Red-Z-Image-AIO-1.5

Community-optimized all-in-one version of Alibaba's Z-Image-Turbo, integrating full components for easy setup, enhanced performance, Chinese prompt adherence, and dedicated NSFW fine-tuning with detail optimization for specific body parts.

Z-Image

Community Optimized

NSFW Fine-tuned

Chinese Prompts

ComfyUI

Stable Diffusion WebUI

Low VRAM

Style Lock

Visit Resource

Overview

Red-Z-Image-AIO-1.5 is a community-enhanced version of the official Z-Image-Turbo model, built on the S³-DiT single-stream diffusion architecture. It addresses key pain points like setup complexity, performance on low-end hardware, and creator-specific needs, including specialized NSFW tuning for realistic rendering of human anatomy.

Features

One-click integration of all components (model, encoder, VAE, plugins) for ComfyUI and SD WebUI
Optimized inference for 21.7% faster generation (1.8s on RTX 4090) and 11.5% lower VRAM usage
94% Chinese prompt adherence with enhanced traditional aesthetics and cultural elements
Dedicated NSFW fine-tuning with detail optimization for specific body parts, improving skin texture, light-shadow, and anatomical accuracy
Built-in style locking for consistent batch generation
Full compatibility with ControlNet, IP-Adapter, LoRA, and 8 built-in prompt templates

Images

Fine-tuned anatomical rendering of human form

Installation

Download from ModelScope or GitHub. Unzip and place in ComfyUI/models/checkpoints or SD WebUI/models/Stable-diffusion. No additional configuration required.

Usage

Use standard prompts with CFG 7-9, 20-30 steps. For NSFW, include descriptive terms for body parts; enable style lock for batches. Supports 512x512 to 1024x1024 resolutions.

Requirements

Python 3.8+
PyTorch 2.0+
ComfyUI or Stable Diffusion WebUI
NVIDIA GPU with 6GB+ VRAM (RTX 3060 compatible)
CUDA 11.8+