Some tasks from @Home require specific image information that ranges from person descriptions to finding certain details within an image (e.g. how many people are raising their hand). These tasks either require a whole script all for themselves or a trained model. Recent Vision-LLM models such as GPT-Vision or Moondream can handle these tasks.
This issue will address Docker + ROS integration of Moondream2 as an alternative for handling these problems, running the model on an Nvidia Jetson Xavier.
Some tasks from @Home require specific image information that ranges from person descriptions to finding certain details within an image (e.g. how many people are raising their hand). These tasks either require a whole script all for themselves or a trained model. Recent Vision-LLM models such as GPT-Vision or Moondream can handle these tasks.
This issue will address Docker + ROS integration of Moondream2 as an alternative for handling these problems, running the model on an Nvidia Jetson Xavier.