tensorflow (version 2.16.0)

as_tensor: as_tensor

Description

Coerce objects to tensorflow tensors (potentially of a specific dtype or shape). The provided default methods will call tf$convert_to_tensor. Depending on arguments supplied it may also call some combination of

• tf$saturate_cast or tf$cast

• tf$fill or tf$reshape

Usage

as_tensor(x, dtype = NULL, ..., name = NULL)
# S3 method for default
as_tensor(x, dtype = NULL, ..., shape = NULL, name = NULL)
# S3 method for double
as_tensor(x, dtype = NULL, ..., name = NULL)

Value

a tensorflow tensor

Arguments

x

object to convert

dtype

NULL, a tensorflow dtype (tf$int32), or something coercible to one (e.g. a string "int32")

...,

ignored

name

NULL or a string. Useful for debugging in graph mode, ignored while in eager mode.

shape

an integer vector, tensor, or tf.TensorShape. Can contain up to 1 unspecified dimension, encoded as a -1 or NA. This will reshape x using row-major (C-style) semantics. It will prefer reshaping using non-graph operations if possible, but will otherwise invoke tf$reshape(). If x is a scalar and the requested shape is fully defined or a tensor, the value of x will be recycled to fill a tensor of the requested shape (it will dispatch to tf$fill()).

Examples

if (FALSE) {
as_tensor(42, "int32")
as_tensor(as_tensor(42))
}