we need to do some performance measurements to see if we introduced lag or slowness. the model comes out at around 12-13 token per second in the overall run and should be higher than that - the website reports over 30 token/secl. we probably also have to measure against ds4 here to see what that is giving us.
be aware that you can only run either against ds4 or against DS4Server for tests. you can't run both at the same time due to memory constraints.
we need to do some performance measurements to see if we introduced lag or slowness. the model comes out at around 12-13 token per second in the overall run and should be higher than that - the website reports over 30 token/secl. we probably also have to measure against ds4 here to see what that is giving us.
be aware that you can only run either against ds4 or against DS4Server for tests. you can't run both at the same time due to memory constraints.
hugo
added the bugcritical labels 2026-08-03 20:27:09 +00:00
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
we need to do some performance measurements to see if we introduced lag or slowness. the model comes out at around 12-13 token per second in the overall run and should be higher than that - the website reports over 30 token/secl. we probably also have to measure against ds4 here to see what that is giving us.
be aware that you can only run either against ds4 or against DS4Server for tests. you can't run both at the same time due to memory constraints.
can't reproduce right now (or well, too much effort right now to get going)