upvote
Yep, that’s the question. I asked just that when Bonsai’s first models got released. Super interesting if we can push the parameter count over 100B with 1.125 bit quantization and still keep pretty good performance versus 16-bit 100B models. That’s a definite sweet spot.
reply
Doing some naive math, the F16 filesize is ~53.8gb, the 1-bit version is ~3.8gb, about 7% of the original size. The F16 size is roughly 2x param count, so that gives a rough ballpark of ~110B.
reply