Steffen Schotthöfer's Projects
Baseline model and job submission templates for the use-case "Energy" of the AI-HERO hackathon 2022 on energy efficient AI
Charmnet Benchmarking Suite
Dynamical low rank training for neural networks
A lightweight package to explore and employ physics informed neural networks for entropy closures.
A Framework to compare different entropy closures
NeurIPS Large Language Model Efficiency Challenge: 1 LLM + 1GPU + 1Day
My profile readme
My personal website
SU2: An Open-Source Suite for Multiphysics Simulation and Design
Test cases for SU2
Differentiable ODE solvers with full GPU support and O(1)-memory backpropagation with Relaxation Term
Transport-based computational strategies for training neural networks
, a project to fuse deep neural networks with high performance numerical solvers for particle transport. The work is summarized in the ICML2022 contribution, which can be found on arxiv
, and Tucker-DLRT a novel approach for adaptive low-rank neural network training and model compression. The work is summarized in the NeurIPS2022 contribution, which can be found on arxiv.
, an high performance radiation transport simulator and multi purpose numerical toolkit. You can also read the docs. The preprint can be found on arxiv.
, a multi-physics optimization and CFD suite. Check out the official website. My contribution is summarized in the AIAA Journal.
, a deep-learning based prediction of continuum breakdown of rarefied gas flows with a hybrid flow-solver. The preprint can be found on arxiv.
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent