Datasets:

---

Qinsi1

/

Vision-Zero-clevr-dataset

like 0

This PR populates the dataset card for the Vision-Zero CLEVR dataset.

It includes:

• A link to the paper (https://huggingface.co/papers/2509.25541).
• A link to the GitHub repository (https://github.com/wangqinsi1/Vision-Zero).
• The paper abstract for full context.
• Relevant task_categories: image-text-to-text and question-answering.
• Additional tags like vlm, self-improvement, gamified-self-play, reinforcement-learning, and clevr.
• A "Sample Usage" section with code snippets for environment setup and training preparation, extracted directly from the GitHub README.
• The BibTeX citation for the paper.