PaddleCFD is a deep learning toolkit for surrogate modeling, equation discovery, shape optimization and flow-control strategy discovery in the field of fluid mechanics. Currently, it mainly supports surrogate modeling, including models based on Fourier Neural Operator (FNO), Transformer, Diffusion Model (DM), Kolmogorov-Arnold Networks (KAN) and DeepONet.
doc: documentationexamples: example scriptsppcfd/data: data-process source codeppcfd/model: model source codeppcfd/utils: utils codesource: source code of paddlepaddle custom operators
# Pull docker image
docker pull ccr-2vdh3abv-pub.cnc.bj.baidubce.com/paddlepaddle/paddle:3.0.0-gpu-cuda11.8-cudnn8.9-trt8.6
# Run docker container
nvidia-docker run --name ppcfd-container -v /home/:/home --network=host -it --shm-size 64g ccr-2vdh3abv-pub.cnc.bj.baidubce.com/paddlepaddle/paddle:3.0.0-gpu-cuda11.8-cudnn8.9-trt8.6 /bin/bash# Clone PaddleCFD
git clone https://github.com/PaddlePaddle/PaddleCFD.git
cd PaddleCFD
# Create conda environment
conda create --name ppcfd python=3.10
conda activate ppcfd
python -m pip install -r requirements.txt -i https://pypi.tuna.tsinghua.edu.cn/simple
python -m pip install paddlepaddle-gpu==3.0.0 -i https://www.paddlepaddle.org.cn/packages/stable/cu118/
# Download and install paddle-backended Open3D
wget https://paddle-org.bj.bcebos.com/paddlecfd/envs/open3d-0.18.0+da239b25-cp310-cp310-manylinux_2_31_x86_64.whl
python -m pip install open3d-0.18.0+da239b25-cp310-cp310-manylinux_2_31_x86_64.whl -i https://pypi.tuna.tsinghua.edu.cn/simple
# Compile customed operator to conda environment
wget -nc https://paddle-org.bj.bcebos.com/paddlescience/cmake-3.23.0-linux-x86_64.tar.gz
tar -zxvf cmake-3.23.0-linux-x86_64.tar.gz
rm -f cmake-3.23.0-linux-x86_64.tar.gz
PATH=$PWD/cmake-3.23.0-linux-x86_64/bin:$PATH
cd source/ppfno_op
python -m pip install --no-build-isolation -v .# Install PaddleCFD from sourcecode at PaddleCFD root directory
python -m pip install -e . -i https://pypi.tuna.tsinghua.edu.cn/simple
# Install PaddleCFD from pypi
python -m pip install ppcfd -i https://pypi.tuna.tsinghua.edu.cn/simple# Run examples
cd PaddleCFD/examples/xxx/xxx
run the example according to the example README.mdFollowing the guidelines of MetaX to run PaddleCFD on MetaX machine.
Join PaddleCFD WeChat group to discuss with us!
PaddleCFD is provided under the Apache-2.0 license