Skip to main content

Quick Start with SL1600 Series

These are self-guided beginner tutorials that will give you hands-on experience and code examples for running edge AI models spanning vision, speech, and even large language models.

Important

These examples are designed to work with Astra SL1680 processor using Astra Machina Dev Kit. Vision examples leverage NPU and other examples leverage CPU.

For Astra SL1640 processor (leveraging NPU), these examples can be still run, after adding required set of packages into the OOBE image via bitbake.

For Astra SL1620 processor (leveraging GPU), Vision - Classification Model is not pre-installed. Other examples can be still run, after adding required set of packages into the OOBE image via bitbake.

Prerequisites​

Before we dive into Edge AI development, you need to make sure you have:

  • A Synaptics Astra Machina development kit (Buy one here)

  • Hardware setup ready (Follow guide here).Follow the guide to boot your board and open a terminal.

SyNAP C++ Framework

The Synap Python API binds the Synap C++ Framework. You can read more about the C++ framework here and find C++ source code for NPU accelerated inference applications on GitHub.

Software​

All the examples below are Python-based, so before proceeding you need to set up the necessary libraries and packages. When you buy a Machina Dev board from distribution, it's already programmed with a OOBE SDK Image with essential packages like pip and python3 pre-installed.

Clone our Examples GitHub repository and Navigate to the Repository Directory:

git clone https://github.com/synaptics-synap/examples.git
cd examples

Set up your Python environment ensuring all required dependencies are installed within a virtual environment:

python3 -m venv .venv --system-site-packages
source .venv/bin/activate

For Astra OOBE SDK 2.0 (scarthgap) and above, Please use Python3.12 packages:

pip install -r requirements-py312.txt

For Astra OOBE SDK 1.8 (kirkstone) and below, Please use Python3.10 packages:

pip install -r requirements.txt

Install latest SynapRT Python package which allows you to run real-time AI pipelines:

Run Edge AI Python Examples​

Image Classification

In this Quick start guide, we will use a MobileNet Model for Image classification.

Object Detection

Object detection is a computer vision technique that identifies objects in images or videos.

Speech-to-text

Convert sound to text and build features like dictation and voice commands.

Text-to-Speech

Here we look at running TTS systems which turn written text into human-sounding speech.

Embeddings

Embedding models transform text, images, or signals into vectors that represent meaning.

AI Text Assistant

Lets build an AI helper that uses embeddings to give you the best answers.

AI Tool Calling Assistant

Lets build an AI helper that uses embeddings to call external functions or APIs.

Large Language Models

A quick guide to run Large language models like Qwen and Deepseek.