README.md

metadata

title: HEA Query
emoji: 🔬
colorFrom: blue
colorTo: green
sdk: gradio
sdk_version: 5.45.0
app_file: app.py
pinned: true

🔬 HEA Query

This Hugging Face Space allows you to query high-entropy alloys (HEAs) using a combination of:

• FAISS vector database for paper embeddings (semantic search)
• Structured CSV datasets with properties like hardness, bulk modulus, yield strength, etc.
• Large language model (Mistral-7B-Instruct) for intelligent answers

---

Features

• Query 3000+ HEA research papers via FAISS embeddings
• Filter and rank HEA datasets based on numeric properties or phase (fcc, bcc, hcp, etc.)
• Interactive Gradio interface with:
  • LLM answer
  • CSV matches table
  • Paper context (FAISS)

---

How to Use

1. Enter your question about HEAs in the text box.
   Examples:
   • "Which FCC HEAs have hardness > 200 HV?"
   • "List high bulk modulus BCC alloys"
2. Click Submit.
3. View the results:
   • LLM Answer: AI summary based on papers and datasets
   • CSV Matches: Table of filtered alloys
   • Paper Context (FAISS): Text excerpts from research papers

---

Setup (for developers)

1. Clone the Space repo:

   git clone https://huggingface.co/spaces/taradutt007/README.md
   cd <space-name>