Why no_shard strategy is deprecated in FSDP

In the code located at pytorch/torch/distributed/fsdp/_init_utils.py at main · pytorch/pytorch · GitHub Line 443, when I use the no-shard strategy in FSDP to switch to DDP, it displays a warning. I’m curious about why the NO-SHARD option is deprecated.

"elif sharding_strategy == ShardingStrategy.NO_SHARD:
“The NO_SHARD sharding strategy is deprecated. If having issues, "
“please use DistributedDataParallel instead.”,
# Level 1 is here, level 2 is from FullyShardedDataParallel, and
# level 3 is from the true caller