mradermacher/FATLLAMA-1.7T-Instruct-i1-GGUF

About

weighted/imatrix quants of https://huggingface.co./RichardErkhov/FATLLAMA-1.7T-Instruct

The imatrix was created from the IQ4_XS quant.

static quants are available at https://huggingface.co./mradermacher/FATLLAMA-1.7T-Instruct-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
P1 P2 P3 P4 P5 P6 P7 P8	i1-IQ1_S	355.4	for the desperate
P1 P2 P3 P4 P5 P6 P7 P8	i1-IQ1_M	390.0	mostly desperate
P1 P2 P3 P4 P5 P6 P7 P8 P9 P10	i1-IQ2_XXS	447.7
P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 P11 P12	i1-IQ2_XS	498.2
P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 P11	i1-IQ2_S	524.0
P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 P11 P12 P13	i1-IQ2_M	570.1
P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 P11 P12 P13	i1-Q2_K	622.5	IQ3_XXS probably better
P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 P11 P12 P13 P14 P15 P16	i1-IQ3_XXS	649.8	lower quality
P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 P11 P12 P13 P14 P15	i1-IQ3_XS	692.1
P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 P11 P12 P13 P14 P15 P16 P17	i1-Q3_K_S	729.5	IQ3_XS probably better
P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 P11 P12 P13 P14 P15	i1-IQ3_S	731.8	beats Q3_K*
P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 P11 P12 P13 P14 P15 P16 P17	i1-IQ3_M	758.1
P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 P11 P12 P13 P14 P15 P16 P17	i1-Q3_K_M	814.9	IQ3_S probably better
P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 P11 P12 P13 P14 P15 P16 P17 P18	i1-Q3_K_L	888.4	IQ3_M probably better
P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 P11 P12 P13 P14 P15 P16 P17 P18 P19	i1-IQ4_XS	904.2
P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 P11 P12 P13 P14 P15 P16 P17 P18 P19 P20	i1-Q4_0_4_4	954.4	fast on arm, low quality
P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 P11 P12 P13 P14 P15 P16 P17 P18 P19 P20	i1-Q4_0_4_8	954.4	fast on arm+i8mm, low quality
P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 P11 P12 P13 P14 P15 P16 P17 P18 P19 P20	i1-Q4_0_8_8	954.4	fast on arm+sve, low quality
P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 P11 P12 P13 P14 P15 P16 P17 P18 P19 P20	i1-Q4_0	957.9	fast, low quality
P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 P11 P12 P13 P14 P15 P16 P17 P18 P19 P20	i1-Q4_K_S	961.5	optimal size/speed/quality
P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 P11 P12 P13 P14 P15 P16 P17 P18 P19 P20 P21 P22	i1-Q4_K_M	1015.5	fast, recommended
P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 P11 P12 P13 P14 P15 P16 P17 P18 P19 P20 P21 P22 P23 P24 P25 P26	i1-Q5_K_S	1166.0
P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 P11 P12 P13 P14 P15 P16 P17 P18 P19 P20 P21 P22 P23 P24 P25 P26	i1-Q5_K_M	1197.5
P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 P11 P12 P13 P14 P15 P16 P17 P18 P19 P20 P21 P22 P23 P24 P25 P26 P27 P28 P29	i1-Q6_K	1390.9	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.