• OmniArch: Building the Foundation Model for Scientific Computing
• • Abstract
  • Quick Start
  • Pretrained Weights
  • Data

OmniArch: Building the Foundation Model for Scientific Computing

This is the Repo of paper OmniArch: Building the Foundation Model for Scientific Computing.

Abstract

Foundation models have revolutionized language modeling, while whether this success is replicated in scientific computing remains unexplored. We present OmniArch, the first prototype aiming at solving multi-scale and multi-physics scientific computing problems with physical alignment. We addressed all three challenges with one unified architecture. Its pre-training stage contains a Fourier Encoder-decoder fading out the disharmony across separated dimensions and a Transformer backbone integrating quantities through temporal dynamics, and the novel PDE-Aligner performs physics-informed fine-tuning under flexible conditions. As far as we know, we first conduct 1D-2D-3D united pre-training on the PDEBench, and it sets not only new performance benchmarks for 1D, 2D, and 3D PDEs but also demonstrates exceptional adaptability to new physics via in-context and zero-shot learning approaches, which supports realistic engineering applications and foresight physics discovery.

Quick Start

1. Choose the right environment

mindspore_2_2_cann7_cb2

2. Install the Mindformers PDE Version

git clone https://openi.pcl.ac.cn/cty315/mindformers.git
cd mindformers
pip uninstall mindformers
bash build.sh

3. Start training

Single Device Training:

python train_single.py

Distributed Training:

bash run_dynamic_cluster.sh

Pretrained Weights

Model	Params.	Size	Huawei-Cloud Link	Google Drive
OmniArch-PT-Torch	672.85M	7.9GB	ckpt	ckpt
OmniArch-PT-MindSpore	672.85M	7.9GB	ckpt	ckpt

Data

Please Download the data from PDEbench abd PDEArena.