mradermacher's picture
auto-patch README.md
c6014ec verified
metadata
base_model: Kotokin/Merged-RP-Stew-V2-51B
language:
  - en
library_name: transformers
license: other
license_link: https://huggingface.co/01-ai/Yi-34B-200K/blob/main/LICENSE
license_name: yi-34b
quantized_by: mradermacher
tags:
  - merge
  - roleplay
  - exl2
  - not-for-all-audiences

About

weighted/imatrix quants of https://huggingface.co/Kotokin/Merged-RP-Stew-V2-51B

static quants are available at https://huggingface.co/mradermacher/Merged-RP-Stew-V2-51B-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 11.1 for the desperate
GGUF i1-IQ1_M 12.1 mostly desperate
GGUF i1-IQ2_XXS 13.8
GGUF i1-IQ2_XS 15.3
GGUF i1-IQ2_S 16.2
GGUF i1-IQ2_M 17.5
GGUF i1-Q2_K 19.1 IQ3_XXS probably better
GGUF i1-IQ3_XXS 19.8 lower quality
GGUF i1-IQ3_XS 21.2
GGUF i1-Q3_K_S 22.3 IQ3_XS probably better
GGUF i1-IQ3_S 22.3 beats Q3_K*
GGUF i1-IQ3_M 23.2
GGUF i1-Q3_K_M 24.8 IQ3_S probably better
GGUF i1-Q3_K_L 27.0 IQ3_M probably better
GGUF i1-IQ4_XS 27.5
GGUF i1-Q4_0 29.1 fast, low quality
GGUF i1-Q4_K_S 29.2 optimal size/speed/quality
GGUF i1-Q4_K_M 30.8 fast, recommended
GGUF i1-Q5_K_S 35.3
GGUF i1-Q5_K_M 36.2
GGUF i1-Q6_K 42.0 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.