fry69 @fry69.dev · May 27 Here is a longer running benchmark (4x), notably the overall runtime went down from ~42 min to ~30 min ->
0 likes 1 replies
?
Replies
20 t/s at 300k context window is nuts.
fry69 @fry69.dev · May 27 fry69 @fry69.dev · May 27