QDepth-VLA: Quantized Depth Prediction as Auxiliary Supervision for Vision-Language-Action Models

Yixuan Li, Yuhui Chen, Mingcai Zhou, Haoran Li, Zhengtao Zhang +1 more
10/16/2025
cs.CVcs.RO

No official code found. Best implementation: ucasmichael/QDepth-VLA (9 stars, MIT, last commit 2025-12-18).

Abstract

Spatial perception and reasoning are crucial for Vision-Language-Action (VLA) models to accomplish fine-grained manipulation tasks. However, existing approaches often lack the ability to understand and reason over the essential 3D structures necessary for precise control. To address this limitation, we propose QDepth-VLA, a general framework that augments VLA models with an auxiliary depth prediction task. A dedicated depth expert is designed to predict quantized latent tokens of depth maps obtained from a VQ-VAE encoder, enabling the model to learn depth-aware representations that capture critical geometric cues. Experimental results on the simulation benchmarks and real-world tasks demonstrate that QDepth-VLA yields strong spatial reasoning and competitive performance on manipulation tasks.

View on arXivView PDF

Code Implementations(1)

ucasmichael/QDepth-VLACommunity69%
MIT

QDepth-VLA: Quantized Depth Prediction as Auxiliary Supervision for Vision-Language-Action Models

90Shell, PythonDec 17, 202510 months agoMIT

Similar Papers

DepthVLA: Enhancing Vision-Language-Action Models with Depth-Aware Spatial Reasoning

88% match2025

Perceptio: Perception Enhanced Vision Language Models via Spatial Token Generation

85% match2026

DepthCues: Evaluating Monocular Depth Perception in Large Vision Models

84% match1 repo232026

SG-MIM: Structured Knowledge Guided Efficient Pre-training for Dense Prediction

84% match2024

DFormerv2: Geometry Self-Attention for RGBD Semantic Segmentation

84% match1 repo5072026

Manipulation as in Simulation: Enabling Accurate Geometry Perception in Robots

83% match2025

Cite this paper

@article{li2025qdepthvla,
  title  = {QDepth-VLA: Quantized Depth Prediction as Auxiliary Supervision for Vision-Language-Action Models},
  author = {Yixuan Li and Yuhui Chen and Mingcai Zhou and Haoran Li and Zhengtao Zhang and Dongbin Zhao},
  year   = {2025},
  eprint = {2510.14836},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2510.14836v2}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2510.14836.svg)](https://codeofpaper.com/paper/2510.14836)

Discussion