synced with main, cleaned outputs

[rahulkulhalli]

@Abby-Wheelis @shankari Please disregard the previous draft PR. I have synced this branch with the current main implementation.

[Abby-Wheelis]

With a few exceptions, I am able to run these now:

• still need to run the models in notebook 03, I'll let those run this evening while I'm at home
• getting an error with script 02 that there is no key transit - this is confusing since the script has not changed and it ran earlier, I will try this one again tonight as well

[Abby-Wheelis]

The models in 03 run for me, though they output all 1's, is this a consequence of the dataset (nicr) that I'm using and how small it is?

[rahulkulhalli]

With a few exceptions, I am able to run these now:

* still need to run the models in notebook 03, I'll let those run this evening while I'm at home

* getting an error with script 02 that there is no key `transit` - this is confusing since the script has not changed and it ran earlier, I will try this one again tonight as well

02 was not updated yesterday. it seems the preprocessing fix might have made introduced some changes in that code. I shall fix it today

[rahulkulhalli]

The models in 03 run for me, though they output all 1's, is this a consequence of the dataset (nicr) that I'm using and how small it is?

Yes, due to the small data size in NICR, they are overfitting. If you look at the cosine similarities, even they are very high.