by Eric Madariaga | January 18, 2017

Introducing the ODBC Reader for TensorFlow

Machine learning enables organizations to leverage today's vast computational power to recognize patterns in data and make meaningful predictions. To do this, machine learning platforms have to crunch enormous amounts of data. From financial fraud detection to enabling self-driving car technologies, data is essential to pattern recognition and discovery.

Unfortunately, meaningful organizational data is often siloed into any number of disparate systems. From CRM, ERP, and SaaS applications to various RDBMS or NoSQL database sources, every data interface is different. Organizations interested in applying machine learning techniques are forced to write data-source-specific custom code or live with static dumps to populate their models.

Our goal in this project is to help streamline the way that organizations connect machine learning with real-time data, regardless of location. By doing so we hope to facilitate the growth of ML integration and ultimately make it more accessible to businesses of all sizes.

The TensorFlow ODBC Reader

The TensorFlow project from Google is one of the most popular open-source libraries for machine learning intelligence. Our ODBC Reader for TensorFlow standardizes the way that users populate TensorFlow with data through ODBC, the most commonly used data interface.

The ODBC Reader for TensorFlow allows users to connect the Google TensorFlow machine learning project with any SQL-92 compliant ODBC Driver. The ODBC Reader is driver-agnostic and can retrieve data from any ODBC Driver source. This includes our CData ODBC Drivers, as well as any other open source or standards-compliant third-party ODBC Driver.

Extending TensorFlow with ODBC connectivity provides developers with an easy way to connect ML with live data from disparate sources, enabling ML applications to deliver meaningful insights in real-time.

Getting Started

The ODBC Reader 'tensorflow-odbc' and full source code are publicly available on GitHub.

The ODBC Reader for TensorFlow is an open-source project released under the GPL 3.0 license. For information about commercial licensing, please contact us.