486
Lavalamp too hot (discuss.tchncs.de)
submitted 4 months ago* (last edited 4 months ago) by swiftywizard@discuss.tchncs.de to c/programmer_humor@programming.dev
77 comments fedilink hide all child comments
you are viewing a single comment's thread
view the rest of the comments
[-] bunchberry@lemmy.world 1 points 3 months ago

This happened to me a lot when I tried to run big models with low context windows. It would effectively run out of memory so each new token wouldn't actually be added to the context so it would just get stuck in an infinite loop repeating the previous token. It is possible that there was a memory issue on Google's end.

[-] FishFace@piefed.social 1 points 3 months ago

There is something wrong if it's not discarding old context to make room for new

[-] bunchberry@lemmy.world 1 points 3 months ago

At least llama.cpp doesn't seem to do that by default. If it overruns the context window it just blorps.

[-] FishFace@piefed.social 1 points 3 months ago

I think there are parameters for that, from googling.

this post was submitted on 25 Jan 2026
486 points (97.5% liked)

Programmer Humor

31559 readers
2011 users here now

Welcome to Programmer Humor!

This is a place where you can post jokes, memes, humor, etc. related to programming!

For sharing awful code theres also Programming Horror.

Rules

founded 2 years ago
MODERATORS
Feyter@programming.dev
anzo@programming.dev
BurningTurtle@programming.dev
pylapp@programming.dev