James Mullenbach

Results 4 comments of James Mullenbach

Only 7 months later :) But in case someone else faces this, make sure you run make in the lib directory of this repo to compile all the cython stuff

Yeah, the lack of “same” padding is frustrating - you are right, technically, with this code the convolution output size will only be the same for odd width filters. One...

> For example, if only 1 true prediction in the top 5, the denominator should be 1 but in this case it would still be 5. Do you mean if...

Ah, I see what you mean now. Yes you can define it that way. I'm not actually sure what the standard practice is in information retrieval, which is where I...