The Ara AI-software development kit (SDK) streamlines deployment of AI models onto Ara-1 and Ara240 discrete neural processing units (DNPUs) and modules. It features an extensible compiler that supports diverse model architectures, including convolutional neural network (CNN), large language models (LLMs) and vision language models (VLMs). The SDK offers efficient dataflow optimization, flexible quantization methods and support for multiple datatypes, ensuring high-performance and low-latency for edge-AI applications.
Choose a diagram:
| Release and Documentation | Description | Build Sources | Supported Platforms/Binaries |
|---|---|---|---|
|
Ara2 Runtime Package |
Start Here - Runtime SDK for AI/ML acceleration with Ara240 DNPU on i.MX SoCs. Means to perform benchmarks with pre-compiled CNN models. |
|
Supported Platforms: |
|
Ara2 Vision Examples |
Package containing GStreamer pipeline for 8 video streams using YOLOv8n model on Ara240 DNPU |
|
Supported Platforms: |
|
LLM Edge Studio |
Qt based launcher for unimodal LLMs running on Ara240 |
|
Supported Platforms: |
|
VLM Edge Studio |
Multimodal GenAI at the Edge - Demo showing Ara240 multimodal capabilities on i.MX with VLM, for Video and Image Analysis |
|
Supported Platforms: |
|
eIQ AAF Connector |
A service that exposes an OpenAI‑compatible REST API to client applications while brokering requests to locally hosted LLM/VLM backends on Ara240 DNPU |
|
Supported Platforms: |
5 downloads
Note: For better experience, software downloads are recommended on desktop.
4 hardware offerings
