## 6. Offload Face detection Node to Remote Workstation
From the workstation, run the face detection node. The face-detection node uses model parameters loaded from the stretch_deep_perception_models directory, whose [path is pulled]([url](https://github.com/hello-robot/stretch_ros2/blob/humble/stretch_deep_perception/stretch_deep_perception/deep_learning_model_options.py#L5)) from HELLO_FLEET_PATH environment variable. In our case, we will set the HELLO_FLEET_PATH environment variable to point to the home folder where the stretch_deep_perception_models directory was cloned.
From the workstation, run the face detection node. The face-detection node uses model parameters loaded from the stretch_deep_perception_models directory, whose [path is pulled](https://github.com/hello-robot/stretch_ros2/blob/humble/stretch_deep_perception/stretch_deep_perception/deep_learning_model_options.py#L5) from HELLO_FLEET_PATH environment variable. In our case, we will set the HELLO_FLEET_PATH environment variable to point to the home folder where the stretch_deep_perception_models directory was cloned.