CASTOR documentation

CASTOR is a transformer-based pipeline for gravitational-wave detection, training, continuous-data inference, and sensitivity evaluation.

Note

CASTOR is under active development. Interfaces may change before the first stable release.

Installation

Choose a PyTorch build and install CASTOR and its optional scientific dependencies.

Installation

Command-line tools

Start with the castor-train and castor-eval workflows.

Quick start

Reproducibility

Training checkpoints record configuration, dataset identities, random number generator states, optimizer state, and scheduler state.

Reproducibility and provenance

API reference

Browse documentation generated from the packaged Python API.

castor_gw