examples: add FDA training examples (pytorch_lightning, pytorch_lightning_distributed) (#1964)

Adds pytorch_lightning/fda.py and pytorch_lightning_distributed/fda.py following
their respective simclr.py templates. The model and loss are identical to SimCLR
(NTXentLoss + SimCLRProjectionHead); only the transform differs. FDATransform does
not accept input_size directly, so view transforms are passed explicitly — same
pattern as byol.py and tico.py.

Also regenerates the corresponding notebooks via make generate-example-notebooks.

Part of #1956

Co-authored-by: Saud Kamran <saud.kamran92@gmail.com>