tumourkit.research_pipe.run_early_topo

tumourkit.research_pipe.run_early_topo(args: Namespace, logger: Logger) → None

It tests the hypothesis that we can use the neural persistence as an early stopping criterion.

This function performs the following steps:

  1. Trains a GNN model with early stopping based on neural persistence (Rips).

  2. Trains a GNN model with early stopping based on neural persistence (Cubical).

  3. Trains a GNN model with early stopping based on validation dataset.

Parameters:

  • args (Namespace) – The arguments for the GNN training.

  • logger (Logger) – The logger object used for logging messages.