Doc typo

[evanmiltenburg]

The documentation here says nlp.update([doc], [gold], drop=0.5, sgd=optimizer). I guess it should be: nlp.update([doc], [gold], drop=0.5, optimizer=sgd)

[honnibal]

No that's correct --- does it conflict with a previous statement?

I'm not that enthused about the sgd keyword argument, but that currently matches Thinc, and at least it's short.

[evanmiltenburg]

It seemingly contradicts the table directly below it. Or the table is confusing.

(Maybe it's just me, I'm a bit sleepy. But I initially thought the table listed the keyword arguments. But optimizer apparently isn't a keyword argument, but rather the value passed to sgd.)

[ines]

Okay yeah, I had a look at that table again and I'm not sure why we did it this way... I guess the optimizer was supposed to describe the variable optimizer declared in the example above. But it definitely makes a lot more sense to the table explain the arguments of nlp.update instead of all kinds of random functions, variables and arguments that happen to occur in the example. Will fix this, thanks! 👍

[evanmiltenburg]

Thanks for the quick response! Reading the page again I should have read the code a bit more carefully. It says optimizer = nlp.begin_training(get_data) right there on the first line...

[ines]

Ideally, docs should even make sense if you're just skimming them 😉 This one was definitely confusing. I mean, for what it's worth, you can call your optimizer unicorn and that's totally fine, as long as you pass it to the sgd keyword argument.

Just pushed an update to the site btw!

[lock]

This thread has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.