Search papers, labs, and topics across Lattice.
VLM3 demonstrates that Vision Language Models (VLMs) can achieve strong 3D understanding without complex task-specific designs or architectural changes. The key ingredients are focal length unification, text-based pixel referencing, and careful data mixture/scaling. VLM3 significantly improves depth estimation accuracy and matches the performance of expert vision models on tasks like pixel correspondence and camera pose estimation, while using standard VLM architectures and text-based training.
Forget complex architectures and task-specific designs: VLMs are already native 3D learners with the right training recipe.
Vision Language Models (VLMs) enable a unified model to solve various vision tasks through prompting. They have shown promising performance in semantic understanding. However, 3D understanding still largely relies on expert vision models with complex task-specific designs. The key argument this work wants to make is that VLMs are native 3D learners. Our in-depth large scale study shows that 1) focal length unification, 2) text-based pixel reference and 3) data mixture and scaling, are all you need for effective 3D learning. Model architecture changes, large models, heavy data augmentations, and complex losses including the regression formulation, many of which form the foundation of expert vision models, are actually not necessary conditions. As a result, we propose VLM3, a scalable method with the simplest design that enables standard VLMs to master diverse 3D tasks. VLM3 not only advances the VLM depth estimation accuracy by a large margin (0.84 ->0.9), but also enables diverse 3D tasks such as pixel correspondence, camera pose estimation and object-level 3D understanding, matching expert vision model accuracy while maintaining standard architectures and text-based training. We believe VLM3 opens up a new paradigm for simple and scalable 3D learning.
