VLM3: Vision Language Models Are Native 3D Learners

Hugging Face Daily Papers Papers
vision-language-models 3d-understanding depth-estimation text-based-training scaling pixel-correspondence camera-pose-estimation

Summary

This paper proposes VLM3, a method that adapts vision language models for 3D understanding tasks through simple architectural modifications and text-based training, achieving competitive performance without complex designs. It demonstrates significant improvements in depth estimation accuracy and enables diverse 3D tasks like pixel correspondence, camera pose estimation, and object-level understanding.

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.
Original Article
View Cached Full Text

Cached at: 06/01/26, 03:18 AM

Paper page - VLM3: Vision Language Models Are Native 3D Learners

Source: https://huggingface.co/papers/2605.30561

Abstract

Vision Language Models can be adapted for 3D understanding tasks through simple architectural modifications and text-based training, achieving performance comparable to specialized vision models without requiring complex designs or extensive data augmentation.

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 understandingstill 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 referenceand 3)data mixtureandscaling, 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 VLMdepth estimationaccuracy by a large margin (0.84 -> 0.9), but also enables diverse 3D tasks such aspixel correspondence,camera pose estimationandobject-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.

View arXiv pageView PDFGitHub1Add to collection

Get this paper in your agent:

hf papers read 2605\.30561

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2605.30561 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2605.30561 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2605.30561 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles