hugging-science

Curate scientific AI datasets, models, and demos from Hugging Face.

4|Updated Mar 2, 2026
One-click install
npx skills add https://github.com/shushuzn/Rairos --skill hugging-science-shushuzn
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: hugging-science
Source: https://github.com/shushuzn/Rairos/tree/main/skills/hugging-science
Command: npx skills add https://github.com/shushuzn/Rairos --skill hugging-science-shushuzn

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

It helps users discover high-signal, curated scientific AI resources (datasets, models, and interactive demos) and then use them correctly for domain-specific research and experimentation.

Core Features & Use Cases

  • Curated scientific catalog for LLM agents: Quickly locate open or gated resources across 17 scientific topics using the Hugging Science index (including llms-full.txt for full coverage).
  • Domain-aligned discovery with actionable pointers: Match the user’s scientific ML need (e.g., proteins, DNA/genomics, climate/earth science, materials, drug discovery, theorem proving) to the right entries and Hugging Face links.
  • Practical execution paths: Load datasets with datasets, run models with transformers or HF Inference API/Providers, and call Gradio Spaces via gradio_client (including common scientific “gotchas” like trust_remote_code and schema inspection).
  • Methodology-aware guidance: Use catalog blog-post entries as citations for training/preprocessing methodology when available.

Quick Start

Ask for a curated dataset and runnable approach for a specific scientific task (for example, “Find me a protein model for mutation effect prediction and tell me how to load it with transformers”).

Frequently Asked Questions about hugging-science

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I find curated scientific ML datasets and models on Hugging Face?

You can find curated scientific ML datasets and models by querying a structured catalog that matches your domain to high-signal Hugging Face resources across 17 scientific topics. It provides actionable pointers to relevant datasets, models, and interactive demos.

How do I load a protein model for mutation effect prediction with transformers?

To load a protein model for mutation effect prediction with transformers, locate the model via the scientific catalog, then execute it using the transformers library while handling common scientific gotchas like trust_remote_code.

Can I run interactive climate modeling demos using gradio_client?

Yes, you can run interactive climate modeling demos using gradio_client. The catalog identifies relevant Gradio Spaces for scientific tasks and provides correct execution paths to call them remotely.

Does Hugging Face support resources for genomics and materials science workflows?

Hugging Face supports genomics and materials science workflows through hosted datasets and models. The catalog indexes these resources across biology, chemistry, and physics to help you locate and operationalize them.

What's the best way to access Hugging Face datasets for scientific machine learning?

The best way to access Hugging Face datasets for scientific machine learning is through a curated index that maps your specific domain need to the right dataset entries. You then load the data using the datasets library.

Why do I need trust_remote_code when running scientific models with transformers?

You need trust_remote_code when running scientific models with transformers because many domain-specific architectures require custom modeling code. The catalog highlights this requirement as a common gotcha for scientific execution.