Keraflow
Deep Learning for Python.
 All Classes Namespaces Functions Pages
keraflow.callbacks.EarlyStopping Class Reference

Stop training when a monitored quantity has stopped improving. More...

Inheritance diagram for keraflow.callbacks.EarlyStopping:
keraflow.callbacks.Callback

Public Member Functions

def __init__
 

Detailed Description

Stop training when a monitored quantity has stopped improving.

Constructor & Destructor Documentation

def keraflow.callbacks.EarlyStopping.__init__ (   self,
  monitor = 'val_loss',
  patience = 0,
  verbose = 0,
  mode = 'auto' 
)
Parameters
monitorstr. Quantity to monitor.
patiencenumber of epochs with no improvement after which training will be stopped.
verboseverbosity mode, 0 or 1.
modeone of {auto, min, max}. Decides if the monitored quantity improves. If set to max, increase of the quantity indicates improvement, and vice versa. If set to 'auto', behaves like 'max' if monitor contains substring 'acc'. Otherwise, behaves like 'min'.