Post by Three Eyes Software in Release 1.7.0 - General

Viewing post in Release 1.7.0 - General

I've also experienced weird memory errors working with Qwen 3.5, though I'm not sure if it's an issue with KoboldCPP or the hardware itself.
Try Gemma-4-Sparse, it's a better model in most aspects anyway.

Guidelau1 day ago

No, it's okay, it seems to be working again.
Even though what I'm saying might sound stupid but, if one AI has more RAM than another, won't it necessarily perform better?
Or is Gemma-4-Sparse really is better than Qwen3.5?

Guidelau1 day ago

Okay, so no, the bug is back.

But I have a feeling it's because Koboldcpp isn't up to date. Because when I changed the version of Koboldcpp to install in the settings.json file, to the new one, it fixed it. Incidentally, because the AI was still glitching and saying weird things.
They were talking to each other in the first and third person... in short, it was strange.

Three Eyes Software22 hours ago

Gemma 4 is much more stable, knowledgeable, and RP oriented than Qwen 3.5. You will notice that it's a sparse model with less active parameters, but it more than makes up for it by being "in its element".

Guidelau20 hours ago (2 edits)

Okay. I'll try that later...

Because right now I can't play the game at all. I'm putting Qwen 3.5 and the bug occurs.
I switch to Gama 4 Sparse, and the characters stay with the three dots for minutes.
So I tried OpenRouter, but it won't accept Apikeys.
So I think I'll wait a bit..

Three Eyes Software5 hours ago

This is just a guess, but make sure your KoboldCPP version is back to 1.111.1.
The OpenRouter key has to have credits on it, and sometimes extra characters slip in while pasting into the API key field, so try quick setup if you haven't already.

Guidelau3 hours ago

The version automatically reverts to version 1.111.1 when I start a game, so... And it's in this version that I'm experiencing the bug.
I'm going to try your thing for the openrouter.

itch.io

Viewing post in Release 1.7.0 - General