tf.autograph.experimental.set_loop_options

View source

Specifies additional arguments to be passed to the enclosing while_loop.

Aliases:

  • tf.compat.v1.autograph.experimental.set_loop_options
  • tf.compat.v2.autograph.experimental.set_loop_options
tf.autograph.experimental.set_loop_options(
    parallel_iterations=UNSPECIFIED,
    back_prop=UNSPECIFIED,
    swap_memory=UNSPECIFIED,
    maximum_iterations=UNSPECIFIED
)

The parameters apply to and only to the immediately enclosing loop. It only has effect if the loop is staged as a TF while_loop; otherwise the parameters have no effect.

Usage example:

@tf.function(autograph=True) def dynamic_rnn(..., parallel_iterations=32): num_steps = ... for t in tf.range(num_steps): tf.autograph.experimental.set_loop_options( parallel_iterations=parallel_iterations) ...

Args:

  • parallel_iterations: See tf.while_loop.
  • back_prop: See tf.while_loop.
  • swap_memory: See tf.while_loop.
  • maximum_iterations: See tf.while_loop.

results matching ""

    No results matching ""