You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I asked Claude to checkout a couple of ideas into cutting notebook startup time. Of all the ideas I had, there seemed to be two which stood out as making an impact.
Cache dependancy Resolution
Project Leyden
Cache Dependancy Resolution
It appears that ammonite has something built-in, but it is disabled in the presence of hooks, which Almond always sets up. Claude's claim si that dependencies re-resolve themself every time and all the XML parsing does actually eat up some wall clock time. The proposal would be to setup something like a persistent "Version Set" cache in the user local. In my tests this shaved about 0.6s off the second and subsequent execution of the first cell.
I haven't put up a PR to almond itself because it's kind of messy and entirely vibe-coded, so here for discussion first. Happy to submit if there is appetite.
Project Leyden
Shaved about 1s (off of 6s) but I don't think it's possible to actually ship this as part of Almond itself? Filing here for posterity or future idea generation - Claude wrote up a script which (seems) to do some of the lifting of a training run.
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
Uh oh!
There was an error while loading. Please reload this page.
I asked Claude to checkout a couple of ideas into cutting notebook startup time. Of all the ideas I had, there seemed to be two which stood out as making an impact.
Cache Dependancy Resolution
It appears that ammonite has something built-in, but it is disabled in the presence of hooks, which Almond always sets up. Claude's claim si that dependencies re-resolve themself every time and all the XML parsing does actually eat up some wall clock time. The proposal would be to setup something like a persistent "Version Set" cache in the user local. In my tests this shaved about 0.6s off the second and subsequent execution of the first cell.
I haven't put up a PR to almond itself because it's kind of messy and entirely vibe-coded, so here for discussion first. Happy to submit if there is appetite.
Project Leyden
Shaved about 1s (off of 6s) but I don't think it's possible to actually ship this as part of Almond itself? Filing here for posterity or future idea generation - Claude wrote up a script which (seems) to do some of the lifting of a training run.
All reactions