Libra

Libra is a deep learning API that allows users to use machine learning in their workflows in fluent one-liners. It is built so that users with no knowledge in preprocessing, modeling, or tuning can build high-performance models with ease without worrying about the details of implementation. End-to-end pipelines containing hundreds of processes are automatically run for the user. The developer only has to consider what they want to accomplish from the task and the location of their initial dataset.

  • Beginner Friendly. Libra is an API designed to be used by developers with no deep learning experience whatsoever. It is built so that users with no knowledge in preprocessing, modeling, or tuning can build high-performance models with ease without worrying about the details of implementation.
  • Quick Integration. With the recent rise of machine learning on the cloud, the developer community has failed to make easy-to-use platforms that exist locally and integrate directly into workflows. Libra allows users to develop models directly in programs with hundreds of API endpoints without having to worry about the transition to the cloud.
  • Automation. End-to-end pipelines containing hundreds of processes are automatically run for the user. The developer only has to consider what they want to accomplish from the task and the location of their initial dataset.
  • Easy Extensibility. Queries are split into standalone modules. Under the dev-pipeline module you can pipeline both different and new modules and integrate them into the workflow directly. This allows newly developed features to be easily tested before integrating them into the main program.

Tech Stack

  • Python

About the Engineer

Palash Shah is studying Computer Science at University of Virginia. He’s currently interested in deep reinforcement learning and building end-to-end API’s.

Connect with Palash via LinkedIn or email to learn how to contribute!

 

company logo
Palash Shah