3dlg-hcvc/revsi

Benchmark and training code for MindCube: spatial mental modeling in vision-language models from limited views.

Code repository for Qwen-3D: A Generalist 3D Vision-Language Model for Spatial Understanding

TempoVista is the official implementation of Dyn-3D, a benchmark and training pipeline for 3D spatial reasoning in vision-language models. The release contains frame-selection tools, SFT and Kinematic-GSPO training code, benchmark evaluation scripts, Dyn-3D annotations, and checkpoints for Qwen3-VL and InternVL3.5.

This is a project about visual spatial reasoning.

[NeurIPS 2026 Reviewed (5/4/4 - A/WA/WA)] 3D-R1: Enhancing Reasoning in 3D VLMs for Unified Scene Understanding

[Awesome-Spatial-VLMs] This repository is the official, community-maintained resource for the survey paper: Spatial Intelligence in Vision-Language Models: A Comprehensive Survey;

Learning Spatial Reasoning through Interaction with the Observable Physical World

[NeurIPS 2026] Official implementation of paper "VLM³: Vision Language Models Are Native 3D Learners".

[CVPR 2026 Highlight] SpatialScore: Towards Comprehensive Evaluation for Spatial Intelligence