arpes.deep_learning.interpret.InterpretationItem#
- class arpes.deep_learning.interpret.InterpretationItem(target, predicted_target, loss, index, parent_dataloader)[source]#
Provides tools to introspect model performance on a single item.
- __init__(target, predicted_target, loss, index, parent_dataloader)#
Methods
__init__(target, predicted_target, loss, ...)decodes_target(value)Pulls the predicted target backwards through the transformation stack.
show(input_formatter, target_formatter[, ...])Plots item onto the provided axes.
Attributes
datasetFetches the original dataset used to train and containing this item.
targetpredicted_targetlossindexparent_dataloader