https://github.com/qualcomm/snapdragon-compute-samples
https://github.com/qualcomm/snapdragon-compute-samples
Last synced: 2 months ago
JSON representation
- Host: GitHub
- URL: https://github.com/qualcomm/snapdragon-compute-samples
- Owner: qualcomm
- License: bsd-3-clause
- Created: 2025-04-16T21:08:47.000Z (about 1 year ago)
- Default Branch: main
- Last Pushed: 2026-02-11T17:27:06.000Z (5 months ago)
- Last Synced: 2026-03-10T22:11:58.067Z (4 months ago)
- Language: Jupyter Notebook
- Size: 2.45 MB
- Stars: 2
- Watchers: 1
- Forks: 1
- Open Issues: 2
-
Metadata Files:
- Readme: README.md
- Contributing: CONTRIBUTING.md
- License: LICENSE.txt
- Code of conduct: CODE-OF-CONDUCT.md
- Security: SECURITY.md
Awesome Lists containing this project
README
# Snapdragon Compute Sample Apps
This repository contains sample apps for running ONNX models efficiently using [ONNX Runtime](https://onnxruntime.ai/), specifically targeting Qualcomm Hexagon NPU with [QNN Execution Provider](https://onnxruntime.ai/docs/execution-providers/QNN-ExecutionProvider.html)
## Table of Contents
1. [Available Apps](#available-apps)
2. [Project Status](#project-status)
3. [Getting Started](#getting-started)
4. [Quick Start](#quick-start)
5. [Contributing](#contributing)
6. [Directory Structure](#directory-structure)
7. [License](#license)
## Available Apps
| App Name | Model Used | Providers | Quick Start | Notebook | Notes |
|------------------------|------------|-----------|--------------------------------------------------------------------------------------|---------------------------------------------------------------------------------------------------|-------|
| 'HRNet Pose Detection' | HRNetPose | CPU |[Complete](https://github.com/DerrickJ1612/qnn_sample_apps/tree/main/src/hrnet_pose) |[Complete](https://github.com/DerrickJ1612/qnn_sample_apps/tree/main/notebooks/pose_detection) | None |
| 'DeepSeek Local' | DeepSeek | QNN |[Complete](https://github.com/DerrickJ1612/qnn_sample_apps/tree/main/src/deepseek_r1) |[Complete](https://github.com/DerrickJ1612/qnn_sample_apps/tree/main/notebooks/reasoning_llm) | None |
| 'Prompt Pong' | Qwen2.5 Coder | CPU | [Complete](./src/pong/) | ❌ | An example using Ollama. Compatible with other LLM models |
These apps demonstrate end-to-end inference using ONNX Runtime on devices with Hexagon NPUs. Each app includes:
- Input preprocessing
- Onnx model inference
- Output postprocessing
Supported features:
- CPU fallback (if you don't have access to Hexagon NPU)
- Hexagon QNN Execution
## Getting Started
#### General Requirements
- Python (version 3.11.+)
- If targeting Hexagon ensure you install ARM64 compatible Python version
- C and C++ support in Visual Studio
- [Installation Instructions](https://learn.microsoft.com/en-us/cpp/build/vscpp-step-0-installation?view=msvc-170)
- Rust Support
- [Installation Instructions](https://rustup.rs/)
#### 1. Clone the Repository
```
>> git clone https://github.com/qualcomm/snapdragon-compute-samples.git
```
#### 2. Setup Virtual Environment
```
>> python -m venv venv
>> venv\Scripts\activate.ps1 # Linux: >> source venv/bin/activate
>> pip install -r \src\\requirements.txt # Ex: pip install -r \src\deepseek_r1\requirements.txt
```
#### 3. Download Models
| Model Name | Description | Download Source |
|-------------|--------------------------|---------------------------------------------------------------------------------------------------------------|
| HRNetPose | Human pose estimation | [AI Hub](https://aihub.qualcomm.com/compute/models/hrnet_pose?domain=Computer+Vision&useCase=Pose+Estimation) |
| DeepSeek R1 | Reasoning Language Model | [s3 Bucket](tbd) |
#### 4. Run models.py
models.py will automatically place models in appropriate destination
```
>> python models.py --model_directory (absolute path to directory where models were downloaded)
```
## Quick Start
| App Name | CLI Command |
|------------------------|---------------------------------------------|
| 'HRNet Pose Detection' | ` >> python ./src/hrnet_pose/main.py ` |
| 'DeepSeek Local' | ` >> python ./src/deepseek_r1/main.py ` |
## Contributing
We welcome contributions to this repository! Please refer to our [contributing guide](CONTRIBUTING.md) for how to contribute.
## Directory Structure
## License
This project is licensed under the [MIT](https://github.com/DerrickJ1612/qnn_sample_apps/blob/main/LICENSE.txt)