---
base_model:
- deepseek-ai/DeepSeek-V3.2-Speciale
---
[<img src="https://raw.githubusercontent.com/csabakecskemeti/devquasar/main/dq_logo_black-transparent.png" width="200"/>](https://devquasar.com)

'Make knowledge free for everyone'

- uploading - 

# !EXPERIMENTAL
Channel wise INT8 weigths for CPU inference. SGLang suppose to support this on CPU with AMX support (AFAIK - Xeon 6!?)
I'd appreciate a test from anyone with the necessary Hardware.

[SGLang how to ](https://lmsys.org/blog/2025-07-14-intel-xeon-optimization/)  

Channel wise INT8 version of [deepseek-ai/DeepSeek-V3.2-Speciale](https://huggingface.co/deepseek-ai/DeepSeek-V3.2-Speciale)

<a href='https://ko-fi.com/L4L416YX7C' target='_blank'><img height='36' style='border:0px;height:36px;' src='https://storage.ko-fi.com/cdn/kofi6.png?v=6' border='0' alt='Buy Me a Coffee at ko-fi.com' /></a>