tf.contrib.training.enqueue_in_queue_dataset(
queue,
components
)
Defined in tensorflow/contrib/training/python/training/tensor_queue_dataset.py
.
Enqueue components into queue from PrependFromQueueAndPaddedBatchDataset
.
The components' dtypes and shapes must be compatible with the output_shapes
attribute of the dataset
created by
prepend_from_queue_and_padded_batch_dataset
. This operation supports both
non-batched and batched modes.
For more details, see the example in the docstring for
prepend_from_queue_and_padded_batch_dataset
.
Args:
queue
:variant
scalar or vector tensor. The tensor emitted by the first component of the iterator associated withprepend_from_queue_and_padded_batch_dataset
. If this is a scalar, then thecomponents
input tensors should not have a prepended batch dimension.components
: Nested tuple of tensors, each with a leading batch dimension ifqueue
is a vector. The structure, dtypes, and shapes (excluding batch dimension) must match the nested tuplesdataset.output_types[1]
anddataset.output_shapes[1]
(the non-queue output types and shapes) of thedataset
emitted by the originalprepend_from_queue_and_padded_batch_dataset
call.
Returns:
An Operation
that enqueues components
into the dataset(s) associated
with entries of queue
.