Google DeepMind has launched the Gemini Robotics On-Device, a language model that allows robots to operate independently from the internet. This model builds upon the earlier Gemini Robotics version from March and showcases its ability to control robot movements using natural language inputs from developers. The performance of Gemini Robotics On-Device is reportedly comparable to the cloud-based model. Demonstrations showed robots executing tasks like unzipping bags and folding clothes. In addition, the company introduces a Gemini Robotics SDK, allowing developers to train robots through novel task demonstrations.
Google DeepMind released the Gemini Robotics On-Device model, enabling local control of robots without internet access, emphasizing adaptability and performance.
Collection
[
|
...
]