Skip to content

YuekaiXuEric/Rotation-Pruning

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

13 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Roation + Pruning

Setup

Below is an example command for pruning LLaMA2-7B with Rotation and Sparsegpt, to achieve unstructured 50% sparsity. Step 1: In the QuaRot Repo download the fast_hadamard_transform QuaRot.

python setup.py install

Step 2: Back to original repo

python main.py
  --model meta-llama/Llama-2-7b-hf
  --prune_method sparsegpt
  --sparsity_ratio 0.5
  --sparsity_type unstructured
  --rotate
  --reorder

We provide a quick overview of the arguments:

  • --rotate: Adding roation to Pruning
  • --reorder: Adding permutation to Pruning

Llama 2 7b model path:

ln -s /data/datasets/llama2 ./meta-llama

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages