Dynam3D

Official implementation of "Dynam3D: Dynamic Layered 3D Tokens Empower VLM for Vision-and-Language Navigation" (NeurIPS'25 Oral)

Features

Navigation Models - Uses dynamic 3D tokens to empower vision-language models for navigation.

InternRobotics' open platform for building generalized navigation foundation models.

Learned Perceptive Forward Dynamics Model

CoRL 2025 Search-TTA: A Multimodal Test-Time Adaptation Framework for Visual Search in the Wild

CoRL 2025 GC-VLN: Instruction as Graph Constraints for Training-free Vision-and-Language Navigation

InternRobotics' open platform for building generalized navigation foundation models.

Learned Perceptive Forward Dynamics Model

CoRL 2025 Search-TTA: A Multimodal Test-Time Adaptation Framework for Visual Search in the Wild

CoRL 2025 GC-VLN: Instruction as Graph Constraints for Training-free Vision-and-Language Navigation