class torch.nn.SyncBatchNorm(num_features: int, eps: float = 1e-05, momentum: float = 0.1, affine: bool = True, track_running_stats: bool = True, process_group: Optional[Any] = None)[source]

Applies Batch Normalization over a N-Dimensional input (a mini-batch of [N-2]D inputs with additional channel dimension) as described in the paper Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift .

y=xE[x]Var[x]+ϵγ+βy = \frac{x - \mathrm{E}[x]}{ \sqrt{\mathrm{Var}[x] + \epsilon}} * \gamma + \beta

The mean and standard-deviation are calculated per-dimension over all mini-batches of the same process groups. γ\gamma and β\beta are learnable parameter vectors of size C (where C is the input size). By default, the elements of γ\gamma are sampled from U(0,1)\mathcal{U}(0, 1) and the elements of β\beta are set to 0. The standard-deviation is calculated via the biased estimator, equivalent to torch.var(input, unbiased=False).

Also by default, during training this layer keeps running estimates of its computed mean and variance, which are then used for normalization during evaluation. The running estimates are kept with a default momentum of 0.1.

If track_running_stats is set to False, this layer then does not keep running estimates, and batch statistics are instead used during evaluation time as well.


This momentum argument is different from one used in optimizer classes and the conventional notion of momentum. Mathematically, the update rule for running statistics here is x^new=(1momentum)×x^+momentum×xt\hat{x}_\text{new} = (1 - \text{momentum}) \times \hat{x} + \text{momentum} \times x_t , where x^\hat{x} is the estimated statistic and xtx_t is the new observed value.

Because the Batch Normalization is done for each channel in the C dimension, computing statistics on (N, +) slices, it’s common terminology to call this Volumetric Batch Normalization or Spatio-temporal Batch Normalization.

Currently SyncBatchNorm only supports DistributedDataParallel (DDP) with single GPU per process. Use torch.nn.SyncBatchNorm.convert_sync_batchnorm() to convert BatchNorm*D layer to SyncBatchNorm before wrapping Network with DDP.

  • num_featuresCC from an expected input of size (N,C,+)(N, C, +)

  • eps – a value added to the denominator for numerical stability. Default: 1e-5

  • momentum – the value used for the running_mean and running_var computation. Can be set to None for cumulative moving average (i.e. simple average). Default: 0.1

  • affine – a boolean value that when set to True, this module has learnable affine parameters. Default: True

  • track_running_stats – a boolean value that when set to True, this module tracks the running mean and variance, and when set to False, this module does not track such statistics, and initializes statistics buffers running_mean and running_var as None. When these buffers are None, this module always uses batch statistics. in both training and eval modes. Default: True

  • process_group – synchronization of stats happen within each process group individually. Default behavior is synchronization across the whole world

  • Input: (N,C,+)(N, C, +)

  • Output: (N,C,+)(N, C, +) (same shape as input)


>>> # With Learnable Parameters
>>> m = nn.SyncBatchNorm(100)
>>> # creating process group (optional)
>>> # process_ids is a list of int identifying rank ids.
>>> process_group = torch.distributed.new_group(process_ids)
>>> # Without Learnable Parameters
>>> m = nn.BatchNorm3d(100, affine=False, process_group=process_group)
>>> input = torch.randn(20, 100, 35, 45, 10)
>>> output = m(input)

>>> # network is nn.BatchNorm layer
>>> sync_bn_network = nn.SyncBatchNorm.convert_sync_batchnorm(network, process_group)
>>> # only single gpu per process is currently supported
>>> ddp_sync_bn_network = torch.nn.parallel.DistributedDataParallel(
>>>                         sync_bn_network,
>>>                         device_ids=[args.local_rank],
>>>                         output_device=args.local_rank)
classmethod convert_sync_batchnorm(module, process_group=None)[source]

Helper function to convert all BatchNorm*D layers in the model to torch.nn.SyncBatchNorm layers.

  • module (nn.Module) – module containing one or more attr:BatchNorm*D layers

  • process_group (optional) – process group to scope synchronization, default is the whole world


The original module with the converted torch.nn.SyncBatchNorm layers. If the original module is a BatchNorm*D layer, a new torch.nn.SyncBatchNorm layer object will be returned instead.


>>> # Network with nn.BatchNorm layer
>>> module = torch.nn.Sequential(
>>>            torch.nn.Linear(20, 100),
>>>            torch.nn.BatchNorm1d(100),
>>>          ).cuda()
>>> # creating process group (optional)
>>> # process_ids is a list of int identifying rank ids.
>>> process_group = torch.distributed.new_group(process_ids)
>>> sync_bn_module = torch.nn.SyncBatchNorm.convert_sync_batchnorm(module, process_group)


Access comprehensive developer documentation for PyTorch

View Docs


Get in-depth tutorials for beginners and advanced developers

View Tutorials


Find development resources and get your questions answered

View Resources