---
exported_from: deepseek-ai/deepseek-llm-67b-chat
language:
- en
library_name: transformers
license: other
license_link: LICENSE
license_name: deepseek
quantized_by: mradermacher
---
## About

<!-- ### quantize_version: 1 -->
<!-- ### output_tensor_quantised: 1 -->
<!-- ### convert_type:  -->
<!-- ### vocab_type:  -->
weighted/imatrix quants of https://huggingface.co/deepseek-ai/deepseek-llm-67b-chat

<!-- provided-files -->
static quants are available at https://huggingface.co/mradermacher/deepseek-llm-67b-chat-GGUF
## Usage

If you are unsure how to use GGUF files, refer to one of [TheBloke's
READMEs](https://huggingface.co/TheBloke/KafkaLM-70B-German-V0.1-GGUF) 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](https://huggingface.co/mradermacher/deepseek-llm-67b-chat-i1-GGUF/resolve/main/deepseek-llm-67b-chat.i1-IQ1_S.gguf) | i1-IQ1_S | 14.8 | for the desperate |
| [GGUF](https://huggingface.co/mradermacher/deepseek-llm-67b-chat-i1-GGUF/resolve/main/deepseek-llm-67b-chat.i1-IQ1_M.gguf) | i1-IQ1_M | 16.1 | for the desperate |
| [GGUF](https://huggingface.co/mradermacher/deepseek-llm-67b-chat-i1-GGUF/resolve/main/deepseek-llm-67b-chat.i1-IQ2_XXS.gguf) | i1-IQ2_XXS | 18.3 |  |
| [GGUF](https://huggingface.co/mradermacher/deepseek-llm-67b-chat-i1-GGUF/resolve/main/deepseek-llm-67b-chat.i1-IQ2_XS.gguf) | i1-IQ2_XS | 20.3 |  |
| [GGUF](https://huggingface.co/mradermacher/deepseek-llm-67b-chat-i1-GGUF/resolve/main/deepseek-llm-67b-chat.i1-IQ2_S.gguf) | i1-IQ2_S | 21.4 |  |
| [GGUF](https://huggingface.co/mradermacher/deepseek-llm-67b-chat-i1-GGUF/resolve/main/deepseek-llm-67b-chat.i1-IQ2_M.gguf) | i1-IQ2_M | 23.2 |  |
| [GGUF](https://huggingface.co/mradermacher/deepseek-llm-67b-chat-i1-GGUF/resolve/main/deepseek-llm-67b-chat.i1-Q2_K.gguf) | i1-Q2_K | 25.2 | IQ3_XXS probably better |
| [GGUF](https://huggingface.co/mradermacher/deepseek-llm-67b-chat-i1-GGUF/resolve/main/deepseek-llm-67b-chat.i1-IQ3_XXS.gguf) | i1-IQ3_XXS | 26.2 | lower quality |
| [GGUF](https://huggingface.co/mradermacher/deepseek-llm-67b-chat-i1-GGUF/resolve/main/deepseek-llm-67b-chat.i1-IQ3_XS.gguf) | i1-IQ3_XS | 28.0 |  |
| [GGUF](https://huggingface.co/mradermacher/deepseek-llm-67b-chat-i1-GGUF/resolve/main/deepseek-llm-67b-chat.i1-Q3_K_S.gguf) | i1-Q3_K_S | 29.4 | IQ3_XS probably better |
| [GGUF](https://huggingface.co/mradermacher/deepseek-llm-67b-chat-i1-GGUF/resolve/main/deepseek-llm-67b-chat.i1-IQ3_S.gguf) | i1-IQ3_S | 29.5 | beats Q3_K* |
| [GGUF](https://huggingface.co/mradermacher/deepseek-llm-67b-chat-i1-GGUF/resolve/main/deepseek-llm-67b-chat.i1-IQ3_M.gguf) | i1-IQ3_M | 30.6 |  |
| [GGUF](https://huggingface.co/mradermacher/deepseek-llm-67b-chat-i1-GGUF/resolve/main/deepseek-llm-67b-chat.i1-Q3_K_M.gguf) | i1-Q3_K_M | 32.8 | IQ3_S probably better |
| [GGUF](https://huggingface.co/mradermacher/deepseek-llm-67b-chat-i1-GGUF/resolve/main/deepseek-llm-67b-chat.i1-Q3_K_L.gguf) | i1-Q3_K_L | 35.7 | IQ3_M probably better |
| [GGUF](https://huggingface.co/mradermacher/deepseek-llm-67b-chat-i1-GGUF/resolve/main/deepseek-llm-67b-chat.i1-IQ4_XS.gguf) | i1-IQ4_XS | 36.3 |  |
| [GGUF](https://huggingface.co/mradermacher/deepseek-llm-67b-chat-i1-GGUF/resolve/main/deepseek-llm-67b-chat.i1-Q4_0.gguf) | i1-Q4_0 | 38.4 | fast, low quality |
| [GGUF](https://huggingface.co/mradermacher/deepseek-llm-67b-chat-i1-GGUF/resolve/main/deepseek-llm-67b-chat.i1-Q4_K_S.gguf) | i1-Q4_K_S | 38.5 | optimal size/speed/quality |
| [GGUF](https://huggingface.co/mradermacher/deepseek-llm-67b-chat-i1-GGUF/resolve/main/deepseek-llm-67b-chat.i1-Q4_K_M.gguf) | i1-Q4_K_M | 40.5 | fast, recommended |
| [GGUF](https://huggingface.co/mradermacher/deepseek-llm-67b-chat-i1-GGUF/resolve/main/deepseek-llm-67b-chat.i1-Q5_K_S.gguf) | i1-Q5_K_S | 46.6 |  |
| [GGUF](https://huggingface.co/mradermacher/deepseek-llm-67b-chat-i1-GGUF/resolve/main/deepseek-llm-67b-chat.i1-Q5_K_M.gguf) | i1-Q5_K_M | 47.8 |  |
| [PART 1](https://huggingface.co/mradermacher/deepseek-llm-67b-chat-i1-GGUF/resolve/main/deepseek-llm-67b-chat.i1-Q6_K.gguf.part1of2) [PART 2](https://huggingface.co/mradermacher/deepseek-llm-67b-chat-i1-GGUF/resolve/main/deepseek-llm-67b-chat.i1-Q6_K.gguf.part2of2) | i1-Q6_K | 55.4 | practically like static Q6_K |


Here is a handy graph by ikawrakow comparing some lower-quality quant
types (lower is better):

![image.png](https://www.nethype.de/huggingface_embed/quantpplgraph.png)

And here are Artefact2's thoughts on the matter:
https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9

## Thanks

I thank my company, [nethype GmbH](https://www.nethype.de/), for letting
me use its servers and providing upgrades to my workstation to enable
this work in my free time.

<!-- end -->