[FSDP()][Easy] Make fully_shard() only FULL_SHARD (pytorch#88260)

We can have a separate API for each of the other sharding strategies.
Pull Request resolved: pytorch#88260
Approved by: https://github.com/mrshenli

torch/distributed/_composable/fully_shard.py

@@ -30,7 +30,6 @@
 def fully_shard(
     module: nn.Module,
     process_group: Optional[dist.ProcessGroup] = None,
-    sharding_strategy: Optional[ShardingStrategy] = None,
     mixed_precision: Optional[MixedPrecision] = None,
     cpu_offload: Optional[CPUOffload] = None,
     auto_wrap_policy: Optional[Callable] = None,
@@ -51,7 +50,7 @@ def fully_shard(
     forward_prefetch_limit = 1
     state = _init_core_state(
         state,
-        sharding_strategy,
+        ShardingStrategy.FULL_SHARD,
         mixed_precision,
         cpu_offload,
         limit_all_gathers,