Lighter and Faster
#1
by
bobig
- opened
So....128k context, and more Flash for faster answers? Like
So far It works with no drama
Yeah this is probably the best merge so far, I'm shortly going to add a Lightest model. The current Lighter model thinks for a short period of time before answering, and seems to be not too far away from QwQ-32B in terms of output quality, at least from my testing.