# QNN 워크플로우

개발자는 AI Runtime을 사용하여 ONNX, TensorFlow, TensorFlow Lite 또는 PyTorch와 같이 지원되는 AI 프레임워크에서 사전 학습된 모델을 선택하고 Snapdragon X 플랫폼의 AI 엔진 하드웨어에서 추론에 사용할 수 있는 형식으로 변환할 수 있습니다.

다음 다이어그램은 모델 추론에서 와트당 최상의 성능을 얻기 위해 따라야 할 일반적인 워크플로우를 보여줍니다.

![../../_images/ai-engine-workflow.png](data:image/png;base64,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)

이 프로세스의 단계는 다음과 같습니다.

1. **모델 학습**: 이 단계는 이 문서의 범위를 벗어납니다. 이미 AI Hub에서 사전 학습된 모델이 있거나 써드파티 AI 프레임워크를 사용하여 모델을 개발했을 것으로 예상됩니다.
2. **모델 변환**: 이 단계에서는 사전 학습된 모델을 그래프에 대한 *model.cpp* 파일과 가중치를 포함하는 *model.bin* 으로 구성된 AI Runtime 형식으로 변환합니다.

    **양자화**: 모델 변환기 도구는 NPU에서 지원되는 정수 정밀도(예: INT4, INT8 또는 INT16)로 모델을 양자화하는 것도 지원합니다. 이 단계는 모델 추론을 위해 FP32 정밀도를 지원하는 NPU 또는 CPU Runtime에서 FP16 모델을 실행하는 데는 필요하지 않습니다.
3. **모델 생성**: 이 단계에서는 이전 단계에서 생성한 model.cpp 파일을 사용하여 *model.dll* 을 생성합니다. 이 *model.dll* 은 AI 엔진 하드웨어에서 모델을 추론하는 데 사용될 수 있습니다.
4. **컨텍스트 바이너리 준비**: NPU 추론의 경우, 이 단계를 오프라인에서 선택적으로 사용하여 3단계에서 생성된 *model.dll* 에서 컴파일된 컨텍스트 바이너리를 생성할 수 있습니다. 이 단계를 오프라인에서 사용하면 모델을 미리 컴파일하여 애플리케이션 시작 시 모델 초기화 시간을 단축하고 사용자 경험을 개선할 수 있습니다.
5. **모델 실행**: 이 단계에서는 3단계와 4단계에서 각각 생성된 *model.dll* 또는 컨텍스트 바이너리를 사용하며 AI 엔진 하드웨어에서 모델을 실행합니다.

참고

모델을 퀄컴 하드웨어에 맞게 변환하고 양자화해야 하므로 추론을 실행하기 전에 모델을 커스터마이즈해야 합니다.

Last Published: Jan 13, 2026

[Previous Topic
Qualcomm AI Rumtime(QNN)](https://docs.qualcomm.com/bundle/publicresource/80-62010-1KO/topics/qnn.md) [Next Topic
설정](https://docs.qualcomm.com/bundle/publicresource/80-62010-1KO/topics/ai-engine-direct.md)