Train PyTorch graph neural networks for molecular and protein property prediction.
npx skills add https://github.com/shushuzn/Rairos --skill torchdrug-shushuzn
Please help me install this Agent Skill. Skill: torchdrug Source: https://github.com/shushuzn/Rairos/tree/main/skills/torchdrug Command: npx skills add https://github.com/shushuzn/Rairos --skill torchdrug-shushuzn
TorchDrug helps you build and train PyTorch-native graph neural network models for molecules, proteins, and biomedical knowledge graphs, so you can predict properties and perform reasoning tasks without stitching together multiple fragile toolchains.
Ask the AI to propose a TorchDrug workflow to train a GIN-based molecular property predictor on BBBP with AUROC/AUPRC metrics, including dataset loading, model configuration, and a scaffold-split training loop.
High-intent search queries and answers about installing and using this skill.
Train graph neural networks for molecular property prediction by defining a YAML workflow that loads curated datasets like BBBP or Tox21, configures a GNN model, and runs a scaffold-split training loop with AUROC and AUPRC metrics.
Model protein structures and sequences for function prediction by applying PyTorch-native sequence models like ESM or structure-aware models like GearNet and SchNet to tackle stability, localization, and function tasks.
Perform link prediction on biomedical knowledge graphs like Hetionet or FB15k-237 by using PyTorch-based embedding and reasoning tasks to complete missing relationships within the graph structure.
PyTorch supports molecular generation and retrosynthesis planning through workflows that generate novel molecules and plan synthetic routes using multi-step retrosynthesis tasks defined via YAML metadata.
Set up dataset splitting and evaluation metrics for graph neural networks by configuring YAML-defined metadata that specifies task, criterion, and metrics setup alongside proper dataset splitting for training workflows.
Use a unified PyTorch toolchain for graph neural networks to predict properties and perform reasoning tasks without stitching together multiple fragile toolchains for molecules, proteins, and biomedical knowledge graphs.
LeonChaoXTrain PyTorch-native graph neural networks on TorchDrug datasets for molecular and protein tasks.
tassiovaleBuild custom GNN architectures for molecular and protein science tasks.
hxk622Develop custom GNN architectures for molecular and protein data using PyTorch.
Zeyuyang-0420Predict molecular properties and model proteins using graph neural networks.
