Class TPUConfig

Defined in tensorflow/contrib/tpu/python/tpu/

TPU related configuration required by TPUEstimator.


  • iterations_per_loop: This is the number of train steps running in TPU system before returning to CPU host for each This means global step is increased iterations_per_loop times in one It is recommended to be set as number of global steps for next checkpoint.
  • num_shards: (Deprecated, ignored by TPUEstimator). The number of model replicas in the system. For non-model-parallelism case, this number equals the total number of TPU cores. For model-parallelism, the total number of TPU cores equals num_cores_per_replica * num_shards.
  • num_cores_per_replica: Defaults to None, which disables model parallelism. An integer which describes the number of TPU cores per model replica. This is required by model-parallelism which enables partitioning the model to multiple cores. Currently num_cores_per_replica must be 1, 2, 4, or 8.
  • per_host_input_for_training: If True, PER_HOST_V1, or PER_HOST_V2, input_fn is invoked once on each host. With the per-core input pipeline configuration, it is invoked once for each core. With a global batch size train_batch_size in TPUEstimator constructor, the batch size for each shard is train_batch_size // #hosts in the True or PER_HOST_V1 mode. In PER_HOST_V2 mode, it is train_batch_size // #cores. In BROADCAST mode, input_fn is only invoked once on host 0 and the tensors are broadcasted to all other replicas. The batch size equals to train_batch_size. With the per-core input pipeline configuration, the shard batch size is alsotrain_batch_size` // #cores.
  • Note: per_host_input_for_training==PER_SHARD_V1 only supports mode.TRAIN.
  • tpu_job_name: The name of the TPU job. Typically, this name is auto-inferred within TPUEstimator, however when using ClusterSpec propagation in more esoteric cluster configurations, you may need to specify the job name as a string.
  • initial_infeed_sleep_secs: The number of seconds the infeed thread should wait before enqueueing the first batch. This helps avoid timeouts for models that require a long compilation time.
  • input_partition_dims: A nested list to describe the partition dims for all the tensors from input_fn(). The structure of input_partition_dims must match the structure of features and labels from input_fn(). The total number of partitions must match num_cores_per_replica. For example, if input_fn() returns two tensors: images with shape [N, H, W, C] and labels [N]. input_partition_dims = [[1, 2, 2, 1], None] will split the images to 4 pieces and feed into 4 TPU cores. labels tensor are directly broadcasted to all the TPU cores since the partition dims is None. Current limitations: This feature is only supported with the PER_HOST_V2 input mode.

  • Raises: * ValueError: If num_cores_per_replica is not 1, 2, 4 or 8.





Alias for field number 5


Alias for field number 6


Alias for field number 0


Alias for field number 2


Alias for field number 1


Alias for field number 3


Alias for field number 4