# Text-Only Data Synthesis for Vision Language Model Training

arXiv: https://arxiv.org/abs/2503.22655 · HTML: https://mlsanity.com/paper/2503.22655/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Placed in VLM benchmarks and visual reasoning, one of 16 directions in Multimodal LLMs / vision-language models.
- Builds on 17 of its 22 references that are in the map.
- First posted on arXiv 28 Mar 2025.

## Placed in

- Multimodal LLMs / vision-language models / [VLM benchmarks and visual reasoning](https://mlsanity.com/mllm/vlm-benchmarks-and-visual-reasoning/)
- Synthetic data and self-improvement for LLMs / [Self-improving MLLMs and synthetic data](https://mlsanity.com/synth/self-improving-mllms-and-synthetic-data/)

## Builds on (most cited first, 17 of 17)

- [Microsoft COCO: Common Objects in Context](https://mlsanity.com/paper/1405.0312/) (2014)
- [Improved Baselines with Visual Instruction Tuning](https://mlsanity.com/paper/2310.03744/) (2023)
- [MMMU: A Massive Multi-discipline Multimodal Understanding and Reasoning Benchmark for Expert AGI](https://mlsanity.com/paper/2311.16502/) (2023)
- [Learn to Explain: Multimodal Reasoning via Thought Chains for Science Question Answering](https://mlsanity.com/paper/2209.09513/) (2022)
- [MME: A Comprehensive Evaluation Benchmark for Multimodal Large Language Models](https://mlsanity.com/paper/2306.13394/) (2023)
- [Flickr30k Entities: Collecting Region-to-Phrase Correspondences for Richer Image-to-Sentence Models](https://mlsanity.com/paper/1505.04870/) (2015)
- [ShareGPT4V: Improving Large Multi-Modal Models with Better Captions](https://mlsanity.com/paper/2311.12793/) (2023)
- [MMMU-Pro: A More Robust Multi-discipline Multimodal Understanding Benchmark](https://mlsanity.com/paper/2409.02813/) (2024)
- [LogicVista: Multimodal LLM Logical Reasoning Benchmark in Visual Contexts](https://mlsanity.com/paper/2407.04973/) (2024)
- [ZeroCap: Zero-Shot Image-to-Text Generation for Visual-Semantic Arithmetic](https://mlsanity.com/paper/2111.14447/) (2021)
- [Efficient Multimodal Learning from Data-centric Perspective](https://mlsanity.com/paper/2402.11530/) (2024)
- [The All-Seeing Project V2: Towards General Relation Comprehension of the Open World](https://mlsanity.com/paper/2402.19474/) (2024)
- [DenseFusion-1M: Merging Vision Experts for Comprehensive Multimodal Perception](https://mlsanity.com/paper/2407.08303/) (2024)
- [Connect, Collapse, Corrupt: Learning Cross-Modal Tasks with Uni-Modal Data](https://mlsanity.com/paper/2401.08567/) (2024)
- [I Can't Believe There's No Images! Learning Visual Tasks Using only Language Supervision](https://mlsanity.com/paper/2211.09778/) (2022)
- [LLM2CLIP: Powerful Language Model Unlocks Richer Cross-Modality Representation](https://mlsanity.com/paper/2411.04997/) (2024)
- [ArcSin: Adaptive ranged cosine Similarity injected noise for Language-Driven Visual Tasks](https://mlsanity.com/paper/2402.17298/) (2024)

## Built on by (most cited first, 3 of 3)

- [Anisotropic Modality Align](https://mlsanity.com/paper/2605.07825/) (2026)
- [CARE: Competence-Aware Reward Shaping for Adaptive Reasoning Length in Video-MLLMs](https://mlsanity.com/paper/2606.19927/) (2026)
- [Text-Centric Post-Training for Omni-Modal Reasoning](https://mlsanity.com/paper/2610.02819/) (2026)

## Abstract (arXiv)

Training vision-language models (VLMs) typically requires large-scale, high-quality image-text pairs, but collecting or synthesizing such data is costly. In contrast, text data is abundant and inexpensive, prompting the question: can high-quality multimodal training data be synthesized purely from text? To tackle this, we propose a cross-integrated three-stage multimodal data synthesis framework, which generates two datasets: Unicorn-1.2M and Unicorn-471K-Instruction. In Stage 1: Diverse Caption Data Synthesis, we construct 1.2M semantically diverse high-quality captions by expanding sparse caption seeds using large language models (LLMs). In Stage 2: Instruction-Tuning Data Generation, we further process 471K captions into multi-turn instruction-tuning tasks to support complex reasoning. Finally, in Stage 3: Modality Representation Transfer, these textual captions representations are transformed into visual representations, resulting in diverse synthetic image representations. This three-stage process enables us to construct Unicorn-1.2M for pretraining and Unicorn-471K-Instruction for instruction-tuning, without relying on real images. By eliminating the dependency on real images while maintaining data quality and diversity, our framework offers a cost-effective and scalable solution for VLMs training.