ZhengaoLi/DISP-LLM-Dimension-Independent-Structural-Pruning

An implementation of the DISP-LLM method from the NeurIPS 2024 paper: Dimension-Independent Structural Pruning for Large Language Models.

32
/ 100
Emerging

This project helps machine learning engineers and researchers reduce the computational cost and memory footprint of large language models like LLaMA 7B/13B and LLaMA-2 7B/13B. By taking a pre-trained LLaMA model and applying a technique called structural pruning, it outputs a smaller, more efficient version of the model that maintains its performance. This is ideal for those deploying LLMs in resource-constrained environments or seeking to optimize existing models.

No commits in the last 6 months.

Use this if you need to make LLaMA 7B/13B or LLaMA-2 7B/13B models smaller and faster without significant performance loss.

Not ideal if you are working with other large language models outside the LLaMA family or if your primary goal is to improve model accuracy rather than efficiency.

large-language-models model-optimization deep-learning AI-efficiency
Stale 6m No Package No Dependents
Maintenance 2 / 25
Adoption 7 / 25
Maturity 16 / 25
Community 7 / 25

How are scores calculated?

Stars

25

Forks

2

Language

Python

License

Apache-2.0

Last pushed

Aug 06, 2025

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/transformers/ZhengaoLi/DISP-LLM-Dimension-Independent-Structural-Pruning"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.