pruning papers, surveys, software, and code across CNNs, ViTs, LLMs, VLMs/MLLMs, diffusion, and deployment-aware compression}
}

awesome

pruning

pytorch

structured-pruning

unstructured-pruning

Similar repos

hrcheng1066/awesome-pruning

309

liyunqianggyn/Awesome-LLMs-Pruning

Awesome LLM pruning papers all-in-one repository with integrating all useful resources and…

180

horseee/Awesome-Efficient-LLM

A curated list for Efficient Large Language Models

2,043

LumosJiang/Awesome-On-Device-LLMs

[Artificial Intelligence Review 2026] Awesome list and resources for on-device LLMs: model…

daixiangzi/Awesome-Token-Compress

A paper list of some recent works about Token Compress for Vit and VLM

957

cedrickchee/awesome-ml-model-compression

Awesome machine learning model compression research papers, quantization, tools, and learning…

546

AIoT-MLSys-Lab/Efficient-LLMs-Survey

[TMLR 2024] Efficient Large Language Models: A Survey

1,257

VainF/Torch-Pruning

[CVPR 2023] DepGraph: Towards Any Structural Pruning; LLMs, Vision Foundation Models, etc.

3,357

Contributors

140 commits

5 commits

4 commits

1 commits

ghimiredhikura/Awasome-Pruning

README

Awesome Pruning

Contents

Legend

Quick Recent Trends

Conference Publications

2026

2025

2024

2023

2022

2021

2020

2019

2018

2017

Journal Publications

2026

2025

2024

2023

2022

2021

2020

Survey Articles

Other Publications

Pruning Software and Toolbox

Citation

Contributors

Similar repos

ghimiredhikura/Awasome-Pruning

README

Awesome Pruning

Contents

Legend

Quick Recent Trends

Conference Publications

2026

2025

2024

2023

2022

2021

2020

2019

2018

2017

Journal Publications

2026

2025

2024

2023

2022

2021

2020

Survey Articles

Other Publications

Pruning Software and Toolbox

Citation

Similar repos

Contributors