by drbaph
Open source · 9k downloads · 45 likes
Z Image Turbo FP8 is an optimized FP8 precision version of the Z Image Turbo model, specializing in generating high-quality images from text. This model stands out for its rapid execution, delivering inference times under one second on high-end GPUs while remaining accessible on consumer graphics cards with just 16GB of VRAM. It excels particularly in creating photorealistic images, handling bilingual text (English and Chinese), and precisely adhering to provided instructions. Its capabilities also extend to image editing, enabling creative modifications based on textual prompts. The model positions itself as a versatile tool for artists, developers, and content creators seeking a high-performance and efficient solution.
This is a quantization of Comfy-Org/z_image_turbo to FP8_E5M2 and FP8_E4M3FN
| Precision | Image 1 | Image 2 |
|---|---|---|
| bf16 | ![]() | ![]() |
| fp8_e4m3fn | ![]() | ![]() |
Z-Image is a powerful and highly efficient image generation model with 6B parameters. It is currently has three variants:
🚀 Z-Image-Turbo – A distilled version of Z-Image that matches or exceeds leading competitors with only 8 NFEs (Number of Function Evaluations). It offers ⚡️sub-second inference latency⚡️ on enterprise-grade H800 GPUs and fits comfortably within 16G VRAM consumer devices. It excels in photorealistic image generation, bilingual text rendering (English & Chinese), and robust instruction adherence.
🧱 Z-Image-Base – The non-distilled foundation model. By releasing this checkpoint, we aim to unlock the full potential for community-driven fine-tuning and custom development.
✍️ Z-Image-Edit – A variant fine-tuned on Z-Image specifically for image editing tasks. It supports creative image-to-image generation with impressive instruction-following capabilities, allowing for precise edits based on natural language prompts.