fix _FullyShardedDataParallelMapping when running test_fsdp.py

[Jose Lopez]

How to reproduce

python -m torch.distributed.run --nproc_per_node=2 --master_port=2333 ./tests/torch/nn/parallel/data_parallel/test_fsdp.py

Environment

• OS :
• Python version :
• Transformers version :
• Whether to use Docker:
• Misc.:

The problem is in _fsdp where _FullyShardedDataParallelMappingForHuggingFace is used instead of _FullyShardedDataParallelMapping
from oslo.transformers.mapping_utils import ( _FullyShardedDataParallelMappingForHuggingFace, )

[Jose Lopez]

I'm creating a PR fixing this