Mythalion (enhanced MythoMax) with Kimiko AND a 8K context buffer? Why no discussion !?

#1
by BoshiAI - opened

nRuaif, is this what I think it is, what it appears to be?

A version of Mythalion with Kimiko, capable of 8K context!?

I don't know why nobody has noticed or talked about this model until now? It's like you quietly slipped it out there with no fanfare or notice and nobody noticed it existed? I'm sure you know Kimiko is considered an important enhancement to MythoMax, so much so that most people promote 'MythoMax Kimiko v2' over the original. Mythalion was also considerd an improvement over MythoMax. Yet here it seems we have BOTH improvements in the same package, and there's an 8K context version out there!?

I'm amazed nobody's tested, reviewed or discussed this model yet. I'm eager to experiment and see what it can do. An 8K context version of MythoMax is like an answer to my prayers.

Out of curiousity, did you spend much time playing with this model yourself? What are your own thoughts on it? How does it compare versus MythoMax Kimiko and Mythalion? Maybe you're just humble, but if I were in your shoes, I'd have wanted a lot more people to be aware this model existed. Some of the key limitations of MythoMax are resolvable with Pygmalion, Kimiko and an 8K context buffer so I'm eager to see what this model can do. I'd love to hear your thoughts, and also why you didn't tell anyone about it!? lol

Also: I'm not sure if Kimiko is what has enabled MythoMax to work in 8K or if it was always capable of that, or even if you're aware there is an 8K version, but I noticed an 8K GGUF at the following address:

https://huggingface.co/actionpace/Mythalion-Kimiko-v2

I've started to test it out, and so far, I'm very happy with my new and enhanced 8K version of MythoMax, with Pygmalion and Kimiko :D

This comment has been hidden

This is now officially my favourite model, and I've tried a LOT. It seems more creative and warm in SFW RP, more creative and spicy in NSFW RP, and the GGUF version I'm using works great in 8K context mode! I've tried models based on MythoMax/REMM, MLewd, etc. but right now, this is my absolute favourite model. At the least, it certainly deserves more attention and I'm trying my best to get it some. :)

@TheBloke Is there any chance of getting GGUF quants of this? It's really impressed me as a model (MythoMax+Pygmalion+Kimiko) and I believe it's a very worthy mix to get the special @TheBloke treatment! :) Lots of people on the Faraday Discord are interested in getting quants of this & may well add to the available Faraday model DL options.

Sign up or log in to comment