OpenWebUI connected tabbyUI’s OpenAI endpoint. I will try reducing temperature and seeing if that makes it more accurate.

Context was set to anywhere between 8k and 16k. It was responding in English properly, and then about halfway to 3/4s of the way through a response, it would start outputting tokens in either a foreign language (Russian/Chinese in the case of Qwen 2.5) or things that don’t make sense (random code snippets, improperly formatted text). Sometimes the text was repeating as well. But I thought that might have been a template problem, because it seemed to be answering the question twice.

Otherwise, all settings are the defaults.

I tried it with both Qwen 14b and Llama 3.1. Both were exl2 quants produced by bartowski.

Perplexica works. It can understand ollama and custom OpenAI providers.

Super useful guide. However after playing around with TabbyAPI, the responses from models quickly become jibberish, usually halfway through or towards the end. I’m using exl2 models off of HuggingFace, with Q4, Q6, and FP16 cache. Any tips? Also, how do I control context length on a per-model basis? max_seq_len in config.json?

Seems to be the only necessary thing in my case! Thanks.

Yeah I definitely have the default GTK chooser. Guess I have some config playing to do later.

Can you explain a bit more about this and how to configure it? When I use FF on gnome, the save dialogue just looks like other dialogues?

Doesn’t gnome already have this?

Cygwin ftw

I would think access to the training data, or at least no restrictions on what you can do with the model, would be a good definition.

Ah right. What I really meant to ask was if it can do protocols other than http.

Which I don’t think it can…

Are you able to tunnel ports other than 80 and 443 through Cloudflare?

The GPL explicitly allows selling the software, like a proprietary software product. You don’t even need to have the code up in a public repo. What you DO need to do, though, is provide a reasonable way for customers to get the source code, and send it to them if they ask. Just because a project is GPL doesn’t mean you’re entitled to the source for free. Of course, if someone buys it and requests the source, they can do whatever they want with it, including uploading it somewhere. Which in the end, essentially makes it available to everyone. Which is the whole point!

All of this only works if the owners of the code respect the license. In this particular case, I don’t believe a contributor agreement was ever created, so if the new owners want to close source the apps, they’ll have to get permission from all contributors, or drastically rewrite those contributions.

But again, this only matters if they respect the license in the first place…

I think “complex” refers to the various dark patterns used by Windows and Mac/iOS to scare and/or force users that know nothing of computers into using the default browsers.

Excellent! You should update your original post with the solution, if you haven’t.

Can you list some more specifics then? Distro, what the theme should be, which applications are not respecting theme and how, are they GTK 3 or GTK 4, how are they installed (package manager, flatpak, snap, etc)?

Disable all extensions first, then reenable individually.

Flatpak and snaps do not integrate directly into system themes without special care. No idea how snap works, but flatpak requires some ‘bridge’ theme installed to properly interface with the system theme.

It’s very different now. Much more smooth. Of course you still configure the system manually. But following the handbook will get you a working system pretty easily.

There is at least a few of us. My gaming desktop, personal laptop, and work laptop are all running Gentoo!