Builds a graph operator that runs a replicated TPU computation.
tf.tpu.replicate( computation, inputs=None, infeed_queue=None, device_assignment=None, name=None, maximum_shapes=None )
computation: A Python function that builds the computation to replicate.
inputs: A list of lists of input tensors or
[]), indexed by
[replica_num][input_num]. All replicas must have the same number of inputs. Each input can be a nested structure containing values that are convertible to tensors. Note that passing an N-dimension list of compatible values will result in a N-dimention list of scalar tensors rather than a single Rank-N tensors. If you need different behavior, convert part of inputs to tensors with
infeed_queue: If not
InfeedQueuefrom which to append a tuple of arguments as inputs to computation.
device_assignment: If not
DeviceAssignmentdescribing the mapping between logical cores in the computation with physical cores in the TPU topology. Uses a default device assignment if
DeviceAssignmentmay be omitted if each replica of the computation uses only one core, and there is either only one replica, or the number of replicas is equal to the number of cores in the TPU system.
name: (Deprecated) Does nothing.
maximum_shapes: A nested structure of tf.TensorShape representing the shape to which the respective component of each input element in each replica should be padded. Any unknown dimensions (e.g. tf.compat.v1.Dimension(None) in a tf.TensorShape or -1 in a tensor-like object) will be padded to the maximum size of that dimension over all replicas. Note that if the input dimension is already static, we won't do padding on it and we require the maximum_shapes to have the same value or None on that dimension. The structure of
maximum_shapesneeds to be the same as
A list of outputs, indexed by
[replica_num] each output can be a nested
structure same as what computation() returns with a few exceptions.
Exceptions include: 1) None output: a NoOp would be returned which control-depends on computation. 2) Single value output: A tuple containing the value would be returned. 3) Operation-only outputs: a NoOp would be returned which control-depends on computation. TODO(b/121383831): Investigate into removing these special cases.
ValueError: If all replicas do not have equal numbers of input tensors.
ValueError: If the number of inputs per replica does not match the number of formal parameters to
ValueError: If the static
inputsdimensions don't match with the values given in
ValueError: If the structure of inputs per replica does not match the structure of