mradermacher/Llama-3-70Bx2-MOE-i1-GGUF

About

weighted/imatrix quants of https://huggingface.co/cloudyu/Llama-3-70Bx2-MOE

static quants are available at https://huggingface.co/mradermacher/Llama-3-70Bx2-MOE-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
GGUF	i1-IQ1_S	26.8	for the desperate
GGUF	i1-IQ1_M	29.4	mostly desperate
GGUF	i1-IQ2_XXS	33.9
GGUF	i1-IQ2_XS	37.6
GGUF	i1-IQ2_S	39.0
GGUF	i1-IQ2_M	42.5
GGUF	i1-Q2_K	46.9	IQ3_XXS probably better
GGUF	i1-IQ3_XXS	49.2	lower quality
PART 1 PART 2	i1-IQ3_XS	52.3
PART 1 PART 2	i1-IQ3_S	55.2	beats Q3_K*
PART 1 PART 2	i1-Q3_K_S	55.2	IQ3_XS probably better
PART 1 PART 2	i1-IQ3_M	56.6
PART 1 PART 2	i1-Q3_K_M	61.2	IQ3_S probably better
PART 1 PART 2	i1-Q3_K_L	66.3	IQ3_M probably better
PART 1 PART 2	i1-IQ4_XS	68.0
PART 1 PART 2	i1-Q4_0	72.1	fast, low quality
PART 1 PART 2	i1-Q4_K_S	72.5	optimal size/speed/quality
PART 1 PART 2	i1-Q4_K_M	76.8	fast, recommended
PART 1 PART 2	i1-Q5_K_S	87.5
PART 1 PART 2	i1-Q5_K_M	90.1
PART 1 PART 2 PART 3	i1-Q6_K	104.2	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.