<h3 id="introduction">Introduction</h3>MOSS Cat Tower is <a target="_blank" rel="ugc" href="https://huggingface.co/OpenMOSS-Team/MOSS-Image-Base">MOSS-Image</a> fine-tuned model.This model features enhancements and adjustments to the default anime style.<h3 id="how-to-get-started-with-the-model">How to Get Started with the Model</h3>In order to perform inference, the following files are required in addition to this model.<ul><li>Text Encoders (use one of them)<ul><li><a target="_blank" rel="ugc" href="https://huggingface.co/Comfy-Org/MOSS-Image/blob/main/text_encoders/qwen3.5_9b_bf16.safetensors">qwen3.5_9b_bf16.safetensors</a></li><li><a target="_blank" rel="ugc" href="https://huggingface.co/Comfy-Org/MOSS-Image/blob/main/text_encoders/qwen3.5_9b_int8_convrot.safetensors">qwen3.5_9b_int8_convrot.safetensors</a></li><li><a target="_blank" rel="ugc" href="https://huggingface.co/Comfy-Org/MOSS-Image/blob/main/text_encoders/qwen3.5_9b_w4a8_gptq.safetensors">qwen3.5_9b_w4a8_gptq.safetensors</a></li></ul></li><li>VAE (use one of them)<ul><li><a target="_blank" rel="ugc" href="https://huggingface.co/Comfy-Org/MOSS-Image/blob/main/vae/flux2-vae_small.safetensors">flux2-vae_small.safetensors</a></li><li><a target="_blank" rel="ugc" href="https://huggingface.co/Comfy-Org/MOSS-Image/blob/main/vae/flux2-vae.safetensors">flux2-vae.safetensors</a></li></ul></li></ul>You need to use the webui that supports MOSS-Image.<ul><li>ComfyUI: v0.42.1 or later <a target="_blank" rel="ugc" href="https://huggingface.co/Comfy-Org/MOSS-Image">Refer here for how to place files in ComfyUI.</a></li><li>Forge Neo: 2.31 or later (I recommend updating to at least commit 0c388f8b26d26d475c2c7104aea4eafdf7bbb3e7)</li></ul><h3 id="recommended-setting">Recommended Setting</h3>The following settings are strongly recommended (especially those related to Skip Layer Guidance). Otherwise, the art style will be inconsistent, and the body may not be drawn correctly.<ul><li>Resolution: 1-2MP</li><li>Steps: 50</li><li>CFG scale: 3</li><li>Sampler: euler</li><li>Scheduler: simple</li><li>Single Layers, Layers to Skip: 22,24,26</li><li>Start Percent: 0.35</li><li>End Percent: 0.7</li></ul>Configrations related to Skip Layer Guidance:ComfyUI:Set the values ​​as shown in the image below to "SkipLayerGuidanceDiTSimple" Node<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/6f43b1d7-0970-4af6-a9d9-cdd129cf15ff/original=true/6f43b1d7-0970-4af6-a9d9-cdd129cf15ff.jpeg" />Forge Neo:1.Select "Settings" -&gt; "Stable Diffusion"2.Set the values ​​as shown in the image below<img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/dfd68ef6-8fef-4f55-ae45-e2d5461441cc/original=true/dfd68ef6-8fef-4f55-ae45-e2d5461441cc.jpeg" />3.Restart Web UI<h3 id="prompting-guidelines">Prompting guidelines</h3><a target="_blank" rel="ugc" href="https://github.com/OpenMOSS/MOSS-Image/blob/main/USERGUIDE.md">Refer to the Prompt Guidelines of MOSS-Image-Base User Guide.</a>Negative Prompt<pre><code>worst quality, bad anatomy, watermark, jpeg artifact</code></pre>

MOSS Cat Tower

<ul><li>This is a merged model with multiple full fine-tuned models derived from MOSS-Image-Base.</li><li>The model was trained using natural language captions that conformed to the MOSS-Image JSON format, and Danbooru tags.</li><li>Training Bucket Size: 2.4MP (1536x1536)</li></ul>

Files with SHA256: ad39fb65e1609e2c6584ac76ad80fcef8028c50e3d023b8da9ae0949074aedef