mradermacher's picture
auto-patch README.md
93044c6 verified
metadata
base_model: mlabonne/Meta-Llama-3-120B-Instruct
language:
  - en
library_name: transformers
license: other
quantized_by: mradermacher
tags:
  - merge
  - mergekit
  - lazymergekit

About

weighted/imatrix quants of https://huggingface.co/mlabonne/Meta-Llama-3-120B-Instruct

static quants are available at https://huggingface.co/mradermacher/Meta-Llama-3-120B-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
GGUF i1-IQ1_S 26.1 for the desperate
GGUF i1-IQ1_M 28.6 mostly desperate
GGUF i1-IQ2_XXS 32.7
GGUF i1-IQ2_XS 36.3
GGUF i1-IQ2_S 38.1
GGUF i1-IQ2_M 41.4
GGUF i1-Q2_K 45.2 IQ3_XXS probably better
GGUF i1-IQ3_XXS 47.1 lower quality
PART 1 PART 2 i1-IQ3_XS 50.3
PART 1 PART 2 i1-Q3_K_S 52.9 IQ3_XS probably better
PART 1 PART 2 i1-IQ3_S 53.1 beats Q3_K*
PART 1 PART 2 i1-IQ3_M 54.8
PART 1 PART 2 i1-Q3_K_M 58.9 IQ3_S probably better
PART 1 PART 2 i1-Q3_K_L 64.1 IQ3_M probably better
PART 1 PART 2 i1-IQ4_XS 65.4
PART 1 PART 2 i1-Q4_0 69.2 fast, low quality
PART 1 PART 2 i1-Q4_K_S 69.5 optimal size/speed/quality
PART 1 PART 2 i1-Q4_K_M 73.3 fast, recommended
PART 1 PART 2 i1-Q5_K_S 84.1
PART 1 PART 2 i1-Q5_K_M 86.3
PART 1 PART 2 PART 3 i1-Q6_K 100.1 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.