LearningRateType enum
A type of learning rate strategies
Constructors
- LearningRateType()
-
const
Values
- decreasingAdaptive → const LearningRateType
- constant → const LearningRateType
-
Learning rate value will be constant throughout the whole fitting process
- timeBased → const LearningRateType
-
Learning rate value will be calculated according to the formula:
where:
is the learning rate value for a new iteration
is the learning rate value from a previous iteration
is the decay parameter
is the iteration step
- stepBased → const LearningRateType
-
Learning rate value will be calculated according to the formula:
where:
is the new learning rate value
is the initial learning rate value
is the decay parameter (in the context of step-based strategy, it describes, how much the learning rate should change at each drop, e.g. 0.5 means a halving)
is the iteration step
is the drop rate (how often the learning rate value should be dropped, r=5 means the rate will be dropped every 5 iterations)
- exponential → const LearningRateType
-
Learning rate value will be calculated according to the formula:
where:
is the new learning rate value
is the initial learning rate value
is the decay parameter
is the iteration step
Properties
Methods
-
noSuchMethod(
Invocation invocation) → dynamic -
Invoked when a nonexistent method or property is accessed.
inherited
-
toString(
) → String -
A string representation of this object.
inherited
Operators
-
operator ==(
Object other) → bool -
The equality operator.
inherited
Constants
-
values
→ const List<
LearningRateType> - A constant List of the values in this enum, in order of their declaration.