There are some experimental models made specifically for use with Home Assistant, for example home-llm.

Even though they are tiny 1-3B I’ve found them to work much better than even 14B general purpose models. Obviously they suck for general purpose questions just by their size alone.

That being said they’re still LLMs. I like to keep the “prefer handling commands locally” option turned on and only use the LLM as a fallback.

To expand: Just configure whatever profile you’re using (dev, release, …) to have link time optimization (lto) enabled:

[profile.release]
lto = "fat"

Reference

I’ve had this exact same gripe and can thankfully report that running EarlyOOM has fixed this for me.

Opening the app for the first time on my Fairphone 5 (listed as unsupported) actually crashed the OS, but after that it seems to be working ok.

Closing out of the in-app gallery causes the app to crash. But that can easily be worked around by using some other gallery app.

I’ll be testing it for a bit to see how it fares against other HDR methods…

I am looking at buying a G1 or maybe even a G2 but I haven’t heard of this screen “issue” yet.

Could you please elaborate on that or post a link to someone who already did? (A quick search mostly turned up replacement parts without much explanation)

Are there any implementations of this out there or is this purely theoretical (at this point in time)?