👉 Update (Nov. 13, 2026) - "Draw" your MOSS-Image prompts with this free, open-source, standalone visual layout builder that works on any modern web browser, on any OS: https://github.com/Rexoplanets/layout-studio-moss-image 🤙
✧ ✧ ✧
👉 Update (Nov. 11, 2026) - Latest versions added (OPD-priming final: update 69800 (EMA, 5 epochs) of the v3 GLM-caption + bbox layout SFT: ComfyUI bf16 + int8 convrot). Added as SFT for the base model as per Lodestones' recommendation below in blue.
✧ ✧ ✧
From checkpoint creator Lodestones (https://huggingface.co/lodestones):
"Kromoss is a full fine-tune of MOSS-Image-Base at 1024 px. It uses the same architecture, inference pipeline, text encoder (Qwen3.5-9B) and VAE (FLUX.2 small decoder) as MOSS-Image-SFT, and this repo has the same layout, so it is a drop-in replacement.
The repo also ships two 8-step turbo LoRAs distilled from MOSS-Image-Base: 8 steps, CFG 1, no negative branch.
8-steps-dmd2-lora.safetensors (recommended): the DMD LoRA below, further trained with DMD2's GAN stage against real images. Slightly more contrast and detail, and the best PartiPrompts score."
The 8-steps DMD2 turbo LoRA can be found here: https://tensor.art/models/1057582086052906158/MOSS-BASE-DMD-8Step
