Calico Cat Tower is <a target="_blank" rel="ugc" href="https://civarchive.com/models/950531">RouWei</a> based checkpoint.VAE is baked in.RouWei has its own specific features and prompting, so please refer to the <a target="_blank" rel="ugc" href="https://civarchive.com/models/950531">RouWei</a> page for instructions on how to use it.You need to use the webui that support v-prediction if you use v-prediction version.<ul><li>ComfyUI (recommended)</li><li>Forge Classic (recommended)</li><li>reForge</li><li>Forge</li><li>AUTOMATIC1111 (place the config file in the same folder as the checkpoint)</li></ul><h3 id="license-yg6rfqo7j">License</h3>This model merges the NoobAI-Xl model, so the license follows <a target="_blank" rel="ugc" href="https://civarchive.com/models/833294">NoobAI-XL License (Modified Fair AI Public License 1.0-SD)</a> and is available for non-commercial use only. This license prohibits any form of commercialization, including but not limited to monetization or commercial use of the model, derivative models, or model-generated products.<h3 id="recommended-setting-for-v-prediction-version.-8rxs8s8oo">Recommended Setting for V-prediction version.</h3><ul><li>Steps: 20-30</li><li>CFG scale: 3-5</li><li>Sampler: Euler a</li><li>Use ADetailer as needed</li></ul><h3 id="recommended-setting-for-epsilon-prediction-version.-6xwbrki00">Recommended Setting for Epsilon-prediction version.</h3><ul><li>Steps: 20-30</li><li>CFG scale: 3-7</li><li>Sampler: Euler a, DPM++ 2M Karras</li><li>Use ADetailer as needed</li></ul><h3 id="positive-prompt-ucuu7dloi">Positive Prompt</h3><pre><code>masterpiece, best quality</code></pre>Enter the quality tag and artist tag as shown in the example below, otherwise the image quality will be reduced.<ul><li>Add 'BREAK' after them (for A1111 and A1111 derivatives such as Forge, recommended)</li><li>Use conditioning concat node (for ComfyUI, recommended)</li><li>Put them in the very end of prompt</li></ul><h3 id="negative-prompt-j8guvivym">Negative Prompt</h3><pre><code>worst quality, bad quality, low quality, lowres, scan artifacts, jpeg artifacts, sketch, light particles, watermark,</code></pre><h3 id="merge-recipes-(v-pred-v2.0)-0clx2sr20">Merge recipes (v-pred v2.0)</h3><ol><li>Merging the difference between <a target="_blank" rel="ugc" href="https://civarchive.com/models/950531?modelVersionId=1882934">Rouwei 0.80 vpred</a> and <a target="_blank" rel="ugc" href="https://huggingface.co/stabilityai/stable-diffusion-xl-base-1.0">stable-diffusion-xl-base-1.0</a>, <a target="_blank" rel="ugc" href="https://civarchive.com/models/312530?modelVersionId=1720768">CyberRealistic XL v5.6</a> and stable-diffusion-xl-base-1.0 to Rouwei 0.80 vpred by "Perpendicular Component" and "Add Difference" (Checkpoint A)</li><li>Merging Checkpoint A to Rouwei 0.80 vpred by "SLERP", alpha = 0.5 (Checkpoint B)</li><li>Merging Checkpoint B to Rouwei 0.80 vpred by "Rotate", alignment = 1.0, alpha = 0.0 (Checkpoint C)</li><li>Replacing the CLIP of Checkpoint C by Rouwei 0.80 vpred (Checkpoint D)</li><li>Finetuning Checkpoint D with anime style dataset(0.5k) by OFT method, dim = 4, alpha = 1e-3, learning rate = 2.5e-6, 15000 steps. And merging the OFT model to Checkpoint D, ratio = 1.0 (Checkpoint E)</li><li>Merging Checkpoint E to Checkpoint D by "Weighted Sum" with block (0,0,0,0,0,0,0,0,0,0,0,1,1,0,0,0,0,0,0,0) (Checkpoint F)</li><li>Merging the <a target="_blank" rel="ugc" href="https://civarchive.com/models/920709?modelVersionId=1495826">Cat Tower vpred v1.7</a> to Checkpoint F by "SLERP" with block (0,0,0,0,0,0,0.5,0.5,0.5,0.5,0.5,0.5,0.5,0.5,0.5,0,0,0,0,0) (Checkpoint G)</li><li>Merging the difference between <a target="_blank" rel="ugc" href="https://civarchive.com/models/1277670?modelVersionId=1772645">JANKU v4.0</a> and <a target="_blank" rel="ugc" href="https://civarchive.com/models/950531?modelVersionId=1330983">RouWei v0.7 epred</a> to Checkpoint G by "Train Difference" with block (0,0,0,0,0,0,0,0,0,0,0,0.2,0.2,0.2,0.2,0,0,0,0,0) (Checkpoint H)</li><li><a target="_blank" rel="ugc" href="https://civarchive.com/models/1027203/noobvpencil-xl">Add v_pred and ztsnr keys to Checkpoint H by noob_v_pencil-XL merge recipes</a> (Calico Cat Tower v-pred v2.0)</li></ol><h3 id="merge-recipes-(e-pred-v2.0)-3oocoa656">Merge recipes (e-pred v2.0)</h3><ol><li>Merging the difference between <a target="_blank" rel="ugc" href="https://huggingface.co/Minthy/RouWei-0.8">rouwei_080_epsilon_fp16</a> and <a target="_blank" rel="ugc" href="https://huggingface.co/stabilityai/stable-diffusion-xl-base-1.0">stable-diffusion-xl-base-1.0</a>, <a target="_blank" rel="ugc" href="https://civarchive.com/models/312530?modelVersionId=1720768">CyberRealistic XL v5.6</a> and stable-diffusion-xl-base-1.0 to rouwei_080_epsilon_fp16 by "Perpendicular Component" and "Add Difference" (Checkpoint A)</li><li>Merging Checkpoint A to rouwei_080_epsilon_fp16 by "SLERP", alpha = 0.5 (Checkpoint B)</li><li>Merging Checkpoint B to rouwei_080_epsilon_fp16 by "Rotate", alignment = 1.0, alpha = 0.0 (Checkpoint C)</li><li>Replacing the CLIP of Checkpoint C by rouwei_080_epsilon_fp16 (Checkpoint D)</li><li>Finetuning Checkpoint D with anime style dataset(0.5k) by OFT method, dim = 4, alpha = 1e-3, learning rate = 2.5e-6, 15000 steps. And merging the OFT model to Checkpoint D, ratio = 1.0 (Checkpoint E)</li><li>Merging Checkpoint E to Checkpoint D by "Weighted Sum" with block (0,0,0,0,0,0,0,0,0,0,0,1,1,0,0,0,0,0,0,0) (Checkpoint F)</li><li>Merging the <a target="_blank" rel="ugc" href="https://civarchive.com/models/920709?modelVersionId=1776050">Cat Tower epred v1.3</a> to Checkpoint F by "SLERP" with block (0,0,0.5,0.5,0.5,0.5,0.5,0.5,0.5,0.5,0.5,0.5,0.5,0.5,0.5,0,0,0,0,0) (Checkpoint G)</li><li>Merging the difference between <a target="_blank" rel="ugc" href="https://civarchive.com/models/1277670?modelVersionId=1772645">JANKU v4.0</a> and <a target="_blank" rel="ugc" href="https://civarchive.com/models/950531?modelVersionId=1330983">RouWei v0.7 epred</a> to Checkpoint G by "Train Difference" with block (0,0,0,0,0,0,0,0,0,0,0,0.2,0.2,0.2,0.2,0,0,0,0,0) (Calico Cat Tower e-pred v2.0)</li></ol><h3 id="merge-recipes-(v-pred-v1.0)-bzax7q4fs">Merge recipes (v-pred v1.0)</h3><ol><li>Finetuning <a target="_blank" rel="ugc" href="https://civarchive.com/models/950531?modelVersionId=1360303">RouWei v0.7 vpred</a> with anime style dataset(0.5k) by OFT method, dim = 8, alpha = 1e-3, learning rate = 1e-4, 15000 steps</li><li>Merging the OFT model to <a target="_blank" rel="ugc" href="https://civarchive.com/models/950531?modelVersionId=1360303">RouWei v0.7 vpred</a>, ratio = 1.0 (Checkpoint A)</li><li>Merging Checkpoiint A to <a target="_blank" rel="ugc" href="https://civarchive.com/models/950531?modelVersionId=1360303">RouWei v0.7 vpred</a> by "Weighted Sum", alpha = 0.3 (Checkpoint B)</li><li>Merging the difference between <a target="_blank" rel="ugc" href="https://civarchive.com/models/860278?modelVersionId=1416446">Cat Carrier v3.0</a> and <a target="_blank" rel="ugc" href="https://civarchive.com/models/1232765?modelVersionId=1389133">Illustrious XL v1.0</a> to Checkpoint B by "Add Difference" with block (0,0,0,0,0,0,0,0,0,0,0,0.5,0.5,0.5,0.5,0,0,0,0,0) (Checkpoint C)</li><li>Merging the difference between <a target="_blank" rel="ugc" href="https://civarchive.com/models/920709?modelVersionId=1443338">Cat Tower vpred v1.6</a> and <a target="_blank" rel="ugc" href="https://civarchive.com/models/833294?modelVersionId=1190596">NoobAI-V-Pred-1.0-Version</a> to Checkpoint C by "Train Difference" with block (0,0,0,0,0,0,0,0,0,0,0,0.2,0.2,0.2,0.2,0,0,0,0,0) (Checkpoint D)</li><li>Merging the difference between <a target="_blank" rel="ugc" href="https://civarchive.com/models/615558">Catloaf</a> and <a target="_blank" rel="ugc" href="https://civarchive.com/models/257749/pony-diffusion-v6-xl?modelVersionId=290640">ponyDiffusionV6XL</a> to Checkpoint D by "Train Difference" with block (0,0,0,0,0,0,0,0,0,0,0,0.15,0.15,0.15,0.15,0,0,0,0,0) (Checkpoint E)</li><li>Merging the difference between <a target="_blank" rel="ugc" href="https://civarchive.com/models/894218?modelVersionId=1222866">copycat-noob Vpred_v1.01</a> and <a target="_blank" rel="ugc" href="https://civarchive.com/models/833294?modelVersionId=1190596">NoobAI-V-Pred-1.0-Version</a> to Checkpoint E by "Train Difference" with block (0,0,0,0,0,0,0,0,0,0,0,0.15,0.15,0.15,0.15,0,0,0,0,0) (Checkpoint F)</li><li>Merging style LoRA to Checkpoint F (Checkpoint G)</li><li><a target="_blank" rel="ugc" href="https://civarchive.com/models/1027203/noobvpencil-xl">Add v_pred and ztsnr keys to Checkpoint G by noob_v_pencil-XL merge recipes</a> (Calico Cat Tower v-pred v1.0)</li></ol>

Calico Cat Tower

<ul><li>RouWei v0.8.0 Epsilon-pred based</li><li>In order to synchronize with the v-pred version and the RouWei version, Epsilon-pred version 1 has been omitted.</li></ul>

Files with SHA256: ea197a9de23938c701f21b851c0ef6c718875a13a56bc8a7294412f222defd3c