Streaming Inference for Infinite Non-Stationary Clustering

Authors: Rylan Schaeffer, Gabrielle Kaili-May Liu, Yilun Du, Ila Rani Fiete

---

This code corresponds to our .

Setup

After cloning the repository, create a virtual environment for Python 3:

python3 -m venv dcrp_venv

Then activate the virtual environment:

source dcrp_venv/bin/activate

Ensure pip is up to date:

pip install --upgrade pip

Then install the required packages:

pip install -r requirements.txt

We did not test Python2, but Python2 may work.

Running

Each experiment (e.g. 00_prior) is in its own directory. Each experiment directory should contain a run_one.py file for running a single configuration. If you want to launch sweeps, we run sweeps via Weights and Biases that we configure via .yaml files. After completing your runs, each experiment directory should also contain an analyze_sweep.py file that reads the results from Weights and Biases and generates the plots in a plots subdirectory (e.g. 00_prior/plots).

Contact

Questions? Comments? Interested in collaborating? Open an issue or email Rylan Schaeffer (rylanschaeffer@gmail.com) and cc Ila Fiete (fiete@mit.edu).