# 默认示例程序

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

*gst-ai-classification* 示例应用程序演示了硬件对视频流执行分类的功能。pipeline 接收来自摄像头、文件源或实时流协议（RTSP）的视频流，负责执行预处理，在 AI 硬件上进行推理，并在屏幕上显示结果。

![分类应用的软件流程图](data:image/png;base64,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)

gst-ai-classification 应用程序是 Qualcomm Intelligent Multimedia Product (QIMP) SDK 的一部分。该应用程序可以在刷写设备后直接执行。用户需要将模型和标签文件推送到设备才能运行该应用程序。

## 下载 Qualcomm Neural Processing SDK 的模型和标签文件

如需下载模型和标签文件并将其推送到设备，可在 Linux 主机上执行以下操作。

1. 从 GitHub 下载模型和标签文件。

        wget https://github.com/quic/sample-apps-for-qualcomm-linux/releases/download/GA1.2-rel/GA1.2-rel.zipCopy to clipboard
2. 解压缩文件。

        unzip GA1.2-rel.zipCopy to clipboard
3. 将模型和标签文件推送到设备。
    1. 启用 SSH 以登录主机设备。相关说明，可参见 [如何使用 SSH？](https://docs.qualcomm.com/bundle/publicresource/topics/80-70015-254/how_to.html#use-ssh)
    2. 推送文件。
        - QCS6490

                scp -r GA1.2-rel/QCS6490/* root@<IP address of target device>:/opt/Copy to clipboard
        - QCS9075

                scp -r GA1.2-rel/QCS9075/* root@<IP address of target device>:/opt/Copy to clipboard

## 从 AI Hub 下载 TFLite 的模型和标签文件

1. 从 [https://aihub.qualcomm.com/iot/models](https://aihub.qualcomm.com/iot/models)下载模型。
    有关更新量化模型的 `q_offset` 和 `q_scale` 常量的信息，请参见 [在应用程序中集成 AI Hub 模型](https://docs.qualcomm.com/doc/80-70014-15BY/topic/integrate-aihub-model.html)。
2. 从 Huggingface 下载 YOLOv8 和 YOLO-NAS TFLite 模型或使用 AI Hub API 生成您自己的模型。
    1. 从 Hugging Face 下载模型
        - [YOLOv8-Detection-Quantized](https://huggingface.co/qualcomm/YOLOv8-Detection-Quantized/blob/main/YOLOv8-Detection-Quantized.tflite)
        - [Yolo-NAS-Quantized](https://huggingface.co/qualcomm/Yolo-NAS-Quantized/blob/main/Yolo-NAS-Quantized.tflite)
    2. 使用以下链接安装 AI Hub
        - [YOLOv8-Detection-Quantized](https://github.com/quic/ai-hub-models/tree/main/qai_hub_models/models/yolov8_det_quantized)
        - [Yolo-NAS-Quantized](https://github.com/quic/ai-hub-models/tree/main/qai_hub_models/models/yolonas_quantized)
    3. 使用 AI Hub API 生成您自己的 TFLite 模型
Note: 安装 AI Model Efficiency Toolkit（AIMET）以生成量化模型。有关说明，请参阅 AIMET 安装。
3. 将模型和标签文件推送到设备。
    1. 启用 SSH 以登录主机设备。相关说明，可参见 [如何使用 SSH？](https://docs.qualcomm.com/bundle/publicresource/topics/80-70015-254/how_to.html#use-ssh)
    2. 推送所需的模型文件。

            scp mobilenet_v2_quantized.tflite root@<IP addr of the target device>:/opt/Copy to clipboard
    3. 推送所需的标签。

            wget https://github.com/quic/ai-hub-models/blob/main/qai_hub_models/labels/imagenet_labels.txtCopy to clipboard

            scp imagenet_labels.txt root@<IP addr of the target device>:/opt/Copy to clipboard

Note: 确保推送与模型匹配的标签文件。

Note: 可以使用 Qualcomm 提供的默认模型构建以下示例程序。如果您想使用自己的模型，请参阅 [在应用程序中集成自定义模型](https://docs.qualcomm.com/doc/80-70014-15BY/topic/integrate-custom-model.html)
有关详细信息，请参阅 [Qualcomm GStreamer 插件](https://docs.qualcomm.com/bundle/publicresource/topics/80-70015-50/qim-sdk-plugins.html)。

## 执行示例程序

如需执行示例程序，可运行以下命令：

1. 从主机推送视频文件。

        scp <filename> root@<IP address of target device>:/opt/Copy to clipboard
2. 使用以下命令通过 Qualcomm Neural Processing SDK runtime 执行推理。

        gst-ai-classification --file-path=/opt/video.mp4 --ml-framework=1Copy to clipboard

    将模型和标签文件添加到命令行参数中。下列命令提供 Qualcomm Neural Processing SDK runtime 的默认模型和标签文件路径。

        gst-ai-classification --file-path=/opt/video.mp4 --ml-framework=1 --model=/opt/inceptionv3.dlc --labels=/opt/classification.labelsCopy to clipboard
3. 要运行 TFLite 用例，请从 AI Hub 下载 inception\_v3\_quantized.tflite 模型;请参阅 [从 AI Hub 下载 TFLite 的模型和标签文件](https://docs.qualcomm.com/doc/80-70014-15BY/topic/reference-app.html#concept_q4t_dqx_4bc__section_dl_model_labels_tflite)。
    有关量化模型的 `q_offset` 和 `q_scale` 常量的信息，请参阅 [集成 AI 中心模型](https://docs.qualcomm.com/bundle/publicresource/topics/80-70015-15B/integrate-aihub-model.html)。
4. 推送模型文件。

        scp inception_v3_quantized.tflite root@<IP addr of the target device>:/opt/Copy to clipboard
5. 在使用 TFLite runtime时，可以使用以下命令时对视频运行推理。

        gst-ai-classification --file-path=/opt/video.mp4 --ml-framework=2Copy to clipboard

    将模型和标签文件添加到命令行参数中。以下命令提供 TFLite runtime 的默认模型和标签文件路径。

        gst-ai-classification --file-path=/opt/video.mp4 --ml-framework=2 --model=/opt/inception_v3_quantized.tflite --labels=/opt/classification.labelsCopy to clipboard

## 备注

- 如需停止用例，可按下 CTRL + C。
- 如需显示可用的帮助选项，可运行以下命令：

        gst-ai-classification -hCopy to clipboard

- GStreamer 调试输出由 GST\_DEBUG 环境变量控制。设置所需级别以启用日志。例如，要记录所有警告，应运行以下命令。

        export GST_DEBUG=2Copy to clipboard
- QCS9075 不支持相机，请使用文件或 RTSP 作为视频输入。

Last Published: Jan 26, 2026

[Previous Topic
AI/ML 示例程序](https://docs.qualcomm.com/bundle/publicresource/80-70014-15BY/topics/ai-ml-reference-apps.md) [Next Topic
定制示例程序](https://docs.qualcomm.com/bundle/publicresource/80-70014-15BY/topics/customize-reference-app.md)