undefined | Better HN

0 pointsgwern3y ago0 comments

> The alternatives are learning at the character level (way more complex

No, BPEs are more complex: you have a whole additional layer of preprocessing, with all sorts of strange and counterintuitive downstream effects and brand new ways to screw up (fun quiz question: everyone knows that BPEs use '<|endoftext|>' tokens to denote document breaks; what does the string '<|endoftext|>' encode to?). BPEs are reliably one of the ways that OA API users screw up, especially when trying to work with longer completions or context windows.

But a character is a character.

> and scales badly in memory/compute)

Actually very competitive: https://arxiv.org/abs/2105.13626#google (Especially if you account for all the time and effort and subtle bugs caused by BPEs.)

0 comments

6gvONxR4sf7o3y ago

Judging from the abstract, it looks like that paper talks about compute tradeoffs, but do they address memory tradeoffs? Because the context length limitations for (standard) transformers is holding them back from a whole host of applications, and memory being quadratic in sequence length seems like a hell of a cost to going from BPE tokens to characters.

gwernOP3y ago

You were paying that price to begin with, the BPEs don't magically resolve the quadratic. BPEs only compress by maybe 3x, and the larger the context window, the worse use a Transformer makes of it so the first 1024 or so characters are the most valuable (part of the problem is that document length drops off drastically in the training corpus). There are also many formulations of Transformer attention which change that quadratic (https://www.gwern.net/notes/Attention).

j / k navigate · click thread line to collapse