class torch.nn.utils.prune.PruningContainer(*args)[source]

Container holding a sequence of pruning methods for iterative pruning. Keeps track of the order in which pruning methods are applied and handles combining successive pruning calls.

Accepts as argument an instance of a BasePruningMethod or an iterable of them.


Adds a child pruning method to the container.


method (subclass of BasePruningMethod) – child pruning method to be added to the container.

classmethod apply(module, name, *args, **kwargs)

Adds the forward pre-hook that enables pruning on the fly and the reparametrization of a tensor in terms of the original tensor and the pruning mask.

  • module (nn.Module) – module containing the tensor to prune

  • name (str) – parameter name within module on which pruning will act.

  • args – arguments passed on to a subclass of BasePruningMethod

  • kwargs – keyword arguments passed on to a subclass of a BasePruningMethod


Simply handles the multiplication between the parameter being pruned and the generated mask. Fetches the mask and the original tensor from the module and returns the pruned version of the tensor.


module (nn.Module) – module containing the tensor to prune


pruned version of the input tensor

Return type

pruned_tensor (torch.Tensor)

compute_mask(t, default_mask)[source]

Applies the latest method by computing the new partial masks and returning its combination with the default_mask. The new partial mask should be computed on the entries or channels that were not zeroed out by the default_mask. Which portions of the tensor t the new mask will be calculated from depends on the PRUNING_TYPE (handled by the type handler):

  • for ‘unstructured’, the mask will be computed from the raveled list of nonmasked entries;

  • for ‘structured’, the mask will be computed from the nonmasked channels in the tensor;

  • for ‘global’, the mask will be computed across all entries.

  • t (torch.Tensor) – tensor representing the parameter to prune (of same dimensions as default_mask).

  • default_mask (torch.Tensor) – mask from previous pruning iteration.


new mask that combines the effects of the default_mask and the new mask from the current pruning method (of same dimensions as default_mask and t).

Return type

mask (torch.Tensor)

prune(t, default_mask=None)

Computes and returns a pruned version of input tensor t according to the pruning rule specified in compute_mask().

  • t (torch.Tensor) – tensor to prune (of same dimensions as default_mask).

  • default_mask (torch.Tensor, optional) – mask from previous pruning iteration, if any. To be considered when determining what portion of the tensor that pruning should act on. If None, default to a mask of ones.


pruned version of tensor t.


Removes the pruning reparameterization from a module. The pruned parameter named name remains permanently pruned, and the parameter named name+'_orig' is removed from the parameter list. Similarly, the buffer named name+'_mask' is removed from the buffers.


Pruning itself is NOT undone or reversed!


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