Lightweight multimodal model for robotics perception and interaction.
Robot environment awareness and human-robot interaction.
Not suitable for high-resolution image processing.
