OneScience-Group/AURORA

AURORA

Model Introduction

AURORA is a foundation model of the Earth system developed by Microsoft Research. It addresses a range of Earth system prediction tasks, including global weather forecasting and air pollution prediction. The paper was published at ICML 2024.

Paper: Aurora: A Foundation Model of the Atmosphere

https://arxiv.org/abs/2405.13063

Model Description

Aurora is a deep learning model with 1.3 billion parameters, composed of a 3D Perceiver encoder, a 3D Swin Transformer processor, and a 3D Perceiver decoder.

Use Cases

Scenario	Description
Weather Forecast Training	Train AURORA using ERA5 HDF5 data
Local Quick Validation	Use synthetic data to verify data loading, model training, fine-tuning, inference, and inference result visualization.
ModelScope / OneCode Execution	Download as a standalone model package, install dependencies, and run scripts directly.
Multi-GPU Training	Launch multi-process training via torchrun.

Usage Guide

1. OneCode Usage

Experience intelligent one-click AI4S programming through the OneCode online environment:

Click to Experience Intelligent One-Click AI4S Programming

2. Manual Installation and Usage

Hardware Requirements

• A GPU or DCU is recommended.
• CPU can be used for import and small-scale connectivity verification; full training and inference will be slow.
• DCU users must install DTK in advance. DTK 25.04.2 or above, or the OneScience recommended version matching your cluster, is recommended.

Download the Model Package

hf download OneScience-Group/AURORA --local-dir ./AURORA
cd AURORA

Install the Runtime Environment

DCU Environment

# Please activate DTK and CONDA first
conda create -n onescience311 python=3.11 -y
conda activate onescience311
# uv installation is supported
pip install onescience[earth-dcu] -i http://mirrors.onescience.ai:3141/pypi/simple/  --trusted-host mirrors.onescience.ai

GPU Environment

# Please activate CONDA first
conda create -n onescience311 python=3.11 -y libstdcxx-ng=12 libgcc-ng=12 gcc_linux-64=12 gxx_linux-64=12
conda activate onescience311
# uv installation is supported
pip install onescience[earth-gpu] -i http://mirrors.onescience.ai:3141/pypi/simple/  --trusted-host mirrors.onescience.ai

Training Data Introduction

The OneScience community provides ERA5 data for training (due to file size limits, the current repository contains a slice of the full dataset). Users can download it with the command below and confirm that the data path in conf/config.yaml is set correctly:

hf download --repo-type dataset OneScience-Group/ERA5 --local-dir ./data

Generate Synthetic Data for Pipeline Validation

Synthetic data is only used to verify the data protocol and end-to-end pipeline; it does not represent forecast quality:

python scripts/fake_data.py

Training

Single GPU:

python scripts/train.py

Multi-GPU:

The command below launches 2 training processes on one machine, each using a single device.

torchrun --nproc_per_node=2 --nnodes=1 --rdzv_id=1000 --rdzv_backend=c10d --max_restarts=0 --master_addr="localhost" --master_port=29500 scripts/train.py

Fine-tuning

The configuration points training.finetune.checkpoint to data/checkpoint/model_bak.pt produced during training, so fine-tuning uses the trained model by default.

python scripts/finetune.py

Training Weights

This repository provides weights trained on ERA5 reanalysis data in the weights/ folder. The weight files will be uploaded soon and are expected to be available in the near future.

Inference

Inference reads data/checkpoint/model_finetune.pt saved by fine-tuning by default.

python scripts/inference.py

Evaluation and Visualization

python scripts/result.py

OneScience Official Information

Platform	OneScience Main Repository	Skills Repository
Gitee	https://gitee.com/onescience-ai/onescience	https://gitee.com/onescience-ai/oneskills
GitHub	https://github.com/onescience-ai/OneScience	https://github.com/onescience-ai/oneskills

Citation & License

• Aurora paper: https://arxiv.org/abs/2405.13063.
• This repository is the OneScience reproduction of the original Aurora paper. For citation or commercial use, please contact AIWeatherClimate@microsoft.com by email; see the official requirements for details: https://microsoft.github.io/aurora/intro.html.
• Copyright (c) Microsoft Corporation. Licensed under the MIT license.