CivArchive
    MossYume - v0.1
    NSFW
    Preview 148889640
    Preview 148889646
    Preview 148889644
    Preview 148889649
    Preview 148889648
    Preview 148889659
    Preview 148889660
    Preview 148889669
    Preview 148889670
    Preview 148889675
    Preview 148889681
    Preview 148889691
    Preview 148889707
    Preview 148889685
    Preview 148889699
    Preview 148889706
    Preview 148889710
    Preview 148889709
    Preview 148889716

    I. Introduction

    MossYume is a text-to-image model fine-tuned from MossImage-base, a high-quality image generation model developed by OpenMOSS-Team.

    II. Information

    For version 0.1:

    • Pure Human-Created Dataset: Trained on 20 million curated images spanning both anime and realistic styles, without any AI-generated images or model distillation (captioned with various type from tags, prose, structure caption).

    • Balanced Domain Distribution: Maintains 1:1 ratio between anime and realistic imagery.

    • Multi-Resolution Training: Trained across dynamic resolutions ranging from 1024px to 2048px.

    • Anime Data Sources: Sourced from Danbooru (updated up to September 4, 2026), e621, Gelbooru, Pixiv, and Konachan.

    • Native 2K Support: Capable of native high-resolution inference up to 2048×2048 (as you can see from my example images).

    • This is a base version without sft or RL

    • For this, I recommend you using complex to generate better images

    • If you want int8 version you can download here (thanks SumomoLee for supportng me to convert my model into int8)

    III. File Information

    • This file contains only the diffusion model and does not include a VAE or text encoder. To use it properly, you will need to download those components from the link here

    IV. Notes & Feedback

    This is an independent, community-driven experiment created out of a passion for exploring the untapped potential of the MOSS architecture. The primary goal of this fine-tune is to resolve anatomical inconsistencies and stabilize irregular weight distributions present in the original base model.


    Your feedback, suggestions, and creative prompt ideas are always welcome, every contribution helps make this model even better!

    V. Acknowledgments

    • Big thanks to narugo1992 for the dataset contributions and Neta.art for sharing the richly detailed dataset originally curated for Neta Lumina.

    • Credit to OpenMOSS-Team for the fantastic base model architecture.

    If you'd like to support my work, you can do so through Ko-fi!

    Description

    FAQ

    Checkpoint
    MOSS Image

    Details

    Downloads
    1,447
    Platform
    CivitAI
    Platform Status
    Available
    Created
    11/13/2026
    Updated
    11/20/2026
    Deleted
    -

    Files

    MossYume_v01_int8_convrot.safetensors

    MossYume_v01.safetensors

    Mirrors

    CivitAI (1 mirrors)