# 部署模型

Source: [https://docs.qualcomm.com/doc/80-70014-15BY/topic/qnn-run-model.html](https://docs.qualcomm.com/doc/80-70014-15BY/topic/qnn-run-model.html)

模型 .so（量化或非量化）可以通过支持 QNN 的应用程序（使用 QNN C/C++ API 编写的应用程序）进行部署。QNN 提供 API 来动态加载模型 .so 并使用所选后端在硬件上运行该模型。

QNN 提供了一个预编译工具 ([qnn-net-run](https://docs.qualcomm.com/bundle/publicresource/topics/80-63442-50/tools.html#qnn-net-run))，该工具可以动态加载此模型 .so 并使用提供的输入在指定后端执行推理。

对于 CPU、GPU 或 HTP 执行，` qnn-net-run` 需要三个参数：

- **模型文件** – 由...生成的文件 **`qnn-model-lib-generator`**
- **后端文件**- 目标后端所需的 .so 文件
    - `libQnnCpu.so `用于 CPU 后端。
    - `libQnnGpu.so` 用于 GPU 后端。
    - `libQnnHtp.so `用于 HTP 后端。

- **输入列表** - 文本文件，如 `input_list.txt `量化期间使用的文件，但此列表中的输入原始文件除外（这些文件用于推理）。在本示例中，为简单起见，将量化时使用的同一输入列表用于推理。

## 在 x86 后端运行 QNN 模型

若要在 x86 CPU 上运行模型， `qnn-net-run` 需要模型 .so、后端库 .so 和输入列表来运行推理以生成输出。

例如，以下命令会加载 libinception\_v3.so 模型并在 x86 CPU 上运行该模型。执行完成后，该 `qnn-net-run` 工具将输出文件写入 `~/models/output_x86`。

    ${QNN_SDK_ROOT}/bin/x86_64-linux-clang/qnn-net-run --model ~/models/libs/x86_64-linux-clang/libinception_v3.so --backend ${QNN_SDK_ROOT}/lib/x86_64-linux-clang/libQnnCpu.so --input_list input_list.txt --output_dir ~/models/output_qnn_x86Copy to clipboard

## 准备 QNN 模型以在设备上运行

在目标设备上运行模型之前，要确保 QNN 二进制文件和库连同模型和输入文件已推送到目标中。

使用来自 `${QNN_SDK_ROOT}/bin/aarch64-oe-linux-gcc11.2` 和的工件 `${QNN_SDK_ROOT}/lib/aarch64-oe-linux-gcc11.2`

对 Qualcomm Linux 开发套件使用正确的 DSP Hexagon 架构库。
- QCS9075：使用 v73 库
- QCS6490：使用 v68 库

| **文件** | **源位置** |
| --- | --- |
| `qnn-net-run` | `${QNN_SDK_ROOT}/bin/aarch64-oe-linux-gcc11.2` |
| `libQnnHtp.so` | `${QNN_SDK_ROOT}/lib/aarch64-oe-linux-gcc11.2/libQnnHtp.so` |
| `libQnnCpu.so` | `${QNN_SDK_ROOT}/lib/aarch64-oe-linux-gcc11.2/libQnnCpu.so` |
| `libQnnGpu.so` | `${QNN_SDK_ROOT}/lib/aarch64-oe-linux-gcc11.2/libQnnGpu.so` |
| `libQnnHtpPrepare.so` | `${QNN_SDK_ROOT}/lib/aarch64-oe-linux-gcc11.2` |
| `libQnnHtpV68Stub.so` | `${QNN_SDK_ROOT}/lib/aarch64-oe-linux-gcc11.2` |
| `libinception_v3_quantized.so` | `~/models/libs/aarch64-oe-linux-gcc11.2` |
| libinception\_v3.so | ~/models/libs/aarch64-oe-linux-gcc11.2 |
| `libQnnHtpV68Skel.so` | `${QNN_SDK_ROOT}/lib/hexagon-v68/unsigned` |
| `libqnnhtpv68.cat` | `${QNN_SDK_ROOT}/lib/hexagon-v68/unsigned` |
| `libQnnSaver.so` | `${QNN_SDK_ROOT}/lib/hexagon-v68/unsigned` |
| `libQnnSystem.so` | `${QNN_SDK_ROOT}/lib/hexagon-v68/unsigned` |

需要在主机上执行以下 `scp` 命令，将 QNN SDK 库和二进制文件复制到设备。
Note: `<dsp-arch>` QCS9075 替换为 73，QCS6490 替换为 68

    scp ${QNN_SDK_ROOT}/bin/aarch64-oe-linux-gcc11.2/qnn-* root@[ip-addr]:/opt/
    scp ${QNN_SDK_ROOT}/lib/aarch64-oe-linux-gcc11.2/libQnn*.so root@[ip-addr]:/opt/           
    scp ${QNN_SDK_ROOT}/lib/hexagon-v<dsp-arch>/unsigned/* root@[ip-addr]:/opt/
          
    scp ~/models/libs/aarch64-oe-linux-gcc11.2/* root@[ip-addr]:/opt/            
    scp ~/models/input_list.txt root@[ip-addr]:/opt/
                
    scp /tmp/RandomInputsForInceptionV3 root@[ip-addr]:/tmp/
    
    ssh root@[ip-addr] 
    
    cd /opt/Copy to clipboard

## 在 Arm CPU 上运行 QNN 模型

在基于 Arm 的 CPU 上执行模型时， `qnn-net-run`需要模型 .so（对于 RB3Gen2 目标，.so 必须与 aarch64-oe-linux-gcc11.2 工具链交叉编译）、后端 .so 库和输入列表来运行推理以生成输出。

例如，以下命令将输出文件写入`/opt/output_cpu`。

    export LD_LIBRARY_PATH=/opt/:$LD_LIBRARY_PATH 
    export PATH=/opt:$PATH
    
    qnn-net-run --model libinception_v3.so --backend libQnnCpu.so --input_list input_list.txt --output_dir output_cpuCopy to clipboard

## 在 GPU 上运行 QNN 模型

在 Adreno GPU 上执行模型时， `qnn-net-run`需要模型 .so、后端 .so 库（libQnnGpu.so）和输入列表来运行推理以生成输出。

例如，以下命令将输出文件写入 `/opt/output_gpu`。

export LD_LIBRARY_PATH=/opt/:$LD_LIBRARY_PATH 
    export PATH=/opt:$PATH
    
    qnn-net-run --model libinception_v3.so --backend libQnnGpu.so --input_list input_list.txt --output_dir output_gpuCopy to clipboard

## 在 HTP 后端运行 QNN 模型

要在 HTP 后端运行模型，请使用 `libquantized_inception_v3.so` 库和 `libQnnHtp.so `后端库并将输出保存在 output\_htp 目录中。

    export LD_LIBRARY_PATH=/opt/:$LD_LIBRARY_PATH 
    export PATH=/opt:$PATH
    export ADSP_LIBRARY_PATH="/opt/;/usr/lib/rfsa/adsp;/dsp"
    
    qnn-net-run --model libinception_v3_quantized.so --backend libQnnHtp.so --input_list input_list.txt --output_dir output_htpCopy to clipboard

## 对输出进行验证

对于馈送到 `qnn-net-run`（通过 `input_list`文件）的每个输入原始文件，都会生成一个输出文件夹，其中包含输出原始文件，其大小将与模型的输出图层匹配，如下图所示。（[Netron](https://netron.app/) 是一种模型可视化工具）

![](data:image/jpeg;base64,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)
在此示例中， `inception_v3`输出的原始文件是一个二进制文件，其中包含 1,000 个分类类的概率。

我们使用 Python 脚本将文件读取为 NumPy 数组，以执行后处理步骤，并对输出进行验证。下方示例用于检查 HTP 预测结果是否与 CPU 预测结果相同。

    ssh root@[ip-addr]
    cd /opt/
    
    scp -r /opt/output_cpu user@host-ip:<path>
    scp -r /opt/output_htp user@host-ip:<path>Copy to clipboard

Note: `<path>` 在以下 python 脚本中使用上述命令。

将工具的 `qnn-net-run` 输出从设备复制到主机后，我们可以创建一个简单的 Python 脚本来加载 CPU 和 HTP 执行的输出，并使用 NumPy 进行比较。

    #python postprocessing script (compare.py)
    
    import numpy as np
    
    htp_output_file_path = "<path>/output_htp/Result_1/875.raw" 
    cpu_output_file_path = "<path>/output_cpu/Result_1/875.raw"
     
    htp_output = np.fromfile(htp_output_file_path, dtype=np.float32)
    htp_output = htp_output.reshape(1,1000)
     
    cpu_output = np.fromfile(cpu_output_file_path, dtype=np.float32)
    cpu_output = cpu_output.reshape(1,1000)
     
    cls_id_htp = np.argmax(htp_output)
    cls_id_cpu = np.argmax(cpu_output)

    # Let's compare CPU output vs HTP output
    print("CPU prediction {} \n HTP prediction {}".format(cls_id_cpu, cls_id_htp))Copy to clipboard

**输出**

![](data:image/png;base64,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)

## 使用 QNN API 部署模型

Qualcomm AI Engine Direct SDK 提供 C/C++ API，用于创建/开发应用程序，以加载完成编译的 .so 模型并在所选后端（CPU、GPU 或 HTP）上加速执行。[此处](https://docs.qualcomm.com/bundle/publicresource/topics/80-63442-50/sample_app.html)是一个使用 SNPE C/C++ API 开发的用于执行模型的示例程序。

## 使用 Qualcomm IM SDK 部署模型

为了改善开发者在构建整个用例 pipeline（来自摄像头的流、预处理图像、执行推理等）时的体验，QNN 已作为 `qtimlqnn` 插件集成到 Qualcomm IM SDK 中。该插件是基于 QNN C API 开发的，提供了动态加载和执行模型的功能。借助该 `qtimlqnn` 插件，开发者可以将他们转换和编译的模型库插入到 Qualcomm IM SDK pipeline 中，以实现他们的用例。

有关如何使用 Qualcomm IM SDK 部署 QNN 模型的说明，可参见 [开发您自己的应用程序](https://docs.qualcomm.com/doc/80-70014-15BY/topic/develop-own-app.html) 章节。

Last Published: Jan 26, 2026

[Previous Topic
QNN 模型移植](https://docs.qualcomm.com/bundle/publicresource/80-70014-15BY/topics/qnn-port-model.md) [Next Topic
AI/ML 示例程序](https://docs.qualcomm.com/bundle/publicresource/80-70014-15BY/topics/ai-ml-reference-apps.md)