FastSurferCNN.utils.misc¶

FastSurferCNN.utils.misc.check_path(path)[source]¶: Create path.

FastSurferCNN.utils.misc.find_latest_experiment(path)[source]¶

Find and load latest experiment.

Parameters:

pathstr: Path to the latest experiment.

Returns:

int: Latest experiments.

FastSurferCNN.utils.misc.plot_confusion_matrix(cm, classes, title='Confusion matrix', cmap=<matplotlib.colors.LinearSegmentedColormap object>, file_save_name='temp.pdf')[source]¶

Plot the confusion matrix.

Parameters:

cmnpt.NDArray: Confusion matrix.
classesList[str]: List of class names.
titlestr: (Default value = “Confusion matrix”).
cmapplt.cm.ColormapRegistry: Colour map (Default value = plt.cm.Blues).
file_save_namestr: (Default value = “temp.pdf”).

Returns:

figmatplotlib.figure.Figure: Matplotlib Figure object with the confusion matrix plot.

FastSurferCNN.utils.misc.plot_predictions(images_batch, labels_batch, batch_output, plt_title, file_save_name)[source]¶

Plot predictions from validation set.

Parameters:

images_batchtorch.Tensor: Batch of images.
labels_batchtorch.Tensor: Batch of labels.
batch_outputtorch.Tensor: Batch of output.
plt_titlestr: Plot title.
file_save_namestr: Name the plot should be saved tp.

FastSurferCNN.utils.misc.update_num_steps(dataloader, cfg)[source]¶

Update the number of steps.

Parameters:

dataloaderFastSurferCNN.data_loader.loader.DataLoader: The dataloader object that contains the training data.
cfgyacs.config.CfgNode: The configuration object that contains the training configuration.