ai-hub-apps
The Qualcomm® AI Hub apps are a collection of state-of-the-art machine learning models optimized for performance (latency, memory etc.) and ready to deploy on Q
What is ai-hub-apps?
The Qualcomm® AI Hub apps are a collection of state-of-the-art machine learning models optimized for performance (latency, memory etc.) and ready to deploy on Qualcomm® devices.
Overview of ai-hub-apps
What it is
Qualcomm® AI Hub Apps is a collection of sample apps and tutorials that help deploy machine learning models on Qualcomm® devices. Each app is designed to work with one or more models from Qualcomm® AI Hub Models. The repository provides open-source app recipes for running models on local devices, covering various machine learning tasks.
Key features
- Collection of sample apps and tutorials for on-device deployment of machine learning models.
- Supports multiple runtimes: TensorFlow Lite, ONNX, and Genie SDK (Generative AI runtime on top of Qualcomm® AI Engine Direct SDK).
- Deployment targets include Android 11 (API v30+), Windows 11, and Ubuntu 24.04+.
- Supports CPU, GPU, and NPU (including Hexagon HTP) compute units.
- NPU acceleration supported on Snapdragon® chipsets (e.g., Snapdragon X Elite, Snapdragon 8 Elite, Snapdragon 8 Gen 3, etc.) with FP16 or INT8/INT16 weight and activation types.
- Apps are organized by operating system (Android, Windows, Ubuntu) and include tasks like image classification, object detection, semantic segmentation, super resolution, speech-to-text, chat, and more.
Use cases
- Deploying image classification, object detection, semantic segmentation, super resolution, and speech-to-text on Android devices using TensorFlow Lite.
- Running image classification, object detection, super resolution, speech-to-text, and image generation on Windows using ONNX.
- Performing hand gesture recognition, pose estimation, audio classification, portrait segmentation, face recognition, 3D object detection, semantic segmentation, LiDAR segmentation, and text recognition on Ubuntu using TensorFlow Lite.
- Deploying large language models (LLMs) on-device using Genie SDK or GenieX runtime.
Getting started
See the repository README for detailed build and installation instructions for each app. The README provides tables listing apps by OS and task, and each app's own README contains specific instructions.
License and pricing
BSD-3-Clause, open source.
Based on the project's public README and repository metadata.