[script.evaluate] Fix second-annotator comparisons with a key-resemblance check; Add assertions in engine.evaluator to ensure proper comparisons

[models,configs/models] Cleanup model implementation to re-use all backbones from torchvision; Simplify checkpointer; Remove own implementation of model_zoo; Implement normalization for torchvision-based backbones; Enables pytorch installation on osx

[engine.*trainer] Set non-blocking operation for CPU->GPU data transfers to make communication asynchronous

[script,engine,test] Allows the user to specify the number of steps for threshold/performance analysis

[utils.metric] Rename as measure as some of the values are not real metrics; Add tests for AUC and improve its implementation

[engine.*] Move model summarization to training; Separate static information from dynamic information logging

[engine.trainer] Flush CSV file as we log it (every epoch); Only output minimal information on the screen

[engine.trainer] Implement thorough resource monitoring; Do not plot in the end - writing a CSV is fine already

[script.*] Implement F1-score a priori where possible; Improve plotting of precision-recall curves with better legends, more consistence; Remove redundant plotting code; Update tests