This is just a bug in their code. The paper they cite uses 1D convolutions. Thou... | Hacker News

HN2new | past | comments | ask | show | jobs | submit

		soraki_soladead on Dec 27, 2018 \| parent \| context \| favorite \| on: What Kagglers Are Using for Text Classification This is just a bug in their code. The paper they cite uses 1D convolutions. Though, I suppose having an unused dimension only really hurts efficiency.

gnulinux on Dec 27, 2018 [–]

> Though, I suppose having an unused dimension only really hurts efficiency.

That might not be true as it might increase bias and thus might need a more careful hyperparameter tuning to avoid overfitting.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact