mradermacher's picture
auto-patch README.md
860bc91 verified
metadata
base_model: llmixer/Xwinter-120b
language:
  - en
library_name: transformers
license: llama2
quantized_by: mradermacher
tags:
  - Xwin
  - WinterGoddess
  - frankenmerge
  - 120b

About

weighted/imatrix quants of https://huggingface.co/llmixer/Xwinter-120b

static quants are available at https://huggingface.co/mradermacher/Xwinter-120b-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 25.1 for the desperate
GGUF i1-IQ1_M 27.2 mostly desperate
GGUF i1-IQ2_XXS 31.6
GGUF i1-IQ2_XS 35.1
GGUF i1-IQ2_S 36.8
GGUF i1-IQ2_M 40.0
GGUF i1-Q2_K 43.7 IQ3_XXS probably better
GGUF i1-IQ3_XXS 45.6 lower quality
GGUF i1-IQ3_XS 48.4
PART 1 PART 2 i1-Q3_K_S 51.1 IQ3_XS probably better
PART 1 PART 2 i1-IQ3_S 51.3 beats Q3_K*
PART 1 PART 2 i1-IQ3_M 53.0
PART 1 PART 2 i1-Q3_K_M 57.0 IQ3_S probably better
PART 1 PART 2 i1-Q3_K_L 62.1 IQ3_M probably better
PART 1 PART 2 i1-IQ4_XS 63.2
PART 1 PART 2 i1-Q4_0 66.7 fast, low quality
PART 1 PART 2 i1-Q4_K_S 67.2 optimal size/speed/quality
PART 1 PART 2 i1-Q4_K_M 71.0 fast, recommended
PART 1 PART 2 i1-Q5_K_S 81.4
PART 1 PART 2 i1-Q5_K_M 83.6
PART 1 PART 2 PART 3 i1-Q6_K 97.0 practically like static Q6_K

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

image.png

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.