Missed TensorFlow World? Check out the recap. Learn more

tf.contrib.tpu.outfeed_dequeue_tuple

Defined in generated file: python/ops/gen_tpu_ops.py

Retrieve multiple values from the computation outfeed.

tf.contrib.tpu.outfeed_dequeue_tuple(
    dtypes,
    shapes,
    device_ordinal=-1,
    name=None
)

This operation will block indefinitely until data is available. Output i corresponds to XLA tuple element i.

Args:

  • dtypes: A list of tf.DTypes that has length >= 1. The element types of each element in outputs.
  • shapes: A list of shapes (each a tf.TensorShape or list of ints). The shapes of each tensor in outputs.
  • device_ordinal: An optional int. Defaults to -1. The TPU device to use. This should be -1 when the Op is running on a TPU device, and >= 0 when the Op is running on the CPU device.
  • name: A name for the operation (optional).

Returns:

A list of Tensor objects of type dtypes.