Release RICE Billion-scale Candidate Region Dataset on Hugging Face

[NielsRogge]

Hi @anxiangsir 🤗

I'm Niels and work as part of the open-source team at Hugging Face. I discovered your work through Hugging Face's daily papers as yours got featured: https://huggingface.co/papers/2507.20025.
The paper page lets people discuss about your paper and lets them find artifacts about it (your dataset for instance),
you can also claim the paper as yours which will show up on your public profile at HF, add Github and project page URLs.

It's fantastic to see that the pre-trained RICE and MLCD models are already available on Hugging Face!

I noticed that your paper also introduces a "billion-scale candidate region dataset" which is crucial for enhancing region-level visual and OCR capabilities. Would you be interested in hosting this new dataset on https://huggingface.co/datasets as well?
Hosting it on Hugging Face will give you more visibility and enable better discoverability for the data, allowing people to easily use it:

from datasets import load_dataset

dataset = load_dataset("your-hf-org-or-username/your-dataset")

If you're interested, here's a guide to get started: https://huggingface.co/docs/datasets/loading. We also support Webdataset, which is useful for image/video datasets: https://huggingface.co/docs/datasets/en/loading#webdataset.

Additionally, there's the dataset viewer which allows people to quickly explore the first few rows of the data directly in the browser. After it's uploaded, we can also link the dataset to the paper page (read here) so people can discover your work more easily.

Let me know if you're interested or need any guidance!

Kind regards,

Niels

[anxiangsir]

Hi @NielsRogge ,

Thank you so much for reaching out and for featuring our paper in Hugging Face's daily papers! It's wonderful to see the recognition and support from the open-source community at Hugging Face.

We really appreciate your interest in hosting our billion-scale candidate region dataset on Hugging Face. Unfortunately, due to some data cleaning processes we performed, there are missing URLs in our current dataset, which prevents us from immediately releasing the original dataset in its complete form. This will require some additional time to resolve.

However, I'm excited to let you know that we will be releasing a new multimodal large-scale model, RiceVL, in a few weeks, which will include:

• All pre-training datasets
• Instruction fine-tuning datasets
• Pre-training logs
• Pre-trained models

Once RiceVL is ready for release, we would love to reach out to you for guidance on how to best host and showcase our work on Hugging Face. Your platform's visibility and discoverability features would be incredibly valuable for the community to access and utilize our datasets and models.

We'll definitely contact you when we're ready to proceed with the release. Thank you again for your support and for making it easier for researchers to share and discover important work!

[NielsRogge]

Thanks for your reply! It looks like https://huggingface.co/papers/2507.20025 has the model linked which is nice.