Migrate fsdp1 -> fsdp2 in Tensor Parallelism Tutorial #4852

Job	Run time
pytorch_tutorial_build_worker (5, 15, linux.g5.4xlarge.nvidia.gpu)	22m 57s
pytorch_tutorial_build_worker (6, 15, linux.g5.4xlarge.nvidia.gpu)	35m 21s
pytorch_tutorial_build_worker (4, 15, linux.g5.4xlarge.nvidia.gpu)	32m 42s
pytorch_tutorial_build_worker (8, 15, linux.g5.4xlarge.nvidia.gpu)	24m 11s
pytorch_tutorial_build_worker (3, 15, linux.g5.4xlarge.nvidia.gpu)	21m 24s
pytorch_tutorial_build_worker (10, 15, linux.g5.4xlarge.nvidia.gpu)	16m 53s
pytorch_tutorial_build_worker (7, 15, linux.g5.4xlarge.nvidia.gpu)	21m 35s
pytorch_tutorial_build_worker (1, 15, linux.g5.12xlarge.nvidia.gpu)	15m 3s
pytorch_tutorial_build_worker (9, 15, linux.g5.4xlarge.nvidia.gpu)	19m 13s
pytorch_tutorial_build_worker (13, 15, linux.g5.4xlarge.nvidia.gpu)	15m 25s
pytorch_tutorial_build_worker (12, 15, linux.g5.4xlarge.nvidia.gpu)	15m 38s
pytorch_tutorial_build_worker (2, 15, linux.g5.4xlarge.nvidia.gpu)	37m 7s
pytorch_tutorial_build_worker (11, 15, linux.g5.4xlarge.nvidia.gpu)	14m 0s
pytorch_tutorial_build_worker (15, 15, linux.g5.4xlarge.nvidia.gpu)	18m 56s
pytorch_tutorial_build_worker (14, 15, linux.g5.4xlarge.nvidia.gpu)	28m 11s
pytorch_tutorial_build_manager	21m 28s
	6h 0m 4s

Provide feedback