hugging-science

SUMMARY

What it does

Hugging Science is a curated, LLM-friendly index of scientific datasets, models, blog posts, and interactive demos for ML researchers. It provides a static, parseable catalog at huggingscience.co across 17 scientific domains, exposing llms.txt, llms-full.txt, and per-domain topic files. It also references a Hugging Face organization (huggingface.co/hugging-science) with community-submitted datasets, models, and ~27 interactive Spaces. The skill guides discovery and usage of resources via standard Hugging Face tooling: datasets library, transformers, Inference API, and gradio_client. It includes a workflow for identifying relevant domains, fetching catalog content, selecting resources, using them appropriately, and citing methodology blogs. It emphasizes authentication via HF_TOKEN from .env, caution with trust_remote_code and sending data to Spaces, and handling large scientific datasets via streaming. The skill is part of Scientific Agent Skills by K-Dense and includes bundled scripts and reference files.