✅ Database of Code Benchmarks

Welcome to my newest initiative:

• Make one-click to run code tutorials for some of the best ML surrogates at the bleeding edge

• Run these on a variety of datasets - fluid mechanics problems, mechanical deformation problems, etc.

• Collect them all into a Google Drive/Google Colab repo

---

Geometry Aware Operator Transformer

• Demonstrating on the AirfRANS dataset — High Fidelity Computational Fluid Dynamics Dataset for Approximating Reynolds-Averaged Navier-Stokes Solutions

Transolver model family

• A Beginner’s Tutorial: the new HiLiftAeroML dataset with code to train a ML surrogate

• This one co-demonstrates “AdaField” with Transolver++ on the DrivAerML dataset

Anchored-Branch Universal Physics Transformer

• Demonstrating on the DrivAer (automotive aerodynamics) dataset

• Training an AB-UPT Point-Cloud Surrogate on HiLiftAeroML

NVIDIA’s DoMINO model

• Demonstrating on the DrivAer (automotive aerodynamics) dataset

• Training DoMINO on Simulated Jet Engine Bracket Designs

“Learning Geometry” Tutorials

• Models that Learn Geometry (Mamba Family)