Parameters of the LBFGS algorithm.
More...
◆ Parameter() [1/2]
Constructs the parameters of the LBFGS algorithm
- Parameters
-
context | Context to manage the parameters of the LBFGS algorithm |
◆ Parameter() [2/2]
Constructs the parameter for the LBFGS algorithm
- Parameters
-
context | Context to manage the LBFGS algorithm |
cObject | Pointer to C++ implementation of the parameter |
◆ getBatchIndices()
Gets the numeric table that represents 32 bit integer indices of terms of the objective function.
- Returns
- The numeric table that represents 32 bit integer indices of terms of the objective function
◆ getCorrectionPairBatchIndices()
Gets the numeric table that represents 32 bit integer indices of terms of the objective function.
- Returns
- The numeric table that represents 32 bit integer indices of terms of the objective function
◆ getCorrectionPairBatchSize()
long getCorrectionPairBatchSize |
( |
| ) |
|
Gets the number of observations to compute the sub-sampled Hessian for correction pairs computation
- Returns
- The number of observations to compute the sub-sampled Hessian for correction pairs computation
◆ getL()
Gets the number of iterations between the curvature estimates calculations
- Returns
- The number of iterations between the curvature estimates calculations
◆ getM()
Gets the memory parameter of the LBFGS algorithm. Which is the maximum number of correction pairs that define the approximation of inverse Hessian matrix
- Returns
- The memory parameter of the LBFGS algorithm
◆ getSeed()
- Deprecated:
- This item will be removed in a future release.
Gets the seed for random generation of 32 bit integer indices of terms in the objective function.
- Returns
- The seed for random generation of 32 bit integer indices of terms in the objective function.
◆ getStepLengthSequence()
Gets the numeric table that contains values of the step-length sequence
- Returns
- The numeric table that contains values of the step-length sequence
◆ setBatchIndices()
Sets the numeric table of size nIterations x batchSize that represent indices that will be used instead of random values for the stochastic gradient computations. If no indices are provided, the implementation will generate random indices.
- Parameters
-
batchIndices | The numeric table that represents 32 bit integer indices of terms of the objective function |
◆ setCorrectionPairBatchIndices()
void setCorrectionPairBatchIndices |
( |
NumericTable |
batchIndices | ) |
|
Sets the numeric table of size (nIterations / L) x correctionPairBatchSize that represent indices that will be used instead of random values for the sub-sampled Hessian matrix computations. If no indices are provided, the implementation will generate random indices.
- Parameters
-
batchIndices | The numeric table that represents 32 bit integer indices of terms of the objective function |
◆ setCorrectionPairBatchSize()
void setCorrectionPairBatchSize |
( |
long |
batchSize | ) |
|
Sets the number of observations to compute the sub-sampled Hessian for correction pairs computation
- Parameters
-
batchSize | The number of observations to compute the sub-sampled Hessian for correction pairs computation |
◆ setEngine()
Sets the engine to be used by the algorithm
- Parameters
-
engine | to be used by the algorithm |
◆ setL()
Sets the number of iterations between the curvature estimates calculations
- Parameters
-
L | The number of iterations between the curvature estimates calculations |
◆ setM()
Sets the memory parameter of the LBFGS algorithm. Which is the maximum number of correction pairs that define the approximation of inverse Hessian matrix
- Parameters
-
m | The memory parameter of the LBFGS algorithm |
◆ setSeed()
void setSeed |
( |
long |
seed | ) |
|
- Deprecated:
- This item will be removed in a future release.
Sets the seed for random generation of 32 bit integer indices of terms in the objective function.
- Parameters
-
seed | The seed for random generation of 32 bit integer indices of terms in the objective function. |
◆ setStepLengthSequence()
void setStepLengthSequence |
( |
NumericTable |
stepLengthSequence | ) |
|
Sets the numeric table that contains values of the step-length sequence
- Parameters
-
stepLengthSequence | The numeric table that contains values of the step-length sequence |
The documentation for this class was generated from the following file:
- optimization_solver/lbfgs/Parameter.java