coremltools

Statsmodels converter (coremltools fork)

Authors: Raghavprasanna Rajagopalan, Marshall White, and Ethan March.

This is an extension of Apple's coremltools Python package which adds support for conversion to Core ML models from statsmodels, a Python package for econometrics and statistical modeling.

The added converter provides support for Ordinary Least Square Regression and Logit Regression models.

This project was completed at HuskyHacks3 at Northeastern University.

Built With

python

Created by

I worked on the model specific conversion functionality with a focus on the linear regression models. I also provided the team with background on the different statistical modeling methods we were writing converters for.

Ethan March
I worked on the general architecture of our extension to the coremltools package. Specifically, I worked on dispatching the statsmodels ML model object to the appropriate model-specific converter built by team members. I worked on the high level aspects of accepting inputs from the statsmodel and getting the CoreML object as a result of the conversion. (Did a lot of reading documentation, and trying to emulate other converters in the package, trying to filter out functionalities that we needed).

Raghavprasanna Rajagopalan
masterswan

Updates

masterswan started this project — Oct 01, 2017 08:29 AM EDT

Leave feedback in the comments!

Log in or sign up for Devpost to join the conversation.