[NeurIPS 2023] Structural Pruning for Diffusion Models

Awesome Pruning. ✅ Curated Resources for Neural Network Pruning.

The framework to prune LLMs to any size and any config.

[AAAI 2024] Fluctuation-based Adaptive Structured Pruning for Large Language Models

OTOv1-v3, NeurIPS, ICLR, TMLR, DNN Training, Compression, Structured Pruning, Erasing Operators, CNN, Diffusion, LLM

Structured pruning, knowledge distillation, and fairness analysis for Large Language Models.

Code for CHIP: CHannel Independence-based Pruning for Compact Neural Networks (NeruIPS 2021).

💍 Efficient tensor decomposition-based filter pruning

This repository is the official implementation of the paper Pruning via Iterative Ranking of Sensitivity Statistics and implements novel pruning / compression algorithms for deep learning / neural networks. Amongst others it implements structured pruning before training, its actual parameter shrinking and unstructured before/during training.

We have implemented a framework that supports developers to structured prune neural networks of Tensorflow Models

Knowledge distillation from Ensembles of Iterative pruning (BMVC 2020)

2SSP: A Two-Stage Framework for Structured Pruning of LLMs

Towards Meta-Pruning via Optimal Transport, ICLR 2024 (Spotlight)

🌠 Enhanced Network Compression Through Tensor Decompositions and Pruning

Code Implementation for "NASH: A Simple Unified Framework of Structured Pruning for Accelerating Encoder-Decoder Language Models" (EMNLP 2023)

Loss-aware automatic selection of structured pruning criteria for deep neural network acceleration. ✅ Published in Image and Vision Computing, 2023.

[ICLR 2025] Probe Pruning: Accelerating LLMs through Dynamic Pruning via Model-Probing

Code repository for paper "Efficient Structured Pruning and Architecture Searching for Group Convolution" https://arxiv.org/abs/1811.09341

LoMACS-SVDNet: PyTorch model for MIMO SVD (no QR/SVD/EVD), orthogonality via NOR, FFT gating, projected attention, structured pruning. Score 63 — 4th (Third Prize).

Research-ready and production-friendly neural network pruning for PyTorch—transparent methods, reproducible baselines, and deployment metrics to compress models for real-world use.