Local Training

在 Securex NVR 中完成采集、标注、训练、实时测试与部署自定义目标检测模型。

完整工作流

1. 创建数据集。
2. 导入 ZIP / 图片或自动采集。
3. 完成标注。
4. 开始训练并查看实时日志 / 进度。
5. 在模型测试中做图片测试。
6. 在实时模型测试中使用摄像头流验证。
7. 导出 ONNX 和标签文件。
8. 部署到 /config/model_cache。
9. 自动生成或更新检测配置，并进行语法验证。

部署文件

/config/model_cache/securex_custom_yolo.onnx
/config/model_cache/securex_custom_yolo_labels.txt

部署后的检测配置

detectors:
  onnx:
    type: onnx

model:
  model_type: yolo-generic
  width: 640
  height: 640
  input_tensor: nchw
  input_pixel_format: rgb
  input_dtype: float
  path: /config/model_cache/securex_custom_yolo.onnx
  labelmap_path: /config/model_cache/securex_custom_yolo_labels.txt

Local Training

Local Training is part of the Securex Plugin Store ecosystem. This page covers installation, configuration, status checks, updates, and recovery for the plugin.

Installation and configuration

1. Install a release compatible with your Securex NVR version.
2. Open Plugin Settings and configure only the required options.
3. Verify the plugin backend, API, and sidebar entry after installation.

Updates and rollback

Review the plugin change history before updating. Back up configuration first and keep historical release entries instead of replacing them.

Troubleshooting

Check plugin logs, host-service logs, permissions, mounted paths, container dependencies, and version compatibility. Reinstall only after exporting the existing settings.

Configuration examples

/config/model_cache/securex_custom_yolo.onnx
/config/model_cache/securex_custom_yolo_labels.txt

detectors:
  onnx:
    type: onnx

model:
  model_type: yolo-generic
  width: 640
  height: 640
  input_tensor: nchw
  input_pixel_format: rgb
  input_dtype: float
  path: /config/model_cache/securex_custom_yolo.onnx
  labelmap_path: /config/model_cache/securex_custom_yolo_labels.txt

Local Training

Local Training forma parte del ecosistema Securex Plugin Store. Esta página cubre instalación, configuración, estado, actualizaciones y recuperación.

Instalación y configuración

1. Instale una versión compatible con Securex NVR.
2. Abra Ajustes del plugin y configure solo lo necesario.
3. Verifique backend, API y entrada de la barra lateral.

Actualización y reversión

Revise el historial antes de actualizar. Haga copia de seguridad y conserve las versiones históricas.

Solución de problemas

Revise registros del plugin y del host, permisos, rutas montadas, dependencias del contenedor y compatibilidad. Exporte los ajustes antes de reinstalar.

Ejemplos de configuración

/config/model_cache/securex_custom_yolo.onnx
/config/model_cache/securex_custom_yolo_labels.txt

detectors:
  onnx:
    type: onnx

model:
  model_type: yolo-generic
  width: 640
  height: 640
  input_tensor: nchw
  input_pixel_format: rgb
  input_dtype: float
  path: /config/model_cache/securex_custom_yolo.onnx
  labelmap_path: /config/model_cache/securex_custom_yolo_labels.txt