Classes for TPU trace events.
An optimizer that averages gradients across TPU shards.
Mapping from logical cores in a computation to the physical TPU topology.
A helper object to build a device infeed queue.
RunConfig with TPU support.
TPU related configuration required by TPUEstimator.
Estimator with TPU support.
Ops and objects returned from a model_fn and passed to TPUEstimator.
Describes a set of TPU devices.
Shards computation along the batch dimension for parallel execution.
Returns the device name for a core in a replicated TPU computation.
An Op to sum inputs across replicated TPU instances. Each
Computes a device_assignment of a computation across a TPU topology.
A placeholder op for a value that will be fed into the computation.
A placeholder op for values fed into the TPU simultaneously as a tuple.
Initializes a distributed TPU system for use with TensorFlow.
An op which emits a single Tensor value from an XLA computation.
An op which emits multiple Tensor values from an XLA computation.
Builds a training loop that executes a fixed number of interations.
Builds a graph operator that runs a replicated TPU computation.
Rewrites computation for execution on a TPU system.
Shards computation for parallel execution.
Shuts down a running a distributed TPU system.
Builds a training loop for TPUs.