Once configured, Tor Hidden Services also just work (you may need to use some fresh bridges in certain countries if ISPs block Tor there though). You don’t have to trust any specific third party in this case.
- 1 Post
- 21 Comments
Audalin@lemmy.worldto
Selfhosted@lemmy.world•Any of you have a self-hosted AI "hub"? (e.g. for LLM, stable-diffusion, ...)English
5·2 years agoIf your CPU isn’t ancient, it’s mostly about memory speed. VRAM is very fast, DDR5 RAM is reasonably fast, swap is slow even on a modern SSD.
8x7B is mixtral, yeah.
Audalin@lemmy.worldto
Selfhosted@lemmy.world•Any of you have a self-hosted AI "hub"? (e.g. for LLM, stable-diffusion, ...)English
7·2 years agoMostly via terminal, yeah. It’s convenient when you’re used to it - I am.
Let’s see, my inference speed now is:
- ~60-65 tok/s for a 8B model in Q_5_K/Q6_K (entirely in VRAM);
- ~36 tok/s for a 14B model in Q6_K (entirely in VRAM);
- ~4.5 tok/s for a 35B model in Q5_K_M (16/41 layers in VRAM);
- ~12.5 tok/s for a 8x7B model in Q4_K_M (18/33 layers in VRAM);
- ~4.5 tok/s for a 70B model in Q2_K (44/81 layers in VRAM);
- ~2.5 tok/s for a 70B model in Q3_K_L (28/81 layers in VRAM).
As of quality, I try to avoid quantisation below Q5 or at least Q4. I also don’t see any point in using Q8/f16/f32 - the difference with Q6 is minimal. Other than that, it really depends on the model - for instance, llama-3 8B is smarter than many older 30B+ models.
Audalin@lemmy.worldto
Selfhosted@lemmy.world•Any of you have a self-hosted AI "hub"? (e.g. for LLM, stable-diffusion, ...)English
10·2 years agoHave been using llama.cpp, whisper.cpp, Stable Diffusion for a long while (most often the first one). My “hub” is a collection of bash scripts and a ssh server running.
I typically use LLMs for translation, interactive technical troubleshooting, advice on obscure topics, sometimes coding, sometimes mathematics (though local models are mostly terrible for this), sometimes just talking. Also music generation with ChatMusician.
I use the hardware I already have - a 16GB AMD card (using ROCm) and some DDR5 RAM. ROCm might be tricky to set up for various libraries and inference engines, but then it just works. I don’t rent hardware - don’t want any data to leave my machine.
My use isn’t intensive enough to warrant measuring energy costs.
Audalin@lemmy.worldto
Sync for Lemmy@lemmy.world•I got a new phone, with all the same accounts, but cannot restore my previous lifetime purchases from sync. I'm not sure how to rectify it.English
2·2 years agoNo idea whether this would work, but have you tried moving app data via
adb?
Audalin@lemmy.worldto
Linux@lemmy.world•Update available! This version is very old.English
5·2 years agoMaybe not as important, but I still like having a fancy futuristic animation when a device is locked and idle.
I hope Wayland gets screensaver support at some point.
Audalin@lemmy.worldto
Linux@lemmy.world•[SOLVED] Bring existing KDE Konsole into focus instead of launching if one existsEnglish
3·2 years agoWell, that’s exactly what I did. My point was rather that there’s no single consistent way to do this across different DEs with different Wayland implementations - and that’s supposedly considered a feature from Wayland design’s perspective.
Audalin@lemmy.worldto
Linux@lemmy.world•[SOLVED] Bring existing KDE Konsole into focus instead of launching if one existsEnglish
8·2 years agoHere’s a KDE-specific script with
kdotool(Wayland always needs custom solutions for simple things):#!/bin/bash WIN="$(kdotool search --class org.kde.konsole | head -1)" if [[ "$WIN" != "" ]]; then kdotool windowactivate "$WIN" else konsole fi
Audalin@lemmy.worldto
Linux@lemmy.world•[SOLVED] Bring existing KDE Konsole into focus instead of launching if one existsEnglish
5·2 years agoIf you don’t find such a setting, you can try writing a script that checks if it’s running already (e.g. with
pgrep), activates the window if found (no idea how to do that in Wayland properly) and launches a new instance otherwise. Then use a custom.desktoplauncher for Konsole.
Disabling root login and password auth, using a non-standard port and updating regularly works for me for this exact use case.
Audalin@lemmy.worldOPto
Piracy: ꜱᴀɪʟ ᴛʜᴇ ʜɪɢʜ ꜱᴇᴀꜱ@lemmy.dbzer0.com•Sheet music resourcesEnglish
3·3 years agoAlas, their IP requirements are too much for me.
Anyway, is there much of xenharmonic sheet music there? Take, for example, Easley Blackwood (but not his books).
Audalin@lemmy.worldOPto
Piracy: ꜱᴀɪʟ ᴛʜᴇ ʜɪɢʜ ꜱᴇᴀꜱ@lemmy.dbzer0.com•Sheet music resourcesEnglish
5·3 years agoDownloading there is straightforward: look at network requests, redownload svg’s of individual pages with
wgetand reassemble those into a pdf. I did that today and the resulting quality wasn’t exactly low - though I didn’t examine it too closely. Readability was perfect.Probably could be automated, but I’m not bothered enough to do so yet.
Audalin@lemmy.worldto
Lemmy.World Announcements@lemmy.world•Lemmy.world (and some others) were hacked
31·3 years agoOr rather Electron apps.
Audalin@lemmy.worldto
Lemmy.World Announcements@lemmy.world•Lemmy.world (and some others) were hacked
24·3 years agoPasswords are safe, it’s temporary session tokens (already invalidated) that were exposed.
Audalin@lemmy.worldto
General Discussion@lemmy.world•How many communities are you subbed to?English
1·3 years agoWow. I’m subscribed to 13, and only one will be left of 5 of those once I figure out which is the best replacement for the defederated !Literature@beehaw.org.
~160 communities would be too much for me. If something isn’t my active interest, I prefer putting it into my locally hosted MediaWiki.
Audalin@lemmy.worldto
Piracy: ꜱᴀɪʟ ᴛʜᴇ ʜɪɢʜ ꜱᴇᴀꜱ@lemmy.dbzer0.com•HowTo: Download Video Content from ANY websiteEnglish
2·3 years agoAlternatively,
ffmpeg -protocol_whitelist file,crypto,data,https,tls,tcp -stats -i <URL.m3u8> -codec copy <FILE.mp4>.Also, some m3u8’s are just files containing redirects to other m3u8’s in various resolutions. You might want to extract the one you need and download that.
Audalin@lemmy.worldto
Lemmy.World Announcements@lemmy.world•Beehaw defederating effective immediately from lemmy.world and sh.itjust.worksEnglish
5·3 years agoAll sorts of things happen when large numbers of new people suddenly appear. Prepare for the second rexxit wave. Some time after it, federation patterns will likely stabilise.
Audalin@lemmy.worldto
Lemmy.World Announcements@lemmy.world•Beehaw defederating effective immediately from lemmy.world and sh.itjust.worksEnglish
15·3 years agoTheir Literature sublemmy was enjoyable. The alternatives are just too small to have discussions at the moment. Let’s see if a proper replacement arises.
…and I joined this instance because it federated with more or less everyone. It didn’t last long.
Audalin@lemmy.worldto
Lemmy.World Announcements@lemmy.world•Lemmy.world improvements and issuesEnglish
4·3 years agoRarely the web version is laggy and things I’m not intending to click are clicked. Just now, a “Report Created” notification appeared after posting a comment - I’m sorry if something was actually reported, and I have no idea what that even was.
It would. But it’s a good option when you have computationally heavy tasks and communication is relatively light.