removing stale imports and forcing averaging of validation errors across ranks

# Not supporting for __all__

from . import inputs

#from .inputs import *

from . import io_utils

#from .io_utils import *

from . import network

#from .network import *

from . import runtime

#from .runtime import *

from . import ops

#from .ops import *

from . import network_utils

#from .network_utils import *

from . import optimizers

from . import mp_wrapper

from . import lr_policies

from . import automatic_loss_scaler

#__all__ = ['inputs', 'io_utils', 'network', 'runtime', 'network_utils', 'ops']

            loss_label= 'ABS_DIFF'

            errors = tf.losses.absolute_difference(tf.cast(labels, tf.float32), tf.cast(n_net.model_output, tf.float32), reduction=tf.losses.Reduction.MEAN)

        errors = tf.expand_dims(errors,axis=0)

        error_averaging = hvd.allreduce(errors)

        error_averaging = hvd.allreduce(errors, average=True)

        if num_batches is not None:

            num_samples = num_batches