General discussion.
Heya @Endevor ! Can we get a Waifu to represent this one?
Experimental as it is, it deserves the love.
Wanted to comment and say I'm loving this model, experimental or not. Cannot wait to see how you build on it.
It is really nice. I've been using it more and more since, great surprise and I look forward to more work from Endevor.
Really good model, I like it!
Just wanted to ask what is the context size of the model? Thanks!
Just wanted to ask what is the context size of the model? Thanks!
It can do --contextsize 8192. Above this and things go badly.
You might want to try this if you want to push longer context:
https://huggingface.co/Lewdiculous/InfinityNoodleRP-7b-GGUF-IQ-Imatrix
Seemed to work at --contextsize 12288 (Koboldcpp), but more testing of the actual quality or outputs is necessary.
Understood, thank you. Would it be possible to merge a infinitely-laydiculus with the Fett-uccine-Long-Noodle-7B-120k-Context? I like the outputs and would be exciting if it can put out for an even longer context!
That would likely be long-noodle-7b-120k + infinitely-laydiculous 7b slerp merge I imagine is what you wanted to try. I'll request to the good chaotic peeps if they are not busy with a big queue cooking of course.
Coming back to say that this model really is pretty darn nice.
Just curious, is the response lengths similiar to those of limarp's or is it different in terms of token count.