Q8 gguf pl0x
https://huggingface.co/anthracite-org/magnum-v4-123b
p.s. I'm looking for something that falls in the close range of ~100Gb. So maybe Q6 instead of Q8? I "only" have 96GB of VRAM now. Selling other gpus to pay for upcoming Macbook Pro M4 (assuming it's > M3).
And I just read your "About Me."
I thought your name was Michael Radermacher. So much so, that I reached out to a guy on Twitter from Germany with the same name to thank him for one of the quants you made. His profile said he was into ML and AI and I thought that had to be you!
No wonder he never responded back!
I appreciate your quants and you're my favorite quanter on here!
I'll make "all" quants of this model, as per my usual method, don't worry. You can follow it at http://hf.tst.eu/status.html - the imatrix quants will be done on "nico1", but since the load is currently high, it will not happen before tomorrow, I guess.
However, the static quants will be done much earlier, and the static Q6_K is very close to the imatrix one, so you could grab that once it is done (they will be done on "db2").
Twitter from Germany with the same name to thank him for one of the quants you made. His profile said he was into ML and AI
That is an uncanny match, but indeed, not me. I just made up some fake (and german) name when I registered on huggingface. I regret, but what's done is done.