[phplovesong]

Is tokenizing really the bottleneck? If we go from 20ms to 15ms does it really matter?

[gnabgib]

You sound like someone who used to write fastruby

[phplovesong]

Never heard of fast ruby. My point is tokenizing is rarely a bottleneck, as 99% of time is spent in inference.

So you speed up 1% of the pipeline by some factor, and the end result is unobservable for a human.

[antonvs]

Time to first token is observable by a human, and they’re reporting up to 10% reduction there.

Plus, inference is not the only place tokenization happens. This can make a big difference during development of ML models.