gguf-z-image

Z-Image-Turbo diffusion weights in GGUF format for stable-diffusion.cpp, mirrored byte-identical from leejet/Z-Image-Turbo-GGUF for JoyFusion's on-device (Android) inference.

File	Quant	Size	SHA-256
z_image_turbo-Q4_0.gguf	Q4_0	3,683,370,944	2bc57986874c84f7ec6d02d9d7070a53b0029954a0e38a6e1342eb91095572f5
z_image_turbo-Q8_0.gguf	Q8_0	6,577,440,704	df1c5baa86d1398c979495a6072dbcee79444fdb884a2445582ba0769c44e9a1

Shared components live in JoyFusionAI/gguf-components: text encoder Qwen3-4B-Instruct-2507-{Q4_0,Q8_0}.gguf and VAE ae.safetensors (FLUX.1 16-channel VAE).

Recommended: 8 steps, CFG 1.0, Euler, --diffusion-fa.

License

Z-Image-Turbo is released by Tongyi-MAI under the Apache License 2.0 (see LICENSE.md). The GGUF conversion was done by the stable-diffusion.cpp project; the files here are unmodified copies.