Hi! I'm Niels from the community science team at Hugging Face.
This PR improves the dataset card for UniRec40M by:
- Adding the `image-to-text` task category and language tags (`en`, `zh`).
- Linking the dataset card to the associated paper: [UniRec-0.1B: Unified Text and Formula Recognition with 0.1B Parameters](https://huggingface.co/papers/2512.21095).
- Adding a sample usage snippet found in the official GitHub repository to help users get started.
- Including relevant tags and the license metadata.