Build, train, and deploy graph neural network workflows for molecules and proteins.
npx skills add https://github.com/Hung-3008/agusta --skill torchdrug-hung-3008
Please help me install this Agent Skill. Skill: torchdrug Source: https://github.com/Hung-3008/agusta/tree/main/.agents/skills/torchdrug Command: npx skills add https://github.com/Hung-3008/agusta --skill torchdrug-hung-3008
TorchDrug provides a unified platform to design, train, and deploy graph neural networks for chemistry and biology, enabling researchers to build advanced models without assembling disparate tools.
Load a molecular dataset, initialize a GIN model, wrap it in a PropertyPrediction task, and train with a scaffold split.
High-intent search queries and answers about installing and using this skill.
To perform molecular property prediction, load a dataset like BBBP, initialize a GNN architecture such as GIN or GCN, wrap it in a PropertyPrediction task, and train the model using an end-to-end pipeline with a scaffold split.
Protein modeling is supported by multiple graph neural network architectures, including GIN, GCN, GAT, and GearNet, with specific extensions available for handling 3D molecular structures.
Yes, knowledge graph reasoning is fully supported for drug discovery workflows, allowing researchers to perform complex biomedical reasoning using datasets like Hetionet directly within the training pipeline.
Retrosynthesis planning is executed through dedicated workflows that leverage graph neural networks, providing modular components for data handling and model training to support end-to-end chemical reaction prediction.
Yes, transfer learning is supported across 40+ datasets like QM9 and ZINC, enabling researchers to pre-train models on specific tasks and transfer learned representations to new biomedical predictions.
hxk622Develop custom GNN architectures for molecular and protein data using PyTorch.
jaechang-hitsTrain GNN models for molecular property prediction using PyTorch and TorchDrug.
jackspaceApply graph neural networks to drug discovery tasks with PyTorch.
sanand0Predict molecular properties with graph neural networks using PyTorch.
