Skip to content

HieuNT91/attention_pruning

Repository files navigation

Official implementation of "Structured Pruning for Diverse Best-of-$N$ Reasoning Optimization" (ACL'25 Short Findings)


teaser

Installation

Requirements:

python 3.12.9
pip install -e .

Inference

To infer on MATH500 and GSM8K, please run those two scripts

./attention_pruning_experiments/scripts/get_accuracy_data_gsm8k.sh
./attention_pruning_experiments/scripts/get_accuracy_data.sh

To reproduce our tables (coming soon, we will release our wandb experiments):

python3 attention_pruning_experiments/draw_tables.py

Availability

Acknowledgement

A substantial portion of the source code has been borrowed from the following repositories:

Contact

If you have any problems, please open an issue in this repository or send an email to hilljun.2000@gmail.com.

About

Official implementation of "Structured Pruning for Diverse Best-of-N Reasoning Optimization" (ACL'25 - Short Findings)

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors