SpecRaGE

This is the official PyTorch implementation of SpecRaGE from the paper "Generalizable and Robust Spectral Method for Multi-view Representation Learning".

For any question, feel free to approach me in this mail - [email protected]

Installation

To run the project, clone this repo and then create a conda environment via:

conda env create -f environment.yml

Subsequently, activate this environment:

conda activate specrage

Clustering

To run a clustering experiment for the BDGP dataset, for example, cd to the src directory and run:

python3 cluster.py bdgp

The output should look like the following:

Classification

To run a classification experiment, cd to the src directory and run:

python3 classify.py bdgp

In this case, the output should look like the following:

Visualization of the unified representation for the BDGP dataset