SAS Demo | Deep Learning with Python (DLPy) and SAS Viya
In this SAS demo, you will learn about SAS Deep Learning Python API, or DLPy for short. DLPy enables data scientists familiar with Python to take advantage of the deep learning features in SAS Viya. DLPy is available at – https://github.com/sassoftware/python-dlpy
With DLPy, you can apply deep learning to computer vision tasks, such as image classification and object detection. DLPy also supports natural language processing (NLP) tasks, such as text classification and text generation.
In addition to these tasks, DLPy has APIs to use SAS’s deep learning features to analyze time series data for forecasting. DLPy also supports the open neural network exchange, or ONNX project, to easily move models between frameworks. SAS is a member of the ONNX community.
Here is what is covered in this demonstration.
00:00 – Introduction to the Deep Learning with Python (DLPy) and SAS Viya video
02:05 – Image Classification with Convolutional Neural Networks
See how to use our high-level Python API to build your own image classification deep learning model using retail product images. Image classification serves a wide range of business use cases from insurance claims, medical imagery, security screenings, et cetera.
06:30 – Object Detection with Tiny YOLOv2
See how to use our high-level Python API, DLPy, to solve an object detection test. The learning objective is to understand how you can use SAS DLPy to create your own object detection training data sets and train and perform object detection with a pre-defined model.
10:27 – Import and Export Deep Learning Models with ONNX
See how to integration between DLPy and ONNX. We will walk through how to train a CNN in SAS and export the trained model to ONNX. We will also see how to import a pre-trained Tiny YOLOv2 model from ONNX into SAS.
15:43 – Text Classification and Text Generation Using Recurrent Neural Networks
See how to use our high-level Python API, DLPy, to show some of our recurrent neural network capabilities. I will demonstrate the Python interface in a Jupyter notebook. Here I will focus on a sentiment analysis problem and build deep learning models for this task.
19:45 – Time Series Forecasting with Recurrent Neural Networks
See how to use our high-level Python API, DLPy, to preprocess time series data into an appropriate format that can be analyzed by a recurrent neural network model. The data will first be described and then you will see how to preprocess the data using DLPy. Then, you will be shown how to use DLPy to build a simple recurrent neural network model for forecasting and visualizing the results.
DLPy is available at – https://github.com/sassoftware/python-dlpy
• Python version 3 or greater is required
• Install SAS Scripting Wrapper for Analytics Transfer (SWAT) for Python – https://github.com/sassoftware/python-swat
• Access to a SAS Viya 3.4 environment with Visual Data Mining and Machine Learning (VDMML) is required – https://www.sas.com/en_us/software/visual-data-mining-machine-learning.html
• A user login to your SAS Viya back-end is required. See your system administrator for details if you do not have a SAS Viya account.
• It is recommended that you install the open source graph visualization software called Graphviz to enable graphic visualizations of the DLPy deep learning models – https://www.graphviz.org/download/
SUBSCRIBE TO THE SAS USERS YOUTUBE CHANNEL #SASUsers #LearnSAS
SAS is a trusted analytics powerhouse for organizations seeking immediate value from their data. A deep bench of analytics solutions and broad industry knowledge keep our customers coming back and feeling confident. With SAS®, you can discover insights from your data and make sense of it all. Identify what’s working and fix what isn’t. Make more intelligent decisions. And drive relevant change.
CONNECT WITH SAS
SAS ► http://www.sas.com
SAS Customer Support ► http://support.sas.com
SAS Communities ► http://communities.sas.com
Facebook ► https://www.facebook.com/SASsoftware
Twitter ► https://www.twitter.com/SASsoftware
LinkedIn ► http://www.linkedin.com/company/sas
Blogs ► http://blogs.sas.com