maybe that version of Macbook Pro is not supported you should ask the developer for supported
Others Multiic [v0.4.81] [Stvoler]
- Thread starter aleksandr unconditional
- Start date
That's exactly what I said, it confirms the previous question
thanks man i hope he gets a bigger better mac laptop so he can enjoy this game to
He always releases the new version to Patreon first before releasing it here obviouslymisteraleksandr unconditional there is a new version of Multiic version 0.4.81
I don't think it's about the qulity of the in-built GPU. All newer Mac Book Pros rely on in-built GPUs, even the most recent ones.
But since this a very Mac-specific solution it might very well be not supported. The Mac Book Pro is from 2021.
thanks man i appreciate it and thanks for the help and now it's waiting time until its releaset here
10 Gs is just too big. Many, if not most of us, won't waste the download time and memory space for more than 2 or 3 Gs. Especially for a new start.
I have almost 4 terabytes of games on my hard drive. Some I have been following through years of updates, and many of my favourites, with excellent art, smooth animation, and absorbing story lines, are still under 2 Gs.
I have almost 4 terabytes of games on my hard drive. Some I have been following through years of updates, and many of my favourites, with excellent art, smooth animation, and absorbing story lines, are still under 2 Gs.
can some one please give me the password for the for base, person, visionary or whatever it is down the settings for sandbox mode
reidanota
Active Member
- Nov 1, 2021
- 625
- 524
The game is 1.8GB, it's only the local version that's nearly 10GB because koboldcpp and the large language model are bundled. Configuring a local llm can be a boring task and it feels good, for a change, to have everything pre-packed and done for you.
If the author separated downloads - they could require users to set up koboldcpp themselves and fetch a gguf model from huggingface - it would be impossible to ensure, on their side, that everything worked for all users. Most applications using koboldcpp to run a local llm require you to correctly configure both kobold and the front-end. I understand their concern, doing it this way prevents potential supporters from giving up on the tech wall.
Mister_M
Engaged Member
- Apr 2, 2018
- 2,590
- 5,356
reidanota
Active Member
- Nov 1, 2021
- 625
- 524
I wouldn't say otherwise, but I don't have the game installed at the moment to confirm it. Could be that the game runs koboldcpp by script and expects a specific model to be in the installation folder, but you might still be able to trick it by insttalling a different one and renaming it? Really can't say without looking at it. If the game's settings handle and let you choose the model, you could probably just point to a different gguf. Perhaps if you download the "online" version, and you have your own koboldcpp installed elsewhere, you can still use game settings to set up and launch with your existing config, saving the need to download a new model?
francisdrake
Member
- Feb 21, 2019
- 106
- 101
Been trying to get this to run with a local 20b model and I keep getting a prompt to Wake Up/Stay during the first conversation. What's up with that?
Ok, maybe my 3rd or 4th times to try play this. Last time i play on my phone is no response on the first dialogue.
I'm trying to run locally. I keep getting "error: no choices in response data", and blank lines to fill in. Along with "error: kobold failed to fetch."
I also get something like "after 4 attempts no response was receive. Please try another model." That is not the exact language.
What am I doing wrong?
I also get something like "after 4 attempts no response was receive. Please try another model." That is not the exact language.
What am I doing wrong?
Last edited:
reidanota
Active Member
- Nov 1, 2021
- 625
- 524
Game uses a 14B model, I'm not sure this is it but my RTX 3090 sometimes chokes on 12B models or even 7B when the prompt gets too long. Don't now how this game handles context and prompting but when I last played, a good few months back, it worked for a while but after a few interactions the llm stopped responding. I'm downloading now to check it, maybe you can set a different model in settings and use something lighter, sacrificing complexity in responses to a faster and perhaps more streamlined experience?