ixFddlmZddlmZddlmZeGddZy)) dataclass)reduce)_remote_deviceceZdZUdZgdZeeed<eeed<edzed< d deedeede ezdzfdZ dZ y) ShardMetadataaH Represents a shard of the overall Tensor including its offsets, lengths and device placement. Args: shard_offsets(List[int]): Offsets in the original tensor indicating the start offsets for this shard. Should have the same rank as the original tensor. shard_sizes(List[int]): Integers indicating the size of each dimension for this shard. Should have the same rank as the original tensor. placement(:class:`torch.distributed._remote_device`): Specifies the placement of this shard. ) shard_offsets shard_sizes placementrr Nr c||_||_t|trt ||_n||_t |jt |jk7r/tdt |jd|jdtt |jD]<}|j|dkr td|j|dks3tdy)NzMshard_offsets and shard_sizes should have the same number of elements, found z and z respectivelyrzshard_offsets should be >=0zshard_sizes should be >= 0) rr isinstancestrrr len ValueErrorrange)selfrr r is [/var/www/html/engine/venv/lib/python3.12/site-packages/torch/distributed/_shard/metadata.py__init__zShardMetadata.__init__s +& i %+I6DN&DN t!! "c$*:*:&; ;669$:L:L6M5NO''( 7  s4--./ ?A!!!$q( !>??"Q& !=>>  ?cd}t||jd}t||j|}|||j}|S)Nc$|dzt|zS)N)hash)abs r _hash_reducez,ShardMetadata.__hash__.._hash_reduce9sFd1g% %r%)rrr r )rrress r__hash__zShardMetadata.__hash__8sF &\4#5#5r:\4#3#3S93/ r)N) __name__ __module__ __qualname____doc__ __slots__listint__annotations__rr rrrrrrsj >I9c$$ 26 ?Cy?#Y?'$. ?2rrN) dataclassesr functoolsrtorch.distributed.remote_devicerrr(rrr,s'!: 66 6r