datasciencemonkey / img_rec_app-ai4Viya

Sir Fred - Flask App for Simple Image Recognition.

Uses Fast & Reliable Event Driven Deployment with ESP. A quick usecase agnostic way to demonstrate the demo assets built with SAS DL & ESP.

This directory has all the required material to successfully try out the application. You can bring your own projects and use the app as a way to demonstrate ASTORE Scoring with ESP.

Folders & Files included:

a. model_creation-jupyter_notebooks

These are example notebooks to create models using dlpy. You may choose to use the action sets directly, if thats your preference, but the example here is to quickly show how things work in training.

b. define_esp_model

This directory has a notebook showing an example of defining an ESP Project.

Files for the web app to function :

• fred_esp.py - this file has the application logic. Reaches out to ESP as needed to complete the request
• static - this folder has all the static content required. Uploads from the client browsers are saved here before being sent to ESP for scoring as a base64 encoded file
• templates - contains a simple html template
• sample_test_images_retail - couple of sample images to test the app