<h1 id="artius-moss">Artius Moss</h1>A photorealistic, versatile MOSS-Image merge with a carefully balanced fine-tune, LoRA and built-in turbo. Strong NSFW capabilities, sharp results in just 8–10 steps, and an optimized INT8 build that runs on GPUs with 8 GB VRAM in ComfyUI.⚡ 8–10 steps · CFG 1 · ~7 GB · runs on 8 GB GPUs · custom text encoder &amp; VAE included⚖️ It's all about balance. The fine-tune and LoRAs were tested together and adjusted to work as a single model, rather than letting one dominate the result. That means no forced signature look across every image. Portraits, landscapes, fashion, fantasy, NSFW, stylized artwork — Artius Moss can handle a wide range of subjects without locking you into one aesthetic.🎨 Built for everyday image creation, whether you're after convincing photographs, unusual artwork, or something more experimental. Describe the scene, choose your style, and see where it takes you.<h2 id="why-it's-worth-a-try">✨ Why it's worth a try</h2><ul><li>Fine-tune + LoRA merge — LoRA weights were chosen through side-by-side testing to find a balance that works in practice.</li><li>Versatile — switch between photorealism, illustration, fantasy, and other styles without changing models.</li><li>🔞 Uncensored — strong NSFW capabilities, with no built-in content filtering. The model follows your prompt.</li><li>Custom text encoder included — an uncensored Qwen3.5-9B encoder made for this model, so your prompt reaches the image without being softened along the way.</li><li>Fast — turbo is already built in, so there's no need to load an extra LoRA.</li><li>Fits a mid-range GPU — the optimized INT8 build takes up around 7 GB and runs on 8 GB cards.</li><li>Ready-made workflow — drag a showcase image into ComfyUI to load the complete workflow and get started.</li></ul><h2 id="what's-inside">🔧 What's inside</h2><ul><li>FP32 LoRA merge — all LoRAs are merged at full 32-bit precision before quantization, helping preserve fine details throughout the process.</li><li>Calibrated layer layout — each block was evaluated for its sensitivity to quantization. The most sensitive layers retain full precision, while the others use INT8 to keep the model compact with minimal quality loss.</li><li>Custom quantized uncensored text encoder — the custom encoder went through the same calibration: its sensitive layers stay in full precision, so prompts are understood almost exactly as by the full-size model. It can also expand short ideas into detailed prompts with ComfyUI's TextGenerate node.</li></ul><h2 id="settings-that-just-work">⚙️ Settings that just work</h2><ul><li>Steps: 8–10 (8 for speed, 10 for a little extra detail)</li><li>CFG: 1 — no negative prompt needed; just describe what you want to see.</li><li>Sampler / scheduler: euler / simple</li><li>Resolution: 1–2 MP (e.g. 1024×1024, 912×1632)</li><li>VAE: Flux 2 VAE (included in the files of this page)</li><li>Text encoder: the custom encoder from this page (recommended; CLIPLoader, type moss_image). The standard MOSS-Image Qwen3.5-9B encoder also works.</li></ul><h2 id="you'll-need">📦 You'll need</h2><ul><li>The latest ComfyUI with MOSS-Image support</li><li>The Load Diffusion Model node</li><li>An NVIDIA RTX 20xx or newer, with at least 8 GB VRAM</li></ul><hr />18+ only — please follow Civitai's content rules.Built on MOSS-Image by OpenMOSS (Apache 2.0). Thanks to the original model authors, the LoRA creators, and the ComfyUI team.

Files with SHA256: 16169cc2c5ec2fb86060caeb24d63d643872822da7138eee97768507a6f0ab08