Host ZINC datasets on Hugging Face

[NielsRogge]

Hi @chengxiang 🤗

I'm Niels from the Hugging Face open-source team. I came across your ICLR 2025 paper, "Graph Transformers Dream of Electric Flow," on arXiv and noticed the ZINC datasets used in your experiments are hosted on Box.com. We at Hugging Face are working to make research artifacts more discoverable and accessible.

We'd love to know if you'd be interested in hosting your ZINC datasets (ZINC_dgl_clean_20k and ZINC_pytorch_clean_20k) on the Hugging Face Datasets hub. This would significantly improve their visibility and allow researchers to easily access and utilize them via the datasets library using a simple load_dataset call. The Hugging Face team has a new feature called Paper Pages (hf.co/papers), which allows researchers to link their published work to related artifacts hosted on the Hugging Face Hub. If you are one of the authors, you can submit it at https://huggingface.co/papers/submit.

Hosting on Hugging Face offers several benefits including:

• Improved discoverability: Researchers can easily find your dataset through the Hugging Face search functionality and relevant dataset filters.
• Simplified access: Users can directly load your dataset using the Hugging Face datasets library, which provides efficient data loading and processing capabilities.
• Better collaboration: Hosting on Hugging Face encourages community contributions and collaboration around your dataset.

If you're interested, you can find a guide on uploading datasets here: https://huggingface.co/docs/datasets/loading

Let me know if you have any questions or would like to discuss this further.

Best regards,

Niels
ML Engineer @ Hugging Face 🤗