DeepSeek-V4-Flash-Vision-Exp-GGUF

Overview

Unsloth's 2-bit cut of DeepSeek's experimental vision variant of V4-Flash - the same 284B MoE with 13B activated, plus image understanding via a separate mmproj projector. Fits 128 GB of unified memory or 2x 96 GB.

Recommended hardware

≈ 128 GB VRAM to run this 284B model at the UD-IQ2_XXS quantization.

Needs 128 GB+ of VRAM or unified memory (e.g. multi-GPU with 2x 96 GB cards, or a 128 GB Mac)

Links

• View on HuggingFace
• Run on Infersec