README.md

metadata

title: POSEIDON Multiscience Demo
emoji: 🔱
colorFrom: blue
colorTo: pink
sdk: gradio
sdk_version: 4.29.0
app_file: app.py
pinned: false

🔱 POSEIDON Playground: Across Scientific Domains 🔱

An interactive Gradio demo exploring how the POSEIDON foundation model for solving Partial Differential Equations (PDEs) could be applied across physics, finance, quantum mechanics, and biology.

Built with love 💖 🔱 for the Hugging Face Community ML Research Engineer take-home assignment.
Inspired by POSEIDON: A Foundation Model for Solving PDEs by CamLab ETH Zürich.
Code from the original repo: github.com/camlab-ethz/poseidon

---

Goal

This app highlights the multidisciplinary potential of pre-trained PDE models like POSEIDON using:

• Intuitive interactive visualizations
• Simple simulations from four real-world domains

🪩 What You Can Do

Feature	Description
✔️ Pick a scientific domain	Finance, Quantum, Fluids, Biology
✔️ Run a mini simulation	See how PDEs behave in each field
✔️ Try POSEIDON inference	Generate predictions from synthetic inputs
✔️ Use real PDE datasets	Compare POSEIDON output vs. ground truth

🚀 Running Locally

1. Clone this repo

git clone https://github.com/YOUR_USERNAME/poseidon_demo.git
cd poseidon_demo

2. (Option A) Using Conda

conda env create -f environment.yml
conda activate poseidon-demo

3. (Option B) Using virtualenv

python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

4. (Run the demo

python -m poseidon_demo.app

🪩 Big Thanks

• Hugging Face for the opportunity and open tools
• ETH Zürich’s CamLab for releasing the POSEIDON repo
• 💖 You, the curious science hacker, for playing with this demo!