sokann/DeepSeek-V4-Flash-0731-GGUF overview
DeepSeek V4 Flash 0731 GGUF This is a GGUF model for DeepSeek V4 Flash, created with the convert hf to gguf.py script, which does 2 things: Repacks routed expe…
Runs locally from ~94.09 GB disk (32 GB+ VRAM class GPUs with llama.cpp / guIDE).
Repository Files & Downloads
Model Details
Model README
---
base_model: deepseek-ai/DeepSeek-V4-Flash-0731
base_model_relation: quantized
license: mit
license_link: https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash-0731/blob/main/LICENSE
tags:
- deepseek_v4
- conversational
- ik_llama.cpp
- llama.cpp
---
DeepSeek-V4-Flash-0731-GGUF
This is a GGUF model for DeepSeek-V4-Flash, created with the convert_hf_to_gguf.py script, which does 2 things:
- Repacks routed experts tensors in FP4 to MXFP4.
- Quantizes other tensors in FP8 to Q8_0.
The GGUF model can be considered as lossless when compared to the original weights in safetensors.
It works fine in both ik_llama.cpp and llama.cpp. It is faster with ik_llama.cpp, but support for SWA is still in progress.
Run sokann/DeepSeek-V4-Flash-0731-GGUF with guIDE
Download guIDE — the AI-native code editor with local LLM inference and 69 built-in tools.
Source: Hugging Face · Compare models