Great — here’s exactly where to put your README.md so that Hugging Face uses it as your dataset card.
📌 Where to put README.md for Hugging Face Datasets
Hugging Face treats the root-level README.md of your dataset repository as the Dataset Card.
This means:
✅ Place README.md in the root directory of your LDData folder, e.g.:
LDData/
├── README.md ← put it here
├── generating-vectors/
├── generating-matrices/
├── examples/
├── some-file.txt
└── upload.py
Your local folder should look like the above before running:
python upload.py --repo-id QMCSoftware/LDData
The upload_folder API automatically uploads that file as:
https://huggingface.co/datasets/QMCSoftware/LDData/blob/main/README.md
…which becomes the dataset’s front-page description.
What if you already have a README.md?
Perfect — the existing one can be used as the dataset card, or I can help you rewrite it in HF Dataset Card format with recommended sections:
Hugging Face Dataset Card Template
I can create a polished README.md using the structure:
- Title + Summary
- Dataset Description
- Motivation & Citation
- Data Structure
- File Format Details
- How to Load with
datasets - Links to your paper (ArXiv 2502.14256)
- Benchmarks or usage examples
- License
- Acknowledgements
Similar to:
- Meta Omnilingual ASR Corpus
- NVIDIA PhysicalAI
- Moondream RefCOCO-m
📌 Just say:
“Generate a dataset card README.md for LDData.”
…and I’ll do the whole draft for you.
If you want, I can generate:
✔️ A polished dataset card
✔️ A lightweight minimal one
✔️ Or a fully detailed professional one
Which style do you want?