An Estimator input_fn for running predict() after evaluate().
evaluation, steps=None, times=None, exogenous_features=None
If the call to evaluate() we are making predictions based on had a batch_size
greater than one, predictions will start after each of these windows
(i.e. will have the same batch dimension).
The dictionary returned by
Estimator.evaluate, with keys
FilteringResults.STATE_TUPLE and FilteringResults.TIMES.
The number of steps to predict (scalar), starting after the
times is specified,
steps must not be; one is required.
A [batch_size x window_size] array of integers (not a Tensor)
indicating times to make predictions for. These times must be after the
corresponding evaluation. If
steps is specified,
times must not be;
one is required. If the batch dimension is omitted, it is assumed to be 1.
Optional dictionary. If specified, indicates exogenous
features for the model to use while making the predictions. Values must
have shape [batch_size x window_size x ...], where
the batch dimension used when creating
steps argument or the
window_size of the
(depending on which was specified).
input_fn suitable for passing to the
predict function of a time
steps are misspecified.