Top AI Repos — open-source AI, indexed and scored
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Top AI Repos tracks AI repositories on GitHub and answers two different questions about each one: is it moving right now, and would you bet a product on it.
Awesome list for LLM pruning.
| Date | Stars |
|---|---|
| 2026-07-31 | 297 |
| 2026-08-06 | 297 |
Today
— stars today
This week
— stars this week
This month
— stars this month
Momentum
0.0
growth rate 0.00%/day
# Awesome-LLM-Prune
<div align='center'>
<img src=https://cdn.rawgit.com/sindresorhus/awesome/d7305f38d29fed78fa85652e3a63e154dd8e8829/media/badge.svg >
<img src=https://img.shields.io/github/stars/pprp/Awesome-LLM-Prune.svg?style=social >
<img src=https://img.shields.io/github/watchers/pprp/Awesome-LLM-Prune.svg?style=social >
<img src=https://img.shields.io/badge/Release-v0.1-brightgreen.svg >
<img src=https://img.shields.io/badge/License-GPLv3.0-turquoise.svg >
</div>
This repository is dedicated to the pruning of large language models (LLMs). It aims to serve as a comprehensive resource for researchers and practitioners interested in the efficient reduction of model size while maintaining or enhancing performance.
We encourage the community to engage with this resource; please leave comments on issues related to papers you’re interested in or corrections where misunderstandings may occur. For further inquiries or to contribute to this project, feel free to submit a pull request or open an issue.
| Taxonomy | Unstructured | Structured | Semi-Structured | Benchmark |
| -------- | ------------------------------------------------------------ | ----------------------------------------------------------- | ------------------------------------------------------------ | ------------------------------------------------------------ |
| Label | <img src=https://img.shields.io/badge/unstructured-turquoise.svg > | <img src=https://img.shields.io/badge/structured-blue.svg > | <img src=https://img.shields.io/badge/semi_structured-brightgreen.svg > | <img src=https://img.shields.io/badge/benchmark-purple.svg > |
- SparseGPT: Massive Language Models Can be Accurately Pruned in One-shot
- Label: <img src=https://img.shields.io/badge/unstructured-turquoise.svg ><img src=https://img.shields.io/badge/semi_structured-brightgreen.svg >
- Author: Elias Frantar, Dan Alistarh
- Link: https://arxiv.org/pdf/2301.00774.pdf
- Code: https://github.com/IST-DASLab/sparsegpt
- Pub: ICML 2023
- Summary: First to prune GPT with at least 50% sparsity without any training. SparseGPT is entirely local, which only focus on weight updates without any global gradient information.
- 摘要:首次在没有任何训练的情况下，以至少50%的稀疏度修剪GPT。SparseGPT完全是局部的，它只关注权重更新，没有任何全局梯度信息。
- Wanda: A Simple and Effective Pruning Approach For Large Language Models
- Label: <img src=https://img.shields.io/badge/unstructured-turquoise.svg ><img src=https://img.shields.io/badge/structured-blue.svg >
- Author: Mingjie Sun, Zhuang Liu, Anna Bair, etc.
- Link: https://arxiv.org/pdf/2306.11695.pdf
- Code: https://github.com/locuslab/wanda
- Pub: ICML23 workshop
- Summary: Wanda simplify the SparseGPT with approximation thus just rely on weight and activation to compute the pruning metric. Wanda can be seen as a simplified version of SparseGPT, as it simplify the Hessian approximation, reducing computation greatly.
- 摘要：Wanda通过近似简化了SparseGPT，因此只依赖权重和激活来计算修剪指标。Wanda可以被看作是SparseGPT的简化版本，因为它简化了Hessian近似，大大减少了计算量。
- Pruner-Zero: Evolving Symbolic Pruning Metric
- Label: <img src=https://img.shields.io/badge/unstructured-turquoise.svg ><img src=https://img.shields.io/badge/structured-blue.svg >
- Author: Peijie Dong, Lujun Li, Zhenheng Tang, Xiang Liu, Xinglin Pan, Qiang Wang, Xiaowen Chu
- Link: [arxiv.org/pdf/2406.02924v1](https://arxiv.org/pdf/2406.02924v1)
- Code: [pprp/Pruner-Zero: Evolving Symbolic Pruning Metric from scratch (github.com)](https://github.com/pprp/Pruner-Zero)
- Pub: ICML24
- Summary: Pruner-Zero formulates the pruning metric as a symbolic discovery problem. They develop an automatic framework for searching symbolic pruning metrics using genetic programming. They model the pruning metric as treExcerpt of 73,263 characters
Read on GitHubWould you bet a product on this? Bounded 0–100 and slow moving.
matched fp:b4f55a128d71a912, llm:Repository description: 'Awesome list for LLM pruning' — curated collection about pruning large language models (LLM pruning).
matched fp:b4f55a128d71a912, llm:Repository description: 'Awesome list for LLM pruning' — curated collection about pruning large language models (LLM pruning).
matched fp:b4f55a128d71a912, llm:Repository description: 'Awesome list for LLM pruning' — curated collection about pruning large language models (LLM pruning).