SCI Skills

EN Index
Training & EvalMachine Learning & Research AIK-Dense-AI/claude-scientific-skillsModel Training & Evaluation
TO

Torch Geometric

Maintainer K-Dense Inc. · Last updated April 1, 2026

PyTorch Geometric is a library built on PyTorch for developing and training Graph Neural Networks (GNNs). Apply this skill for deep learning on graphs and irregular structures, including mini-batch processing, multi-GPU training, and geometric deep learning applications.

Claude CodeOpenClawNanoClawTrainingEvaluationtorch-geometricmachine-learningpackagemachine learning & deep learning

Original source

K-Dense-AI/claude-scientific-skills

https://github.com/K-Dense-AI/claude-scientific-skills/tree/main/scientific-skills/torch-geometric

Maintainer
K-Dense Inc.
License
MIT license
Last updated
April 1, 2026

Skill Snapshot

Key Details From SKILL.md

2 min

Key Notes

  • PyG is the standard library for Graph Neural Networks built on PyTorch. It provides data structures for graphs, 60+ GNN layer implementations, scalable mini-batch training, and support for heterogeneous graphs.
  • Install: uv add torch_geometric (or uv pip install torch_geometric; requires PyTorch). Optional: pyg-lib, torch-scatter, torch-sparse, torch-cluster for accelerated ops.
  • edge_index = edge_pairs.t().contiguous().
  • For undirected graphs, include both directions: edge (0,1) needs both [0,1] and [1,0] in edge_index.
  • For heterogeneous graphs, use HeteroData — see the Heterogeneous Graphs section below.

Source Doc

Excerpt From SKILL.md

Graph Data: Data and HeteroData

A graph lives in a Data object. The key attributes:

edge_index format is critical: it's a [2, num_edges] tensor where edge_index[0] = source nodes, edge_index[1] = target nodes. It is NOT a list of tuples. If you have edge pairs as rows, transpose and call .contiguous():


## Datasets

PyG bundles many standard datasets that auto-download and preprocess:

```python
from torch_geometric.datasets import Planetoid, TUDataset

## Single-graph node classification (Cora, Citeseer, Pubmed)

dataset = Planetoid(root='./data', name='Cora')
data = dataset[0]  # single graph with train/val/test masks

Use cases

  • Use Torch Geometric in research workflows aligned with this subject area.
  • Follow the upstream documentation for the full working procedure.
  • **Graph-based machine learning**: Node classification, graph classification, link prediction.
  • **Molecular property prediction**: Drug discovery, chemical property prediction.

Not for

  • Do not rely on this catalog entry alone for installation or maintenance details.

Related skills

Related skills

Back to directory
CI
Training & EvalMachine Learning & Research AI

cirq

Cirq is Google Quantum AI's open-source framework for designing, simulating, and running quantum circuits on quantum computers and simulator…

Claude CodeTraining
K-Dense-AI/claude-scientific-skillsView
GT
Training & EvalMachine Learning & Research AI

gtars

Gtars is a high-performance Rust toolkit for manipulating, analy.

Claude CodeOpenClawTraining
K-Dense-AI/claude-scientific-skillsView
PE
Training & EvalMachine Learning & Research AI

PennyLane

PennyLane is a quantum computing library that enables training quantum computers like neural networks. It provides automatic differentiation…

Claude CodeTraining
K-Dense-AI/claude-scientific-skillsView
PU
Training & EvalMachine Learning & Research AI

PufferLib

PufferLib is a high-performance reinforcement learning library designed for fast parallel environment simulation and training. It achieves t…

Claude CodeTraining
K-Dense-AI/claude-scientific-skillsView