ds.bm.defineLayer: Define training parameters for one RBM layer in a DBM or DBN

In stefan-m-lenz/dsBoltzmannMachinesClient: Client for dsBoltzmannMachines

Description

The call stores the given parameters for training one RBM-layer in a DBM or DBN on the server side in a Julia TrainLayer object. The parameters rbmtype, nhidden, epochs, learningrate/learningrates, categories, batchsize, pcd, cdsteps, startrbm and monitoring of this object are passed to ds.monitored_fitrbm. For a detailed description, see there. Values of NULL indicate that a corresponding default value should be used.

Usage

ds.bm.defineLayer(
  datasources,
  newobj,
  epochs = NULL,
  learningrate = NULL,
  learningrates = NULL,
  sdlearningrate = NULL,
  sdlearningrates = NULL,
  categories = NULL,
  monitoring = NULL,
  rbmtype = NULL,
  nhidden = NULL,
  nvisible = NULL,
  batchsize = NULL,
  pcd = NULL,
  cdsteps = NULL,
  startrbm = NULL
)

Arguments

datasources	A list of Opal object(s) as a handle to the server-side session
newobj	The name of the server-side object where the parameters are stored
nvisible	The number of visible units in the RBM. Only relevant for partitioning. This parameter is derived as much as possible by ds.monitored_stackrbms. For multimodal DBMs with a partitioned first layer, it is necessary to specify the number of visible nodes for all but at most one partition in the input layer.

stefan-m-lenz/dsBoltzmannMachinesClient documentation built on May 2, 2021, 10:46 a.m.