Boogu Image 0.1 Edit Turbo — GGUF (Flat Quants) GGUF quantizations of Boogu/Boogu Image 0.1 Edit Turbo, the distilled (Turbo) reference image edit variant of the Boogu Image family. Quantized for low VRAM ComfyUI use — an 8GB card (RTX 3070 class) with 16GB system RAM can run these. Quantized by realrebelai. These are the DiT only — you supply the text encoder and VAE separately (see below). Files File Quant Size boogu edit turbo dit Q8 0.gguf Q8 0 ~11.6 GB boogu edit turbo dit Q5 1.gguf Q5 1 ~8.64 GB boogu edit turbo dit Q5 0.gguf Q5 0 ~8.04 GB boogu edit turbo dit Q4 1.gguf Q4 1 ~7.44 GB boogu edit turbo dit Q4 0.gguf Q4 0 ~6.84 GB On 8GB VRAM, Q4 0 is the recommended sweet spot for the balance of VRAM savings and quality. Step up to Q5 1 or Q8 0 if you have the headroom and want maximum fidelity. Why only flat quants (Q4 0 / Q4 1 / Q5 0 / Q5 1 / Q8 0)? This repo provides flat quants only . Standard K quants (Q2 K, Q3 K M, etc.) require a hardcoded architectural mapping blueprint inside the llama.cpp source. Because the Boogu/OmniGen architecture is brand new, those K quant blueprints do not exist in the compiler yet. Flat quants bypass this requirement by forcing all 2D tensors…
We use cookies for essential functionality and analytics. You can accept or reject analytics cookies.Cookie policy