This sample provides the supporting SQL and R scripts for the blogpost How six lines of code + SQL Server can bring Deep Learning to ANY App.

Data: Galaxy Zoo project was used as source of labeled training data.

Scripts: The following scripts are provided

• createTables.sql: create tables for trained models and scored data.
• train_NN_model.sql: stored procedure for training NN model with Microsoft ML
• predict_NN_model.sql: stored procedure for scoring
• trigger_predict_model.sql: script to invoke scoring.

For end-to-end training and prediction the images could be downloaded from public storage. See this links for more info:

• Morphological classifications of main-sample spectroscopic galaxies from Galaxy Zoo 2
• Galaxy Zoo 2: detailed morphological classifications for 304,122 galaxies from the Sloan Digital Sky Survey
• SkyServer SQL Search