Datasetpublic

yuanqianhao/Vision-OPD-6K

Vision-OPD-6K: Training Data for Vision-OPD Overview Vision-OPD proposes a regional-to-global self-distillation framework that transfers the model's own privileged regional perception to its full-image policy, without external teacher models, ground-truth labels, reward verifiers, or inference-time tool use. Vision-OPD instantiates two conditional policies from the same MLLM: A crop-conditioned teacher that observes the evidence-centered crop as a privileged… See the full description on the dataset page: https://huggingface.co/datasets/yuanqianhao/Vision-OPD-6K.

sourceHugging Faceapache-2.0updated 4mo agoView on Hugging Face

10likes2.1kdownloads

Dataset Card

Vision-OPD-6K: Training Data for Vision-OPD

Overview

Vision-OPD proposes a regional-to-global self-distillation framework that transfers the model's own privileged regional perception to its full-image policy, without external teacher models, ground-truth labels, reward verifiers, or inference-time tool use.

Vision-OPD instantiates two conditional policies from the same MLLM:

—A crop-conditioned teacher that observes the evidence-centered crop as a privileged input
—A full-image-conditioned student that observes the full image as in standard inference

The student generates on-policy rollouts, and Vision-OPD minimizes token-level divergence between the teacher and student next-token distributions along these rollouts.

This dataset contains 6.2K fine-grained visual question answering samples used for Vision-OPD training. Each sample includes a full image with red bounding box overlay (student input), a cropped region image (teacher input), and the original image with bounding box coordinates.

Dataset Statistics

Attribute	Value
Total Samples	6,241
Image Types	Full image with bbox overlay (student) + cropped region (teacher) + original image

Data Structure & Extraction

1. Extract Student Images

bash

cd images/
cat images.tar.gz* | tar -xvf - -C ./

2. Extract Teacher Images

bash

cd teacher_images/
tar -xvf teacher_images.tar.gz -C ./

3. Extract Original Images (Optional)

bash

cd original_images/
tar -xvf original_images.tar.gz -C ./

Data Format

Each record in train.jsonl contains:

Field	Type	Description
`images`	`list[str]`	Relative path to bbox overlay image (student input)
`teacher_images`	`list[str]`	Relative path to cropped region image (teacher input)
`original_images`	`list[str]`	Relative path to original image (without bbox overlay)
`bbox`	`list[int]`	Bounding box coordinates `[x1, y1, x2, y2]` in the original image
`problem`	`str`	Question prompt with `<image>` token
`answer`	`str`	Ground truth answer
`extra_info`	`dict`	Contains `answer` and `question` (without `<image>` prefix)

Intended Use

—On-policy self-distillation training of MLLMs for fine-grained visual understanding
—Research on regional-to-global perception transfer without inference-time tool use

Related Resources

Resource	Link
Paper	arXiv:2605.18740
Code	GitHub: VisionOPD/Vision-OPD

Citation

bibtex

@article{yuan2026vision,
  title={Vision-OPD: Learning to See Fine Details for Multimodal LLMs via On-Policy Self-Distillation},
  author={Yuan, Qianhao and Lou, Jie and Yu, Xing and Lin, Hongyu and Sun, Le and Han, Xianpei and Lu, Yaojie},
  journal={arXiv preprint arXiv:2605.18740},
  year={2026}
}

License

Apache-2.0 License