Rules for Using Downloadable Models:
1. Local Use (Allowed ✅)
You may use the models for local generation and personal (home) use.
2. Merging (Forbidden ⛔)
You are strictly prohibited from using these models for merging!
3. Distribution & Penalties (⚠️)
Do not re-upload them to other platforms, under different names, etc.
- Consequence of violation: If this occurs, no future models or requests will be made available for download. (Online generation will only remain possible via the Seaart website).
4. License Check (ℹ️)
Always check the "License Scope". Verify permitted and restricted uses for each specific model before use.
___________________________________________________________________________________________________________
🎬🅼🆈 🆂🅴🆃🆃🅸🅽🅶🆂🎬
➥〔📸CLIP SKIP - 2 📸 〕
➥〔🎲CFG 3,5-4-4.5-5-5.5-6 - 𝗢𝗽𝘁𝗶𝗺𝗮𝗹🎲〕
➥〔🧼𝗘𝘂𝗹𝗲𝗿 𝗔, 𝗥𝗲𝘀𝘁𝗮𝗿𝘁, 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〕
🔝▶️𝐓𝐡𝐞𝐬𝐞 𝐩𝐚𝐫𝐚𝐦𝐞𝐭𝐞𝐫𝐬 𝐚𝐫𝐞 𝐫𝐞𝐜𝐨𝐦𝐦𝐞𝐧𝐝𝐞𝐝, 𝐛𝐮𝐭 𝐲𝐨𝐮 𝐜𝐚𝐧 𝐜𝐡𝐨𝐨𝐬𝐞 𝐲𝐨𝐮𝐫 𝐨𝐰𝐧 𝐒𝐜𝐡𝐞𝐝𝐮𝐥𝐞𝐫 etc... - [i will not recommended do this].
𝐘𝐨𝐮 𝐜𝐚𝐧 𝐚𝐥𝐬𝐨 𝐚𝐝𝐣𝐮𝐬𝐭 𝐂𝐅𝐆 (𝐂𝐥𝐚𝐬𝐬𝐢𝐟𝐢𝐞𝐫-𝐅𝐫𝐞𝐞 𝐆𝐮𝐢𝐝𝐚𝐧𝐜𝐞) 𝐭𝐨 𝐚 𝐠𝐫𝐞𝐚𝐭𝐞𝐫 𝐨𝐫 𝐥𝐞𝐬𝐬𝐞𝐫 𝐞𝐱𝐭𝐞𝐧𝐭—𝐭𝐡𝐞 𝐬𝐞𝐭𝐭𝐢𝐧𝐠𝐬 𝐈 𝐮𝐬𝐞𝐝 𝐰𝐞𝐫𝐞 𝐞𝐱𝐚𝐜𝐭𝐥𝐲 𝐚𝐬 𝐥𝐢𝐬𝐭𝐞𝐝 𝐚𝐛𝐨𝐯𝐞.◀️
Description
🧬𝑭𝒖𝒍𝒍 𝑫𝒆𝒔𝒄𝒓𝒊𝒑𝒕𝒊𝒐𝒏:🧬
The model is based on: Special [ProjectQ_Base Model]
Special Thanks to: @waikitty @HARU_owo
✍️【𝘼𝙙𝙙𝙞𝙩𝙞𝙤𝙣】: If you have any questions, please don't hesitate to write!✒️
▁ ▂ ▄ ▅ ▆ ▇ █ 📜𝑴𝒐𝒅𝒆𝒍 𝑫𝒆𝒔𝒄𝒓𝒊𝒑𝒕𝒊𝒐𝒏:📜 █ ▇ ▆ ▅ ▄ ▂ ▁
[EN]
❝This is my first model that fully embraces an anime style. It includes special effects, features enhanced overall quality, and possesses a deep color palette. The model demonstrates good prompt comprehension, though it may interact negatively with some LoRAs due to being a V-Pred model.
I forgot to add that the model is modified and fine-tuned.❞
[RU]
❝Это моя первая модель, содержащая в себе полностью аниме стиль, добавлены эффекты, улучшено общее качество, имеет глубокую цветовую палитру, хорошо понимает подсказки, но может негативно взаимодействовать с некоторыми LoRA, так как модель V-Pred.
Забыл добавить, что модель является модифицированной и до-обученной Fine-Tuned❞
🎬🅼🆈 🆂🅴🆃🆃🅸🅽🅶🆂🎬
➥〔📸CLIP SKIP - 2 📸 〕
➥〔🎲CFG 4.5-5-5.5-6 - 𝗢𝗽𝘁𝗶𝗺𝗮𝗹🎲〕
➥〔🧼𝗘𝘂𝗹𝗲𝗿 𝗔, 𝗥𝗲𝘀𝘁𝗮𝗿𝘁, SGM Uniform - Karras🧼〕
➥〔👣𝟮𝟬-𝟮𝟴 𝗦𝘁𝗲𝗽𝘀👣〕
➥〔𝗛𝗶-𝗥𝗲𝘀 ♨️𝟰𝘅-𝗨𝗹𝘁𝗿𝗮𝗦𝗵𝗮𝗿𝗽 or 𝗙𝗼𝗼𝗹𝗵𝗮𝗿𝗱𝘆 𝗥𝗲𝗺𝗮𝗰𝗿𝗶♨️〕 - i didn't use for this model.
➥〔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 - [will not recommended do this]. 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.