mradermacher's picture
auto-patch README.md
f1e410b verified
|
raw
history blame
7.71 kB
metadata
base_model: NousResearch/Hermes-3-Llama-3.1-405B
language:
  - en
library_name: transformers
license: llama3
quantized_by: mradermacher
tags:
  - Llama-3
  - instruct
  - finetune
  - chatml
  - gpt4
  - synthetic data
  - distillation
  - function calling
  - json mode
  - axolotl
  - roleplaying
  - chat

About

weighted/imatrix quants of https://huggingface.co/NousResearch/Hermes-3-Llama-3.1-405B

static quants are available at https://huggingface.co/mradermacher/Hermes-3-Llama-3.1-405B-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
PART 1 PART 2 i1-IQ1_S 85.3 for the desperate
PART 1 PART 2 i1-IQ1_M 93.6 mostly desperate
PART 1 PART 2 PART 3 i1-IQ2_S 125.7
PART 1 PART 2 PART 3 i1-IQ2_M 136.8
PART 1 PART 2 PART 3 PART 4 i1-Q2_K 149.4 IQ3_XXS probably better
PART 1 PART 2 PART 3 PART 4 i1-Q3_K_S 175.3 IQ3_XS probably better
PART 1 PART 2 PART 3 PART 4 i1-Q3_K_M 195.5 IQ3_S probably better
P1 P2 P3 P4 P5 i1-Q4_K_S 230.6 optimal size/speed/quality
P1 P2 P3 P4 P5 i1-Q4_K_M 243.2 fast, recommended
P1 P2 P3 P4 P5 P6 P7 i1-Q6_K 333.0 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. 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.