Shard

class paddle.distributed. Shard

The Shard describes how Tensor splitted across multiple devices according to specified dimensions.

Parameters

dim (int) – specify the slicing dimension of the tensor.

Examples

>>> import paddle
>>> import paddle.distributed as dist
>>> mesh = dist.ProcessMesh([[2, 4, 5], [0, 1, 3]], dim_names=['x', 'y'])
>>> a = paddle.to_tensor([[1,2,3],[5,6,7]])
>>> 
>>> # distributed tensor
>>> d_tensor = dist.shard_tensor(a, mesh, [dist.Shard(0), dist.Shard(1)])
get_dim ( self: paddle.base.libpaddle.Shard ) int

get_dim

is_partial ( self: paddle.base.libpaddle.Placement ) bool

is_partial

is_replicated ( self: paddle.base.libpaddle.Placement ) bool

is_replicated

is_shard ( self: paddle.base.libpaddle.Placement, dim: Optional[int] = None ) bool

is_shard