Deidra is a LoRA that helps you produce a beautiful image of a woman with pale skin.<h2 id="recommended-usage-dxij872qw">Recommended usage</h2><ol><li>Put "score_9, score_8_up, score_7_up, realistic, ccdeidra, 1girl," at the beginning of the positive prompt.</li><li>Set LoRA weight to 0.5 and tweak from there (Note that 1.0 is too strong and image quality will suffer).</li><li>Use highres fix.</li><li>In ComfyUI, use a1111 for weight interpretation in the text encoding node instead of comfy weight interpretation.</li><li>Use with CyberRealistic Pony v7 as the main model (<a target="_blank" rel="ugc" href="https://civarchive.com/models/443821/cyberrealistic-pony">https://civarchive.com/models/443821/cyberrealistic-pony</a>) as it was used for LoRA training.</li></ol><h2 id="example-prompt-2dar0j5om">Example prompt</h2>Positive<pre><code>score_9, score_8_up, score_7_up, realistic, ccdeidra, 1girl
professional portrait, beautiful (cute, sweet) 22yo woman wearing a spring dress,
(cowboy shot), (downturned eyes),
golden hours, bouque of flowers,
sunset,
dramatic lighting, light and deep shadow,
best quality, masterpiece,
photorealistic, sharp focus, masterpiece, best quality, highly detailed, raw photo, 4k, 8k, uhd,
shallow depth of field, bokeh,
shot on Canon 5d mark iv, professional lighting
&lt;lora:ccdeidra_v1:0.5&gt;</code></pre>Negative<pre><code>score_6, score_5, score_4,
lowres, low quality, worst quality, 
blurry, fuzzy, hazy,
illustration, cartoon, drawing, render,
(busty), (fat, obese:1.4), (ugly:1.5),
(disfigured), masculine, man, gorilla, ape,
child, childlike, 
strong jawline,ugly,
(long torso:1.2), long neck,
bad anatomy, disfigured, tiling, deformed, mutated, out of frame, cloned,
dark skin, asian, indian, latino, monolid eyes,
mutation, mutated, disgusting, amputation, deformed, distorted, disfigured,
poorly drawn, bad anatomy, wrong anatomy,
poorly drawn faces, poorly drawn hands, 
extra limb, missing limb, floating limbs,disconnected limbs,mutated hands and fingers,tattoo,
watermark, text</code></pre><h3 id="dataset-used-for-lora-x66t4lohn">Dataset used for LoRA</h3>Main dataset<ul><li>87 synthetic safe-for-work images of adult women were generated using Flux-Fill FP8 (<a target="_blank" rel="ugc" href="https://civarchive.com/models/969431/flux-fill-fp8">https://civarchive.com/models/969431/flux-fill-fp8</a>).</li></ul>Regularization dataset<ul><li>199 synthetic safe-for-work images of adult women were generated using the epiCRealismXL v7-Final Destination model (<a target="_blank" rel="ugc" href="https://civarchive.com/models/277058?modelVersionId=489217">https://civarchive.com/models/277058?modelVersionId=489217</a>).</li></ul>

Files with SHA256: d6871cbcfd5e32f058f325e989019c60198cf8253f8086460a3c61f205eab96c