# On Differentially Private Subspace Estimation in a Distribution-Free Setting
A Python Implementation.

Instructions
===

Download the FriendlyCore python code from https://media.icml.cc/Conferences/ICML2022/supplementary/tsfadia22a-supp.zip and place the friendly_core folder inside this folder.

To run the subspace demo, run 
`python subspace_demo.py`


In order to run the averaging experiments:
Varying d: `python test_avg_d_new.py`
Varying k: `python test_avg_k_new.py`
Varying gap: `python test_avg_g_new.py`
