mradermacher's picture
auto-patch README.md
c9f4bb4 verified
metadata
base_model: wolfram/miquliz-120b-v2.0
language:
  - en
  - de
  - fr
  - es
  - it
library_name: transformers
license: other
quantized_by: mradermacher
tags:
  - mergekit
  - merge

About

static quants of https://huggingface.co/wolfram/miquliz-120b-v2.0

While other static and imatrix quants are available already, I wanted a wider selection of quants available for this model.

weighted/imatrix quants are available at https://huggingface.co/mradermacher/miquliz-120b-v2.0-i1-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 Q2_K 44.6
GGUF IQ3_XS 49.3
PART 1 PART 2 Q3_K_XS 49.3
PART 1 PART 2 IQ3_S 52.1 beats Q3_K*
PART 1 PART 2 Q3_K_S 52.2
PART 1 PART 2 IQ3_M 53.8
PART 1 PART 2 Q3_K_M 58.2 lower quality
PART 1 PART 2 Q3_K_L 63.4
PART 1 PART 2 IQ4_XS 64.9
PART 1 PART 2 Q4_K_S 68.7 fast, recommended
PART 1 PART 2 IQ4_NL 68.8 prefer IQ4_XS
PART 1 PART 2 Q4_K_M 72.6 fast, recommended
PART 1 PART 2 Q5_K_S 83.2
PART 1 PART 2 Q5_K_M 85.4
PART 1 PART 2 PART 3 Q6_K 99.1 very good quality
PART 1 PART 2 PART 3 Q8_0 128.2 fast, best quality

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.