NeoCyan

🍐𝑵𝒆𝒐𝑪𝒚𝒂𝒏 𝑴𝒐𝒅𝒆𝒍𝒔🍐

🧬𝑭𝒖𝒍𝒍 𝑫𝒆𝒔𝒄𝒓𝒊𝒑𝒕𝒊𝒐𝒏:🧬

The model is based on: RouWei [v0.7 Epsilon]

【𝘚𝘩𝘦 𝘸𝘢𝘴 𝘮𝘪𝘹𝘦𝘥 𝘢𝘯𝘥 𝘮𝘦𝘳𝘨𝘦𝘥 𝘸𝘪𝘵𝘩 2 𝘢𝘯𝘰𝘵𝘩𝘦𝘳 𝘮𝘰𝘥𝘦𝘭'𝘴】

According to the requirements of the authors, they will be included in the list.:

@Ocean3

🧬Lora used for Mix-Merge:🧬

SDXL FaeTastic Details v24 - Author: @Faeia

CyariYoda Lora v1 and v2 - Author: @NeoNi

⚠️【𝘼𝙙𝙙𝙞𝙩𝙞𝙤𝙣】: If you have any questions, please don't hesitate to write.

Stable Model - With beautiful style⚠️

🥝𝑴𝒊𝒙𝒆𝒅 𝒂𝒏𝒅 𝑴𝒆𝒓𝒈𝒆𝒅 𝒘𝒊𝒕𝒉 𝑳𝒐𝒓𝒂 𝒇𝒐𝒓 𝑫𝒆𝒕𝒂𝒊𝒍 𝒂𝒏𝒅 𝑴𝒚 𝑻𝒓𝒂𝒊𝒏𝒆𝒅 𝑳𝒐𝒓𝒂 𝒇𝒐𝒓 𝑺𝒕𝒚𝒍𝒆 +/- alpha/beta/gamma work🥝

▁ ▂ ▄ ▅ ▆ ▇ █ 📜𝑴𝒐𝒅𝒆𝒍 𝑫𝒆𝒔𝒄𝒓𝒊𝒑𝒕𝒊𝒐𝒏:📜 █ ▇ ▆ ▅ ▄ ▂ ▁

❝This model has its own distinct style and aesthetic. The style was personally trained by me and hardcoded into the model. Shortly after the model’s release, I’ll publish the LoRA itself—either alongside the release or later; I haven’t decided yet. Now, more details about the style: it’s soft, slightly fairy-tale-like, tailored for anime characters, but you can also experiment with semi-realistic outputs, etc. All specifics will be in the description. The model remains semi-flexible, compatible with Danbooru authors and their triggers. I’d describe it as designed for lo-fi content.❞

+ NeoCyan [RU] (добавлено; - [EN] added)

Brief description of NeoCyan:

The first and softest version; blurred style, less emphasis on details.

+ NeoCyan_v2 [RU] (добавлено; - [EN] added)

Brief description of NeoCyan_v2:

The second version slightly steps back from softness, becoming brighter and bolder with a stronger focus on details.

+ NeoCyan_v3 [RU] (добавлено; - [EN] added)

Brief description of NeoCyan_v3:

The third and most precise version, offering more detail while stepping away from softness, slightly losing some of its original style.

💭𝒀𝒐𝒖 𝒄𝒂𝒏 𝒔𝒆𝒆 𝒎𝒐𝒓𝒆 𝒅𝒆𝒕𝒂𝒊𝒍𝒔 𝒊𝒏 𝒕𝒉𝒆 𝒑𝒓𝒆𝒗𝒊𝒆𝒘 𝒊𝒎𝒂𝒈𝒆𝒔.

🎬🅼🆈 🆂🅴🆃🆃🅸🅽🅶🆂🎬

➥〔📸CLIP SKIP - 2 📸 〕

➥〔🎲CFG 4.5-5-5.5-6 - 𝗢𝗽𝘁𝗶𝗺𝗮𝗹🎲〕

➥〔🧼𝗘𝘂𝗹𝗲𝗿 𝗔, SGM Uniform - Karras🧼〕

➥〔👣𝟮𝟬-𝟮𝟴 𝗦𝘁𝗲𝗽𝘀👣〕

➥〔𝗛𝗶-𝗥𝗲𝘀 ♨️𝟰𝘅-𝗨𝗹𝘁𝗿𝗮𝗦𝗵𝗮𝗿𝗽 or 𝗙𝗼𝗼𝗹𝗵𝗮𝗿𝗱𝘆 𝗥𝗲𝗺𝗮𝗰𝗿𝗶♨️〕

➥〔𝗔𝗱𝗲𝘁𝗮𝗶𝗹𝗲𝗿 𝗳𝗼𝗿 𝗙𝗮𝗰𝗲 - 𝗘𝘆𝗲𝘀👁〕

➥〔𝗤𝘂𝗮𝗹𝗶𝘁𝘆 𝗣𝗮𝗿𝗮𝗺𝗲𝘁𝗲𝗿𝘀 𝗶𝗹𝗹𝘂𝘀𝘁𝗿𝗶𝗼𝘂𝘀: ✔️𝗺𝗮𝘀𝘁𝗲𝗿𝗽𝗶𝗲𝗰𝗲, 𝗯𝗲𝘀𝘁 𝗾𝘂𝗮𝗹𝗶𝘁𝘆〕✔️

➥〔Negative: 𝘄𝗼𝗿𝘀𝘁 𝗾𝘂𝗮𝗹𝗶𝘁𝘆, 𝗹𝗼𝘄 𝗾𝘂𝗮𝗹𝗶𝘁𝘆, 𝘄𝗮𝘁𝗲𝗿𝗺𝗮𝗿𝗸〕

➥〔𝑮𝒆𝒏𝒆𝒓𝒂𝒍 𝒔𝒕𝒚𝒍𝒆𝒔: 2.5d, anime screencap, bold line, sketch, cgi, digital painting, flat colors, smooth shading, minimalistic, ink style, oil style, pastel style〕

🔝▶️𝐓𝐡𝐞𝐬𝐞 𝐩𝐚𝐫𝐚𝐦𝐞𝐭𝐞𝐫𝐬 𝐚𝐫𝐞 𝐫𝐞𝐜𝐨𝐦𝐦𝐞𝐧𝐝𝐞𝐝, 𝐛𝐮𝐭 𝐲𝐨𝐮 𝐜𝐚𝐧 𝐜𝐡𝐨𝐨𝐬𝐞 𝐲𝐨𝐮𝐫 𝐨𝐰𝐧 𝐒𝐜𝐡𝐞𝐝𝐮𝐥𝐞𝐫. 𝐘𝐨𝐮 𝐜𝐚𝐧 𝐚𝐥𝐬𝐨 𝐚𝐝𝐣𝐮𝐬𝐭 𝐂𝐅𝐆 (𝐂𝐥𝐚𝐬𝐬𝐢𝐟𝐢𝐞𝐫-𝐅𝐫𝐞𝐞 𝐆𝐮𝐢𝐝𝐚𝐧𝐜𝐞) 𝐭𝐨 𝐚 𝐠𝐫𝐞𝐚𝐭𝐞𝐫 𝐨𝐫 𝐥𝐞𝐬𝐬𝐞𝐫 𝐞𝐱𝐭𝐞𝐧𝐭—𝐭𝐡𝐞 𝐬𝐞𝐭𝐭𝐢𝐧𝐠𝐬 𝐈 𝐮𝐬𝐞𝐝 𝐰𝐞𝐫𝐞 𝐞𝐱𝐚𝐜𝐭𝐥𝐲 𝐚𝐬 𝐥𝐢𝐬𝐭𝐞𝐝 𝐚𝐛𝐨𝐯𝐞.◀️

👉🅰️🅳🅳🅸🆃🅸🅾️🅽👈

✔️Baked into Model VAE.

✔️I advise not to use [𝗺𝗮𝘀𝘁𝗲𝗿𝗽𝗶𝗲𝗰𝗲, 𝗯𝗲𝘀𝘁 𝗾𝘂𝗮𝗹𝗶𝘁𝘆, 𝗮𝗺𝗮𝘇𝗶𝗻𝗴 𝗾𝘂𝗮𝗹𝗶𝘁𝘆] it for male characters, otherwise they will turn out feminine. In this case, use other quality triggers.