Outputs

Short explanation of the different output files. The convention is that the original name of the image is kept and some keywords are appended to it or its extension.

• test_loss_vs_epochs.png: Plot of the loss values as function of the training epochs. Plot made in otrain-train

• test_accuracy_vs_epochs.png: Plot of the Accuracy values as function of the training epochs. Plot made in otrain-train

• model_name.h5: CNN model file to be executed by otrain-infer

• prob_distribution.png: Plot the probability distribution of the cutouts used for the training. Plot made in otrain-plot-results

• model_F1_MCC.png: Plot the f1-mcc curve and save figure in path_plot. Plot made in otrain-plot-results

• mag_ROC.png: Plot the ROC curve corresponding to a certain range of magnitude. Plot made in otrain-plot-results

• errmag_ROC.png: Plot the ROC curve corresponding to a certain range of magnitude uncertainties. Plot made in otrain-plot-results

• mag_recall.png: Plot the precision-recall curve corresponding to a certain range of magnitude. Plot made in otrain-plot-results

• errmag_recall.png: Plot the precision-recall curve corresponding to a certain range of magnitude uncertainties. Plot made in otrain-plot-results

• well_classified: Folder storing 30 cutouts of well-classified candidates. folder created by otrain-train

• misclassified: Folder storing 30 cutouts of misclassified candidates. folder created by otrain-train

• cube_validation.npz: datacube of cutouts used to validate the performance of the CNN. folder created by otrain-train