ML tutorials: FFNN, torch->heyoka, differentiable drag

[Sceki]

This PR aims at:

• cleaning up a bit the ML tutorials, separating a FFNN tutorial from its application to dynamical systems
• creating a tutorial on how to port models from torch
• differentiable drag tutorial (stay tuned :)) --> this also includes Cartesian to geodetic transformation

[bluescarni]

@Sceki thanks for the PR! There are some failures in the CI which are not related to the new notebooks, I will need to sort them out.

[Sceki]

@bluescarni Ahh! Makes sense :) I was trying to understand why this mpp issue was appearing here and not in the push you made in the main

[bluescarni]

@bluescarni Ahh! Makes sense :) I was trying to understand why this mpp issue was appearing here and not in the push you made in the main

yep the usual random breakage with dependency versions...

[darioizzo]

@bluescarni this is good to go for me. @Sceki give it a last read and confirm (or push last mods). Then if the CI passes can be merged.

[Sceki]

@bluescarni Good to go as far as I am concerned!

[bluescarni]

@Sceki looking great, thanks a lot!