There's a popular free ID app that's mainly for forest trees of the north or something. The name doesn't give an indication of that restriction but it couldn't ID anything I have in Southern California. The database of that one is really limited.
Being able to ID tamarind is pretty impressive! Sounds like a good app.
Also got Physalis peruviana and Monstera deliciosa as its #1 guesses (tried M. deliciosa both by the stalk and leaves, both worked). It thought my Carica monoica was Carica papaya, but that's pretty close (nobody had uploaded C. monoica pictures). Got my acerola flowers top hit, although by leaves it was a fair ways down the list. My bananas were the third hit (I think the number one hit it guessed was some sort of Canna). Coffee leaves were fairly high on the list as well (but it couldn't identify it by its trunk; I added some pictures to hopefully help with that
). Got my Pachira aquatica first hit by leaves, but not my trunk (most trunks in the database were actually for P. glabra... I corrected that). Annona muricata was something like 6th or 7th on its guess list. It strangely wasn't able to get my coconut by its trunk, despite how distinct coconut trunks are (added some pictures), but it had no trouble with the leaves. I actually had a plant I didn't know what it was and it helped me figure it out (Eugenia uniflora - thanks, Plantnet, I thought I had one of those around here somewhere!). Can't remember what other plants I tried, but... I was very impressed, this is a super-difficult task it's doing.
Whenever I'm bored (for example, in the bathroom or waiting on something or whatnot) I also take part in rating other peoples' photos, to help the neural net. I'd guess I've rated maybe 500 or so so far, it's kind of addictive
I've gotten some major seed orders in route right now (five different sellers) and I plan to take lots of good training pictures of them, every stage from seed to maturity.