tumourkit.classification.train_graphs.compute_neural_persistence

tumourkit.classification.train_graphs.compute_neural_persistence(model: Module, writer: SummaryWriter | None = None, epoch: int | None = None, log_suffix: str | None = None, use_cubical: bool | None = False) → float

Computes normalized neural persistence of the model using Gudhi. Logs to TensorBoard if writer is provided. Returns the normalized neural persistence score.