tf_function
constructs a callable that executes a TensorFlow graph created by tracing the
TensorFlow operations in f
.
This allows the TensorFlow runtime to apply optimizations and exploit parallelism in the
computation defined by f
.
tf_function(
f,
input_signature = NULL,
autograph = FALSE,
experimental_autograph_options = NULL
)
the function to be compiled
A possibly nested sequence of tf$TensorSpec
objects specifying the shapes
and dtypes of the tensors that will be supplied to this function.
If NULL
, a separate function is instantiated for each inferred input signature.
If input_signature
is specified, every input to f
must be a tensor.
Whether autograph should be applied on f
before tracing a graph.
This allows for dynamic control flow (if's, loops etc.) in the traced graph.
See https://www.tensorflow.org/guide/autograph for more information.
Note: We set the default to FALSE
until this functionality is available from R.
Experimental knobs (in the form of a tuple of
tf$autograph$Feature
values) to control behavior when autograph = TRUE.