roselan

joined 1 year ago
[–] roselan@alien.top 1 points 11 months ago

for Deepseek there is https://chat.deepseek.com/ , for qwen I don't know.

[–] roselan@alien.top 1 points 11 months ago

This is not a test this is a military drill! poor LLMs :'(

[–] roselan@alien.top 1 points 11 months ago

I got starling-11b-q8_0.gguf in LM studio and can't get a decent output. Or I get a one liner like "here is a possible response: hello" with defaut or chatml preset, or i get pages of various "possible" responses filed with smileys most of the time (alpaca preset).

What chat format can I use?

[–] roselan@alien.top 1 points 11 months ago

I'm just testing it casually rn in lmsys, and really like it's flow and tone. It's pretty pleasant to speak to from the get go, it's a good start and I can't wait to dig a bit harder on it.

[–] roselan@alien.top 1 points 11 months ago

stupid stuff and silly scenarios. My latest:

Jane, Marc bratty over energetic sister, really wants to borrow Marc shiny new convertible. Marc is not so sure...

Write their over the top bickering. Jane is relentless and stops at nothing, to the exasperation of Marc.

For all serious stuff I use gpt4 of course.

[–] roselan@alien.top 1 points 11 months ago (2 children)

Yi-34B-Chat

It's not the most uncensored, and probably not the best, but I really like it's prose and coherence.

And Q4_K_M guff runs on my 32gb ram laptop.

(and yes it's slow)

[–] roselan@alien.top 1 points 11 months ago (1 children)

and of course TheBloke already prepped everything for our fine consumption.