This documentation is for Machine Learner 1.1.0. View documentation for the latest release.
||
Skip to end of metadata
Go to start of metadata

 The configurations included in a sample hyper parameters definition JSON file are as follows.

[
    {
        "key": "Iterations",
        "value": "100"
    },
    {
        "key": "Learning_Rate",
        "value": "0.1"
    },
    {
        "key": "SGD_Data_Fraction",
        "value": "1"
    }
]

Parameter definitions

The definitions of the parameters in the above sample hyper parameters definition JSON file are as follows.

Parameter
Description
Required
Default value
Example
Iterations
Number of iterations of gradient descent to run.YesNone

100

Learning_Rate
A constant (usually a small number less than 1.0 ), which is used in machine learning algorithms to control the speed of the learning.  YesNone

0.1

SGD_Data_Fraction
The fraction of data to be used per iteration in the Stochastic Gradient Descent (SGD) optimization algorithms. If this is set to 1.0, the machine learning algorithm runs with Full-batch Gradient Descent optimizer.YesNone

1

  • No labels