Skip to main content

Example models

The public github repository sa.example-models contains larger examples of OSQL models. These can be used as a reference to learn how to create models with different concepts.

This section walks through these example models and explain the concepts they contain. It can be read stand-alone, but preferably together with the repository.

Fuel efficency

This model predicts fuel efficency of automobiles from the Auto MPG dataset using trained regression models. The ML models are trained using Python and Tensorflow, and inferred using OSQL.

You will learn about:

  • Using Python foreign functions and passing data between Python and OSQL
  • Inference of trained models in OSQL
  • Storing weights for deploying and inferring trained models on devices without Python.

Statistical analysis of large datasets

This guide shows how to explore and create models on a large dataset with Accelerometer, Tachometer and PWM data.

You will learn about:

  • Transforming a dataset to a streaming format
  • Exploring datasets and calculating statistics with OSQL
  • Creating advanced plots in SA Engine