base_model: OpenBuddy/openbuddy-zen-56b-v21.2-32k
language:
- zh
- en
- fr
- de
- ja
- ko
- it
- fi
library_name: transformers
license: other
license_link: https://llama.meta.com/llama3/license/
license_name: llama3
quantized_by: mradermacher
tags:
- llama-3
About
weighted/imatrix quants of https://huggingface.co/OpenBuddy/openbuddy-zen-56b-v21.2-32k
static quants are available at https://huggingface.co/mradermacher/openbuddy-zen-56b-v21.2-32k-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 | 12.3 | for the desperate |
| GGUF | i1-IQ1_M | 13.4 | mostly desperate |
| GGUF | i1-IQ2_XXS | 15.3 | |
| GGUF | i1-IQ2_XS | 17.0 | |
| GGUF | i1-IQ2_S | 17.9 | |
| GGUF | i1-IQ2_M | 19.4 | |
| GGUF | i1-Q2_K | 21.1 | IQ3_XXS probably better |
| GGUF | i1-IQ3_XXS | 22.0 | lower quality |
| GGUF | i1-IQ3_XS | 23.4 | |
| GGUF | i1-Q3_K_S | 24.7 | IQ3_XS probably better |
| GGUF | i1-IQ3_S | 24.8 | beats Q3_K* |
| GGUF | i1-IQ3_M | 25.7 | |
| GGUF | i1-Q3_K_M | 27.5 | IQ3_S probably better |
| GGUF | i1-Q3_K_L | 30.0 | IQ3_M probably better |
| GGUF | i1-IQ4_XS | 30.5 | |
| GGUF | i1-Q4_0 | 32.2 | fast, low quality |
| GGUF | i1-Q4_K_S | 32.3 | optimal size/speed/quality |
| GGUF | i1-Q4_K_M | 34.1 | fast, recommended |
| GGUF | i1-Q5_K_S | 39.2 | |
| GGUF | i1-Q5_K_M | 40.2 | |
| GGUF | i1-Q6_K | 46.6 | practically like static Q6_K |
Here is a handy graph by ikawrakow comparing some lower-quality quant types (lower is better):
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.