Z-Image Turbo GGUF inference engine written in C++/CUDA

Z-Image Turbo inference engine written in C++/CUDA: prompt -> 1024x1024 PNG - Optimized for 12GB 3060RTX

• uses an 8bit quantized zimage model
• uses a 4bit quantized CLIP model

Requires:

./MODEL/zimageturbo-q8_0.gguf

./CLIP/Z-Image.Q4_K_M.gguf

./VAE/ae.safetensors