<h2 id="description">Description</h2><p>The model makes 128x128 (upscaled to 1024x1024) or 160x160 (1280x1280) images on Krea2 Turbo with 8x8 logical pixel size (INT8 ConvRot, INT4, GGUF formats stay usable)</p><p></p><p>Use the '<u>Pixel Grid: Median Fixer</u>' node with 'Grid size 8' to post-process the image for absolute</p><p><a target="_blank" rel="ugc" href="https://registry.comfy.org/nodes/comfyui-pixelgrid-helpers">https://registry.comfy.org/nodes/comfyui-pixelgrid-helpers</a></p><p></p><p>Combining the LoRA with other LoRAs keep the 8x8 pixel sizes, but its best if other LoRAs are given a slightly lower strength.</p><p></p><p>Workflow:</p><p><img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/aec2b73d-3f64-493b-a594-09bc8c65cb49/original=true/aec2b73d-3f64-493b-a594-09bc8c65cb49.jpeg" /></p><h2 id="prompting-guide">Prompting guide</h2><h3 id="objective">Objective</h3><p>Write a direct, factual caption describing what is clearly visible in the image. Avoid guessing identities, emotions, background stories, or artistic mediums.</p><h3 id="description-order">Description Order</h3><ol><li><p><strong>Foreground Subjects:</strong> List main people, animals, or prominent objects in order of visual impact. Separate each subject with a semicolon.</p></li><li><p><strong>Background:</strong> Add the label <code>Background:</code> immediately after the subjects, then describe the environment, key surroundings, arrangement, and lighting.</p></li></ol><h3 id="formatting-template">Formatting Template</h3><p><code>&lt;subject 1 description&gt;; &lt;subject 2 description&gt;. Background: &lt;setting, surroundings, arrangement, and lighting&gt;</code></p><p><em>(If the image lacks a clear main subject, start directly with </em><code>No distinct subject.</code><em> followed by </em><code>Background: ...</code><em>)</em></p><h3 id="guidelines">Guidelines</h3><ul><li><p>Focus on distinct physical traits, clothing, poses, or actions to identify subjects.</p></li><li><p>Describe the scene clearly without listing every minor detail or repeating location terms.</p></li><li><p>Keep the output to a single plain-text paragraph with no extra labels, quotes, or markdown</p></li></ul><p><img src="https://image.civitai.com/xG1nkqKTMzGDvpLrqFT7WA/35bec796-4912-4342-88d7-917872f22eab/original=true/35bec796-4912-4342-88d7-917872f22eab.jpeg" /></p>

Files with SHA256: 773b1784d77e8dbc97cb8f3886957be970f884c277c8b28a0a74e66e95ad9701