Load hypes from the logdir. Namely the modules loaded are: input_file, architecture_file, objective_file, optimizer_file Parameters

Load a JSON file which contains a list of {'raw': 'xy', 'mask': 'z'}. Parameters ---------- json_datafile_path : str Path to

Load all modules from the files specified in hypes. Namely the modules loaded are: input_file, architecture_file, objective_file, optimizer_f

Load hypes from the logdir. Namely the modules loaded are: input_file, architecture_file, objective_file, optimizer_file Parameters

Load all TensorVision plugins.

Load a segmentation mask from an image. The mask is an integer array with shape (height, width). The integer values range from 0 to N-1,

Load the weights of a model stored in saver. Parameters ---------- checkpoint_dir : str The directory of checkpoints. se

Calculate the loss from the logits and the labels. Args: decoded_logits: output of decoder labels: Labels tensor; Output from data_in

Orchestrate the evaluation of a model in the default training dir.

Run main function.

Run main function.

Run main function for continue Training.

Merge two confusion matrices. Parameters ---------- cm1 : dict Confusion matrix which has integer keys 0, ..., nb_classes -

Overlay input_image with a hard segmentation result. Store the result with the same name as segmentation_image, but with `-overlay`.

Add directories to hypes. Parameters ---------- hypes : dict Hyperparameters hypes_fname : str Path to hypes_fil

Set the gpus to use.

Overlay image with propability map. Overlays the image with a colormap ranging from blue to red according to the probability map giv

Start enqueuing threads.

Run one evaluation against the full epoch of data. Parameters ---------- hypes : dict Hyperparameters Returns -----

image orientation

false negatives

Sets up the training Ops. Creates a summarizer to track the loss over time in TensorBoard. Creates an optimizer and applies the gradients to