mxLSTMsetup: mxLSTMsetup
In MarkusBonsch/mxLSTM: LSTM model for regression

initialize the weights and input matrices with random number to create an executable model

1 2	mxLSTMsetup(model, num.features, num.outputs, num.hidden, optimizeFullSequence, seq.length, batch.size, initializer, initialModel = NULL, gammaInit)

`model`	mxSymbol as returned by mxLSTMcreate
`num.features`	number of input features.
`num.outputs`	number of target variables
`num.hidden`	see `mxLSTM`
`optimizeFullSequence`	see `mxLSTM`
`seq.length`	see `mxLSTM`
`batch.size`	see `mxLSTM`
`initializer`	see `mxLSTM`
`initialModel`	see `mxLSTM`
`gammaInit`	see `mxLSTM`