Schedules
|
Base class for all the adapters |
|
This adapter keep the current value equals to the initial_value it's mainly used to have an uniform interface when defining a parameter as constant vs as an adapter |
|
Adapts the initial value towards the end value using an exponential "decay" function |
|
Adapts the initial value towards the end value using a "decay" function of the form 1/x |
|
Adapts the initial value towards the end value using a potential "decay" function |
- class sklearn_genetic.schedules.base.BaseAdapter(initial_value, end_value, adaptive_rate, **kwargs)[source]
Base class for all the adapters
- Parameters:
- initial_valuefloat,
Initial value to be adapted
- end_valuefloat,
The final (asymptotic) value that the initial_value can take
- adaptive_ratefloat,
Controls how fast the initial_value approaches the end_value
- kwargsdict,
Possible extra parameters, None for now
- Attributes:
- current_stepint,
The current number of iterations that the adapter has run
- current_valuefloat,
The transformed initial_value after current_steps changes
- class sklearn_genetic.schedules.ConstantAdapter(initial_value, end_value, adaptive_rate)[source]
This adapter keep the current value equals to the initial_value it’s mainly used to have an uniform interface when defining a parameter as constant vs as an adapter
- Parameters:
- initial_valuefloat,
Initial value to be adapted
- end_valuefloat,
The final (asymptotic) value that the initial_value can take
- adaptive_ratefloat,
Controls how fast the initial_value approaches the end_value
- Attributes:
- current_stepint,
The current number of iterations that the adapter has run
- current_valuefloat,
Same as the initial_value
- class sklearn_genetic.schedules.ExponentialAdapter(initial_value, end_value, adaptive_rate)[source]
Adapts the initial value towards the end value using an exponential “decay” function
- Parameters:
- initial_valuefloat,
Initial value to be adapted
- end_valuefloat,
The final (asymptotic) value that the initial_value can take
- adaptive_ratefloat,
Controls how fast the initial_value approaches the end_value
- Attributes:
- current_stepint,
The current number of iterations that the adapter has run
- current_valuefloat,
The transformed initial_value after current_steps changes
- class sklearn_genetic.schedules.InverseAdapter(initial_value, end_value, adaptive_rate)[source]
Adapts the initial value towards the end value using a “decay” function of the form 1/x
- Parameters:
- initial_valuefloat,
Initial value to be adapted
- end_valuefloat,
The final (asymptotic) value that the initial_value can take
- adaptive_ratefloat,
Controls how fast the initial_value approaches the end_value
- Attributes:
- current_stepint,
The current number of iterations that the adapter has run
- current_valuefloat,
The transformed initial_value after current_steps changes
- class sklearn_genetic.schedules.PotentialAdapter(initial_value, end_value, adaptive_rate)[source]
Adapts the initial value towards the end value using a potential “decay” function
- Parameters:
- initial_valuefloat,
Initial value to be adapted
- end_valuefloat,
The final (asymptotic) value that the initial_value can take
- adaptive_ratefloat,
Controls how fast the initial_value approaches the end_value
- Attributes:
- current_stepint,
The current number of iterations that the adapter has run
- current_valuefloat,
The transformed initial_value after current_steps changes