mradermacher/Tess-3-Mixtral-8x22B-i1-GGUF

About

weighted/imatrix quants of https://huggingface.co/migtissera/Tess-3-Mixtral-8x22B

llama.cpp crashes when creating some of the imatrix quants. Only the ones where it did not crash will be provided. quality is likely reduced. When in doubt, compare with the static quants, which should be safe.

static quants are available at https://huggingface.co/mradermacher/Tess-3-Mixtral-8x22B-GGUF

Usage

If you are unsure how to use GGUF files, refer to one of TheBloke's READMEs for more details, including on how to concatenate multi-part files.

Provided Quants

(sorted by size, not necessarily quality. IQ-quants are often preferable over similar sized non-IQ quants)

Link	Type	Size/GB	Notes
PART 1 PART 2	i1-Q2_K	52.2	IQ3_XXS probably better
PART 1 PART 2	i1-IQ3_XXS	55.0	lower quality
PART 1 PART 2	i1-IQ3_XS	58.3
PART 1 PART 2	i1-IQ3_S	61.6	beats Q3_K*
PART 1 PART 2	i1-Q3_K_S	61.6	IQ3_XS probably better
PART 1 PART 2	i1-IQ3_M	64.6
PART 1 PART 2	i1-Q3_K_M	67.9	IQ3_S probably better
PART 1 PART 2	i1-Q3_K_L	72.7	IQ3_M probably better
PART 1 PART 2	i1-IQ4_XS	75.6
PART 1 PART 2	i1-Q4_0	80.0	fast, low quality
PART 1 PART 2	i1-Q4_K_S	80.6	optimal size/speed/quality
PART 1 PART 2	i1-Q4_K_M	85.7	fast, recommended
PART 1 PART 2	i1-Q5_K_S	97.1
PART 1 PART 2 PART 3	i1-Q5_K_M	100.1
PART 1 PART 2 PART 3	i1-Q6_K	115.6	practically like static Q6_K

Here is a handy graph by ikawrakow comparing some lower-quality quant types (lower is better):

And here are Artefact2's thoughts on the matter: https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9

FAQ / Model Request

See https://huggingface.co/mradermacher/model_requests for some answers to questions you might have and/or if you want some other model quantized.

Thanks

I thank my company, nethype GmbH, for letting me use its servers and providing upgrades to my workstation to enable this work in my free time. Additional thanks to @nicoboss for giving me access to his private supercomputer, enabling me to provide many more imatrix quants, at much higher quality, than I would otherwise be able to.