tumourkit.classification.train_graphs

Script to train and save several GNN configurations.

This program is free software: you can redistribute it and/or modify it under the terms of the GNU Affero General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Affero General Public License for more details.

You should have received a copy of the GNU Affero General Public License along with this program. If not, see <https://www.gnu.org/licenses/>.

Contact information: joseperez2000@hotmail.es

Functions

`append_results`	Appends result to given filename.
`compute_neural_persistence`	Computes normalized neural persistence of the model using Gudhi.
`create_results_file`	Creates header of .csv result file to append results.
`evaluate`	Evaluates model in loader.
`fuse_loaders`	Concatenates both loaders into one.
`generate_configurations`	Generates a grid in the search space with no more than max_confs configurations.
`load_dataset`	Creates Torch dataloaders for training. Folder structure: node_dir: - train - graphs - file1.nodes.csv ... - validation - graphs - file1.nodes.csv ... - test - graphs - file1.nodes.csv .
`load_model`	Available models: GCN, ATT, HATT, SAGE, GIN Configuration space: NUM_LAYERS, DROPOUT, NORM_TYPE
`main`
`main_with_args`
`name_from_conf`	Generates a name from the configuration object.
`save_model`	Save model weights and configuration file to SAVE_DIR
`train`	Train the model with early stopping on either: F1 score (weighted) or neural persistence, or until 1000 iterations.
`train_one_conf`
`train_one_iter`	Trains for one iteration, as the name says.