MOSS-Image SFT in one clean node: flip between 8-step turbo and full 50-step quality, with an optional built-in prompt enhancer that also writes your negative prompt.What's inside<ul><li>Turbo switch. One toggle swaps everything at once:<ul><li>ON: turbo LoRA, 8 steps, CFG 1. About 20 s per image on a 16 GB card.</li><li>OFF: no LoRA, 50 steps, CFG 3, negative prompt active. About 80 s.</li></ul></li><li>Prompt enhancer switch. Qwen3.5 9B (the same file Moss already uses as its text encoder, so nothing extra to download) turns a short idea into a full detailed prompt. It also writes a negative prompt tailored to that image, which gets used automatically when turbo is off.</li><li>Same seed = same composition in both modes. Explore fast with turbo, then flip turbo off on the seed you like for the final render.</li><li>VRAM-friendly. The text encoder is unloaded after prompting, so 50-step runs stay at full speed on 16 GB cards even with the enhancer on.</li><li>Built as a subgraph, Klein-template style. Double-click the node to see the enhancer, its system prompt, and the final prompts it sent.</li><li>The options on the node are tagged [turbo ON] / [turbo OFF], so you can see which ones each mode uses.</li></ul>Downloads<ul><li>Diffusion model: <a target="_blank" rel="ugc" href="https://civarchive.com/models/3005406/moss-image">MOSS Image SFT (Civitai)</a>, or <a target="_blank" rel="ugc" href="https://huggingface.co/Comfy-Org/MOSS-Image/resolve/main/diffusion_models/moss_image_sft_int8_convrot_attn.safetensors">moss_image_sft_int8_convrot_attn.safetensors</a> / <a target="_blank" rel="ugc" href="https://huggingface.co/Comfy-Org/MOSS-Image/resolve/main/diffusion_models/moss_image_sft_bf16.safetensors">bf16</a>. Goes in <code>models/diffusion_models</code>.</li><li>Text encoder: <a target="_blank" rel="ugc" href="https://huggingface.co/Comfy-Org/MOSS-Image/resolve/main/text_encoders/qwen3.5_9b_w4a8_gptq.safetensors">qwen3.5_9b_w4a8_gptq.safetensors</a>. Goes in <code>models/text_encoders</code>.</li><li>VAE: <a target="_blank" rel="ugc" href="https://huggingface.co/Comfy-Org/MOSS-Image/resolve/main/vae/flux2-vae.safetensors">flux2-vae.safetensors</a>, or optionally the <a target="_blank" rel="ugc" href="https://huggingface.co/black-forest-labs/FLUX.2-small-decoder/resolve/main/full_encoder_small_decoder.safetensors">FLUX.2 small decoder</a>. Goes in <code>models/vae</code>.</li><li>Turbo LoRA: <a target="_blank" rel="ugc" href="https://huggingface.co/lodestones/MOSS-Image-SFT-8-steps-dmd/resolve/main/lora_latest.safetensors">lora_latest.safetensors</a> by lodestones, at strength 0.7. Goes in <code>models/loras</code>.</li><li>Custom node: <a target="_blank" rel="ugc" href="https://github.com/Deno2026/comfyui-deno-custom-nodes">comfyui-deno-custom-nodes</a>, for the text-encoder unload step. If you don't use the enhancer, you can bypass it.</li></ul>Needs an up-to-date ComfyUI with native MOSS-Image support.Settings used<ul><li>Sampler euler, scheduler sgm_uniform, 1 MP (832×1248 portrait by default).</li><li>Skip-layer guidance on block 22 for the uncond pass (the official setting).</li><li>Shift: no ModelSampling node is needed, because ComfyUI already applies shift 5 to MOSS-Image. I tested it, and adding it produced pixel-identical images.</li></ul>Tips<ul><li>Be concrete. Moss takes prompts very literally. If something keeps going wrong (feet, poses, object size), describe the exact position instead of using a negative ("both feet resting on the wire floor of the basket").</li><li>Put text you want rendered in "quotes". It handles posters and signs really well.</li><li>The negative prompt only matters with turbo off. CFG 1 ignores it.</li></ul>Tested on: AMD RX 9060 XT 16 GB (ROCm, Windows).<ul><li>Turbo: about 20 s per image, or about 30–55 s with the enhancer.</li><li>Full 50 steps: about 80 s, or about 2 min with the enhancer.</li></ul>Credits: MOSS-Image by OpenMOSS, ComfyUI repack by Comfy-Org, 8-step DMD turbo LoRA by lodestones, unload node by Deno.

Files with SHA256: d02ddb79844b8f5aecb8a83878076fa5baf788fc6c868796e3f6aa526cea1516