AI Software Suite(Linux-Ubuntu)

AI Software Suite

AI Software Suite offers breakthrough AI accelerators and Vision processors uniquely designed to accelerate embedded deep learning applications on edge devices.

Qualcomm devices are accompanied by a comprehensive AI SDK that enables the compilation of deep learning models and the implementation of AI applications in production environments. The model build environment seamlessly integrates with common ML frameworks to allow smooth and easy integration in existing development ecosystems. The runtime environment enables integration and deployment in host processors, such as ARM based products, when utilizing IQ9075 vision processor.

---

• SNPE

  SNPE is a SDK designed to streamline the development and deployment of edge applications demanding high AI performance. This reference application software package empowers users to expedite their time-to-market by minimizing the development workload. SNPE encompasses a user-friendly set of fully operational application examples based on GStreamer, featuring pipeline elements and pre-trained AI tasks. These examples leverage advanced Deep Neural Networks, highlighting Qualcomm's AI processors' top-notch throughput and power efficiency. Furthermore, SNPE serves as a demonstration of Qualcomm's system integration capabilities, showcasing specific use cases on predefined software and hardware platforms. Utilizing SNPE simplifies integration with Qualcomm's runtime software stack and offers a starting point for users to fine-tune their applications. By demonstrating Qualcomm's system integration scenarios on both predefined software and hardware platforms, it can be used for evaluations, reference code, and demos. This approach effectively accelerates time to market, streamlines integration with Qualcomm's runtime software stack, and provides customers with a foundation to fine-tune their applications.

---

Application

• VisionAI

  • AIR-055

  Application (FPS)	CPU	iGPU	NPU	Note
  Object Detection	10	40	200	YOLOv11n-w8a16
  Face Detection	12	20	55	Lightweight-Face-Detection-w8a8
  Pose Estimation	1	4	18	YOLOv11n-w8a16 +HRNetPose-w8a8

  

• Generate AI

  Platform	Version	Application	Model	Note
  AIR-055	v3.6.1	GenAI Chatbot	Qwen2.5-7B-Instruct	Supports only the Qwen2.5-7B-Instruct model.

  

---

Benchmark

It can quickly evaluate computing performance for the NPU, and provides runtime results inferenced with ML.

• Run on NPU ( Terminal )

  QAIRT_ROOT="/opt/qcom/aistack/qairt/2.41.0.251128/lib" ADSP_LIBRARY_PATH="/opt/qcom/aistack/qairt/2.41.0.251128/lib/hexagon-v73/unsigned" LD_LIBRARY_PATH="/opt/qcom/aistack/qairt/2.41.0.251128/lib/aarch64-oe-linux-gcc11.2:$LD_LIBRARY_PATH" PATH="/opt/qcom/aistack/qairt/2.41.0.251128/lib/aarch64-oe-linux-gcc11.2:$PATH" /opt/Advantech/EdgeAI/System/Qualcomm_IQ9/SDK/benchmark/snpe-throughput-net-run --duration 1 --perf_profile burst --use_dsp --userbuffer_auto --container /opt/Advantech/EdgeAI/System/Qualcomm_IQ9/SDK/benchmark/mobilenet_v2-mobilenet-v2-w8a16.dlc

---

Optimization

Performance

Platform	Model	CPU	NPU	Optimization	Note
AIR-055 (IQ9075)	YOLOv11	8	200	How-To	CPU: ONNX FP32, NPU: YOLOv11 w8a16, OS: Ubuntu

Accelerator

Accelerator	Description	Note
SNPEBuilder	Builds the SNPE network, selects the runtime processor order, output layers, and performance profile.	LINK
ITensor	SNPE-managed tensor buffer used to create the model input tensor and copy input data before inference.	LINK
SNPE::execute()	Executes inference with input and output tensor maps on the selected SNPE runtime.	LINK

---

System Monitoring

---

Advantech AI Devices

AI System