SRCL_cpp_train_network_relu: Function used as part of other functions
In ekstroem/SRCL: Synergistic Risk Contributions Learning

Used as part of other functions.

SRCL_cpp_train_network_relu(
  x,
  y,
  testx,
  testy,
  W1_input,
  B1_input,
  W2_input,
  B2_input,
  IPCW,
  lr = 0.01,
  maxepochs = 100
)

`x`	A matrix of predictors for the training dataset
`y`	A vector of output values for the training data with a length similar to the number of rows of x
`testx`	A matrix of predictors for the test dataset
`testy`	A vector of output values for the test data with a length similar to the number of rows of x
`W1_input`	Input-hidden layer weights
`B1_input`	Biases for the hidden layer
`W2_input`	Hidden-output layer weights
`B2_input`	Bias for the output layer (the baseline risk)
`IPCW`	Inverse probability of censoring weights (Warning: not yet correctly implemented)
`lr`	Initial learning rate
`maxepochs`	The maximum number of epochs