[Paper] [Slide] [Demo]

This is the official implementation of "MPerformer: An SE(3) Transformer-based Molecular Perceptron"

MPerformer is a universal learning-based molecular perception method to construct 3D molecules with complete chemical information purely based on molecular 3D atom clouds.

An illustration of MPerformer and its learning paradigm.

• Uni-Core, you can check its Installation Documentation.

• pip install rdkit-pypi==2021.9.4
  pip install dpdata

• Please download the checkpoint and place it to the fold ./weight
• Given XYZ file/fold, you can use the following command to get the corresponding SDF file

  XYZ_PATH='PATH OF YOUR XYZ FILE/FOLD'
  SDF_PATH='PATH TO SAVE SDF FILE'
  python predict.py --filename $XYZ_PATH --outputs_path $SDF_PATH

You can also try MPerformer online by clicking on this link

If our work can help you, please cite it

@inproceedings{wang2023mperformer,
  title={MPerformer: An SE (3) Transformer-based Molecular Perceptron},
  author={Wang, Fanmeng and Xu, Hongteng and Chen, Xi and Lu, Shuqi and Deng, Yuqing and Huang, Wenbing},
  booktitle={Proceedings of the 32nd ACM International Conference on Information and Knowledge Management},
  pages={2512--2522},
  year={2023}
}

This code is built upon Uni-Mol and Uni-Core. Thanks for their contribution.