base_model: lightblue/DeepSeek-R1-Distill-Qwen-14B-Multilingual
datasets:
- lightblue/reasoning-multilingual-R1-Llama-70B-train
language:
- am
- ar
- bn
- zh
- cs
- nl
- en
- fr
- de
- el
- ha
- he
- hi
- id
- it
- ja
- jv
- km
- ko
- lo
- ms
- mr
- fa
- pl
- pt
- ro
- ru
- es
- sw
- sv
- tl
- ta
- te
- th
- tr
- uk
- ur
- vi
library_name: transformers
license: apache-2.0
quantized_by: mradermacher
tags:
- reasoning
About
weighted/imatrix quants of https://huggingface.co/lightblue/DeepSeek-R1-Distill-Qwen-14B-Multilingual
static quants are available at https://huggingface.co/mradermacher/DeepSeek-R1-Distill-Qwen-14B-Multilingual-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_M | 4.0 | mostly desperate |
| GGUF | i1-IQ2_XXS | 4.4 | |
| GGUF | i1-IQ2_XS | 4.8 | |
| GGUF | i1-IQ2_M | 5.5 | |
| GGUF | i1-Q2_K_S | 5.5 | very low quality |
| GGUF | i1-Q2_K | 5.9 | IQ3_XXS probably better |
| GGUF | i1-IQ3_XXS | 6.0 | lower quality |
| GGUF | i1-Q3_K_S | 6.8 | IQ3_XS probably better |
| GGUF | i1-IQ3_M | 7.0 | |
| GGUF | i1-Q3_K_M | 7.4 | IQ3_S probably better |
| GGUF | i1-Q3_K_L | 8.0 | IQ3_M probably better |
| GGUF | i1-IQ4_XS | 8.2 | |
| GGUF | i1-IQ4_NL | 8.6 | prefer IQ4_XS |
| GGUF | i1-Q4_K_S | 8.7 | optimal size/speed/quality |
| GGUF | i1-Q4_K_M | 9.1 | fast, recommended |
| GGUF | i1-Q6_K | 12.2 | 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.