samber

Joined 12 years ago

Karma
4,519
Submissions
792
Member since
Apr 11
@samuelberthe https://github.com/samber meet.hn/city/47.2186371,-1.5541362/Nantes
View on Hacker News

Recent submissions

Comment · 21 days ago
Do you need to disclose it when everybody do it ?

[EDIT] It is disclosed in contributors

Comment · 4 months ago
Comparing compute cost versus FlashAttention-2 is not very honest to me.

FlashAttention-2 is not used anymore for at least 2y.

This architecture would have been a massive improvement 3 years ago, but it is a ~solved~ problem IMO.

Comment · 4 months ago
According to Subquadratic, Needle in a Haystack is strong up to 12m tokens, but RULER has not been tested above 128k tokens ??
Comment · 5 months ago
OMG. I'm going to recode some of my libraries.
Comment · 7 months ago

Live from the official Hacker News API

A modern reader for Hacker News · SvelteKit + Skeleton