Parallel-SGD

Parallel implementation of Stochastic Gradient Descent using SciKit-Learn library in Python.

Please see Benchmark notebook for characterization of 5 techniques showing their speed-up and accuracy. PSGD5, based on 'Hogwild!', performs the best with no loss of accuracy.

This project is a work-in-progress for a feature request on SciKit-Learn: #1174

Team Members

• Angad Gill
• Sharath Rao