Jerome0207/DeepSeek-V4-Flash-Abliterated-DS4-Q2-Single-GPU

DeepSeek V4 Flash Abliterated — DS4 Q2 single-file cache mirror

This repository contains an unchanged server-side copy of Huihui-DeepSeek-V4-Flash-BF16-abliterated-ds4-Q2.gguf from `huihui-ai/Huihui-DeepSeek-V4-Flash-abliterated-ds4-GGUF`.

It exists as a dedicated one-file repository because RunPod cached-model storage downloads every weight file in a selected Hugging Face repository. Keeping only the 80.8 GiB Q2 artifact avoids caching the much larger Q2K, Q4K, and optional MTP files.

The model weights are not modified. Refer to the upstream model card for license, usage warnings, and attribution.