Videos: Building computer vision models with Python and SAS Deep Learning


This series of videos spotlights a very powerful API that lets you use Python while also having access to the power of SAS Deep Learning. Our team of data scientists have been busy enhancing the SAS Deep Learning with Python (DLPy) API with new computer vision models. Each video provides an introductory lesson to the type of model as well as a demo of the model in action. Even if you are a newbie to computer vision, you’ll be able to follow along.

The capabilities spotlighted in this series focus on computer vision with support for popular models such as Functional API, U-NET, Faster R-CNN, MobileNet and Shufflenet. We even delve into multi-class learning for the classification of more than one label and show both training and inference using event stream processing.

DLPy is an open source package that data scientists can download to apply SAS deep learning algorithms to image, text and audio data. And you don’t need to write SAS code to reap the benefits of deep learning. DLPy is a toolset in a Python-style shell to SAS scripting language and the SAS deep learning actions from SAS® Visual Data Mining and Machine Learning.

DLPy is available in SAS Viya and accessed via Juypter Notebook. DLPy is designed to provide an efficient way to apply deep learning functionalities to solve computer vision, natural language processing, forecasting, and speech processing problems. Key features of DLPy were previously discussed in this blog, SAS deep learning with Python made easy using DLPy These APIs are created following the Keras APIs with a touch of PyTorch flavor. This library is available on GitHub.

We’ve compiled the series of videos into one to introduce these computer vision models. You can also find timestamps for specific areas of interest below.

00:00 - Introduction to the Deep Learning with Python (DLPy) and SAS Viya for Computer Vision video

01:55 - Leverage Functional APIs to Build Complex Models

06:39 - Image Segmentation with U-Net

11:28 - Object Detection with Faster R-CNN

19:18 - Image Classification with ShuffleNet and MobileNet

27:28 - Multi-class Deep Learning for Image Tagging – Training

35:28 - Multi-class Deep Learning for Image Tagging – Inferencing

Model deployment

In the multi-class learning video for inferencing, we use SAS Event Stream Processing for Python (ESPPy). ESPPy is an open source package for building SAS event streaming projects using Python, which is well suited to environments where the data is streaming. The ESP module enables you to connect to a SAS ESP server, manage and construct projects, and interact with windows in your projects. This package is also available on GitHub.

Seeing is believing, so try it for yourself

DLPy makes it easy to take advantage of computer vision with its preconfigured models and pretrained weights. Simply choose your models, modify them and begin deep learning. And if you want to contribute to the DLPy library, create a pull request on GitHub, as SAS gladly accepts them.

Just a note, to use the package for model development, a SAS Visual Data Mining and Machine Learning license is required. If you do not yet have a license, consider a 14-free trial at

Learn more

To learn more about computer vision and how to build computer vision models using SAS Viya, this white paper is a great place to start, How to Do Deep Learning With SAS. To learn more about image embedding, please check out Video: Image embedding using deep learning with Python (DLPy) and SAS Viya.


About Author

Susan Kahler

Global Product Marketing Manager for AI

Susan is a Global Product Marketing Manager for AI at SAS. She has her Ph.D. in Human Factors and Ergonomics, having used analytics to quantify and compare mental models of how humans learn complex operations. Throughout her well-rounded career, she has held roles in user centered design, product management, customer insights, consulting and operational risk. Susan recently completed her Master of Science in Analytics, focusing on healthcare analytics. She also holds a patent for a software navigation system to guide users through dynamically changing systems.

Related Posts

Comments are closed.

Back to Top