InternRobotics/InternVLA-N1-System2
InternRobotics/InternVLA-N1-System2 is a 7 billion parameter navigation foundation model developed by InternRobotics, designed for high-level reasoning in autonomous navigation. This model, part of a multi-system architecture, focuses on System 2 capabilities for planning and instruction-following. It excels in zero-shot sim-to-real generalization, achieving leading results on VLN benchmarks for robust navigation in both simulated and real-world environments.
Loading preview...
InternVLA-N1-System2: A Dual-System Navigation Foundation Model
InternVLA-N1-System2 is a 7 billion parameter navigation foundation model developed by InternRobotics, specifically designed as the System 2 component within a novel multi-system architecture. This model focuses on high-level reasoning and planning, working asynchronously with a System 1 for low-level action and control. It represents the first Dual-System Vision-Language Navigation (VLN) Foundation Model, enabling robust instruction-following navigation.
Key Capabilities
- Modular Multi-System Design: Integrates System 2 (reasoning/planning) with System 1 (action/control) for efficient and robust navigation.
- Zero-Shot Sim2Real Generalization: Demonstrates effective generalization to real-world scenarios despite being trained exclusively on simulation data (InternData-N1).
- State-of-the-Art Performance: Achieves leading results on various VLN benchmarks, including VLN-CE R2R/RxR and VLN-PE.
- Asynchronous Inference: Supports smooth execution and dynamic obstacle avoidance during navigation tasks.
Good for
- Developing autonomous navigation systems requiring advanced reasoning and instruction following.
- Applications needing robust sim-to-real transfer capabilities for robotics.
- Research and development in Vision-Language Navigation (VLN) and multi-system AI architectures.