Skip to main content
AI/ML reference applications demonstrate how to deploy and run models across common real-world scenarios, including live camera feeds, video files, and RTSP streams on Qualcomm evaluation kits. These applications can be executed on your local EVK, depending on your development workflow and hardware availability. Select the option that best aligns with your hardware setup, development environment, and use case.

Experience AI with Qdemo

Run preinstalled AI use cases through the Qdemo graphical interface on your EVK.

Run Qualcomm IM SDK Sample Application

Run AI/ML sample applications on a Qualcomm Dragonwing IoT platform using the Qualcomm IM SDK.

Run LiteRT / TFLite models

Run quantized LiteRT models on the NPU of a Qualcomm Dragonwing device using AI Engine Direct delegates in Python or C++.

Run an ONNX model on the NPU

Run an ONNX model on the NPU of a Qualcomm Dragonwing device using ONNX Runtime.

Run LLMs/VLMs using llama.cpp

Run large language and vision-language models locally using llama.cpp with CPU, GPU, or NPU backends.