Watch keynotes, product sessions, workshops, and more from Google I/O See playlist


Creates a tfl.layers.RTL layer.

This function expects that all features defined in model_config.feature_configs are used and present in calibration_outputs.

calibration_outputs A mapping from feature name to calibration output.
model_config Model configuration object describing model architecture. Should be one of the model configs in tfl.configs.
submodel_index Corresponding index into submodels.
average_outputs Whether to average the outputs of this layer.
dtype dtype

A tfl.layers.RTL instance.

ValueError If model_config.parameterization is not one of 'all_vertices' or 'kronecker_factored'.