---
base_model: deepseek-ai/DeepSeek-V3-0324
language:
- en
library_name: transformers
license: mit
tags:
- deepseek_v3
- deepseek
- unsloth
- transformers
---

Expert pruning from 256 to 160

Attn: Q4_K

Experts: IQ1_S

Please refer to [unsloth](https://huggingface.co/unsloth/DeepSeek-V3-0324-GGUF-UD) for running this model.