View source on GitHub

Build a scale-and-shift function using a multi-layer neural network.


This will be wrapped in a make_template to ensure the variables are only created once. It takes the d-dimensional input x[0:d] and returns the D-d dimensional outputs loc ('mu') and log_scale ('alpha').

The default template does not support conditioning and will raise an exception if condition_kwargs are passed to it. To use conditioning in Real NVP bijector, implement a conditioned shift/scale template that handles the condition_kwargs.


  • hidden_layers: Python list-like of non-negative integer, scalars indicating the number of units in each hidden layer. Default: `[512, 512].
  • shift_only: Python bool indicating if only the shift term shall be computed (i.e. NICE bijector). Default: False.
  • activation: Activation function (callable). Explicitly setting to None implies a linear activation.
  • name: A name for ops managed by this function. Default: 'real_nvp_default_template'.
  • *args: tf.layers.dense arguments.
  • **kwargs: tf.layers.dense keyword arguments.


  • shift: Float-like Tensor of shift terms ('mu' in [Papamakarios et al. (2016)][1]).
  • log_scale: Float-like Tensor of log(scale) terms ('alpha' in [Papamakarios et al. (2016)][1]).


  • NotImplementedError: if rightmost dimension of inputs is unknown prior to graph execution, or if condition_kwargs is not empty.


[1]: George Papamakarios, Theo Pavlakou, and Iain Murray. Masked Autoregressive Flow for Density Estimation. In Neural Information Processing Systems, 2017.