feat(examples): TorchRL - MAML integration

[vmoens]

Brand-new version of the old PR

One can test this by running

python maml_torchrl.py --parallel --seed 1

I have added the --parallel flag, to allow users to have more than one environment used.

It can run on cuda if devices are present.

FYI: This should run on latest main in torchrl

torchrl example is not working

[vmoens]

Yes it is on train, just needs better sync on test wrt the seed setting. (At least on my machine, let me update the plot).

Results aren't quite different tbh, and we're still far from significant learning improvement in the original version (the error bars are hugely overlapping)
I'm happy to do some more hyperparameter optimisation to match the other results better.