SpatialTraceGen: High-Fidelity Traces for Efficient VLM Spatial Reasoning Distillation

TL;DR

SpatialTraceGen creates high-quality, step-by-step reasoning datasets for vision-language models, enabling more efficient spatial reasoning training through automated verification and distillation from large models.

Contribution

The paper introduces SpatialTraceGen, a novel framework that distills reasoning traces from large models into high-quality datasets with automated fidelity verification.

Findings

Improves trace quality score by 17%

Reduces variance of trace quality by over 40%

Provides structured reasoning datasets for better fine-tuning

Abstract

While Vision-Language Models (VLMs) excel in many areas, they struggle with complex spatial reasoning, which requires problem decomposition and strategic tool use. Fine-tuning smaller, more deployable models offers an efficient path to strong performance, but this is hampered by a major bottleneck: the absence of high-quality, step-by-step reasoning data. To address this data-efficiency gap, we introduce SpatialTraceGen, a framework to distill the reasoning processes of a large teacher model into a high-quality dataset of multi-hop, multi-tool reasoning traces. A key innovation is our automated Verifier, which scalably ensures the fidelity of each reasoning step, providing a cost-effective alternative to manual human annotation. On the CLEVR-Humans benchmark, this verifier-guided process improves the average quality score of traces by 17\% while reducing quality variance by over 40\%. SpatialTraceGen delivers a dataset of expert traces, providing the structured, step-by-step examples of tool use necessary for effective fine-tuning and sample-efficient offline reinforcement learning.