Генератор отрицательных подсказок Leonardo AI
К сожалению не всегда потрясающие образы, которые создаёт Leonardo AI выглядят немного не так.
Попрощайтесь с уродствами, лишними пальцами, размытыми городскими пейзажами, и странно выглядящими ассиметрично выглядящими людьми.
С этим промптом для ChatGPT у вас больше не будет этих проблем.
Сначала вы опишете в каком стиле вы хотите, чтобы были созданы подсказки, например фотореалистичные, портретные пейзажные и так далее...
Затем вы предоставите тип изображения, которое желаете получить, например льва в джунглях или современную женщину в футуристическом костюме.
Затем ChatGPT спросит, желаете получить серию отрицательных ключевых слов для использования в строке Negative prompts.
И все, что вам нужно будет сделать это сказать да - "yes"
И все это в одном комплекте. Вот почему это называется двойной генератор.
Отправляем полностью весь этот скрипт в ChatGPT 🔻
I want you to become the best prompt generator in the world for Leonardo AI. Leonardo AI is an artificial intelligence program similar to Midjourney. Leonardo AI generates images based on given prompts.
Your job is to engineer the most amazingly detailed and creative descriptions that will craft unique, eye-catching realistically rendered, and interesting images. I will provide you basic information required to make a Stable Diffusion prompt.
Here are some important points to note in order to make this masterful Leonardo AI prompt.
-Examples of camera shot types are long shot, close-up, POV, medium shot, extreme close-up, and panoramic. Camera lenses could be EE 70mm, 35mm, 85 mm, 135mm+, 300mm+, 800mm,
-Example Cameras to use are: Nikon D850, Canon EOS R6 Mark II, Canon EOS R5, LENS: Sigma 50mm f/1.4 DG HSM Art Lens, FILM STOCK / RESOLUTION: High Resolution
- Helpful keywords related to resolution and detail are: 4K, 8K, 64K, 3D Rendering, photorealistic, photography realistic, hyperrealistic, detailed, highly detailed, high resolution, hyper-detailed, HDR, UHD, professional, realistic rending, etc.
Examples of renders are Octane render, cinematic, low poly, Unreal Engine 5, Unity Engine
I will provide you with a command to create one of the following:
/photorealistic, niji 5, unreal engine, octane render, photorealism
The command I send will be followed by a keyword and you will generate a style of prompt with as many details as provided in the prompt structure. For example, my command can look like this: /person a modern woman or /scenery a futuristic city
Your first response will be a question back to me asking what you want my prompt to be about.
Based on my input, you will then start the process. You will write the prompt in a simple format. This is important: The prompt will be no more than 1,000 characters.
After you have provided me with my prompts, you will then ask me if I would like you to generate a list of negative prompt keywords for me to use that correspond to my chosen command.
If I say "yes" you will start the negative prompt keyword process. Negative prompt keywords are keywords or descriptions that may make an image look distorted, unrealistic, blurry, disfigured, etc. Here are the negative prompts to use for each /command. You can also add additional negative prompts that are similar to the ones I am providing you with.
The following are the negative keywords for each command.
/person: plastic, Deformed, blurry, bad anatomy, bad eyes, crossed eyes, disfigured, poorly drawn face, mutation, mutated, extra limb, ugly, poorly drawn hands, missing limb, blurry, floating limbs, disconnected limbs, malformed hands, blur, out of focus, long neck, long body, mutated hands and fingers, out of frame, blender, doll, cropped, low-res, close-up, poorly-drawn face, out of frame double, two heads, blurred, ugly, disfigured, too many fingers, deformed, repetitive, black and white, grainy, extra limbs, bad anatomy
/scenery:blurry, boring, close-up, dark (optional), details are low, distorted details, eerie, foggy (optional), gloomy (optional), grains, grainy, grayscale (optional), homogenous, low contrast, low quality, lowres, macro, monochrome (optional), multiple angles, multiple views, opaque, overexposed, oversaturated, plain, plain background, portrait, simple background, standard, surreal, unattractive, uncreative, underexposed
/cityscapes: asymmetrical buildings, blurry, close-up, creepy, deformed structures, grainy, jpeg artifacts, low contrast, low quality, low-res, macro, multiple angles, multiple views, overexposed, oversaturated, plain background, scary, solid background, surreal, underexposed, unreal architecture, unreal sky, weird colors
/portrait: plastic, Deformed, blurry, bad anatomy, bad eyes, crossed eyes, disfigured, poorly drawn face, mutation, mutated, extra limb, ugly, poorly drawn hands, missing limb, blurry, floating limbs, disconnected limbs, malformed hands, blur, out of focus, long neck, long body, mutated hands and fingers, out of frame, blender, doll, cropped, low-res, close-up, poorly-drawn face, out of frame double, two heads, blurred, ugly, disfigured, too many fingers, deformed, repetitive, black and white, grainy, extra limbs, bad anatomy, extra fingers, fused hand, bad proportions, out of focus, unnatural body, unnatural skin
/photorealistic: 3D render, aberrations, abstract, anime, cartoon, collapsed, conjoined, creative, drawing, extra windows, harsh lighting, illustration, jpeg artifacts, low saturation, multiple levels, overexposed, oversaturated, painting, photoshop, rotten, sketches, surreal, twisted, UI, underexposed, unnatural, unreal engine, unrealistic, video game
/illustration: 3d, bad art, bad artist, bad fan art, CGI, grainy, inaccurate sky, inaccurate trees, kitsch, lazy art, less creative, low-res, noise, photorealistic, poor detailing, realism, realistic, render, stacked background, stock image, stock photo, text, unprofessional, unsmooth
When you provide the negative prompt keywords, do not list them with bullet points. Please list them in paragraph form, each keyword separated by a comma.
PROMPT 2 (дополнение): WRITE THE PROMPT OUT IN A SIMPLER FORMAT
Этот дополнительный промпт #2 может потребоваться для сокращения полученного результата, если созданная подсказка оказалась слишком большой (больше 1000 знаков).
ЗАКЛЮЧИТЕЛЬНАЯ КОМАНДА ДЛЯ ОТРИЦАТЕЛЬНЫХ ПОДСКАЗОК: please convert the bulleted negative keywords into comma-separated
После ввода основного промта мы получим запрос от чата:
Absolutely, I understand your request. Please tell me which prompt you would like to start with. Would it be /person, /scenery, /cityscapes, /portrait, /photorealistic, or /illustration? And what specific keyword or theme would you like to be included in the prompt?
Абсолютно, я понимаю вашу просьбу. Пожалуйста, скажите мне, с какой подсказки вы хотели бы начать. Будет ли это / человек, / пейзаж, /городской пейзаж, /портрет, /фотореалистичный или /иллюстрация? И какое конкретное ключевое слово или тему вы хотели бы включить в приглашение?
И здесь важно использовать символ косой черты в качестве команды и далее через пробел следует указать объект для генерации
/photorealistic, niji 5, unreal engine, octane render, photorealism
И так. В данном примере я укажу команду, которая касается человека, затем через пробел вводите все что считаете нужным.
/person woman hiking in the mountains
Как и предполагалось, чат выдал слишком большую подсказку.
Поэтому следует отправить в чат команду
WRITE THE PROMPT OUT IN A SIMPLER FORMAT
и в результате получим короткий промпт
Woman hiking on mountain trail, medium shot, Canon EOS R6 Mark II, Sigma 50mm lens, 8K HDR resolution, golden hour lighting, stylish outdoor gear, lightweight backpack, snowy peaks, lush green slopes, clear blue sky, crystal-clear stream.
Далее для получения негативного промпта отправим в чат слово
И вот какую негативную подсказку мы получили
For a person prompt, especially a woman hiking in the mountains, you may want to avoid the following to maintain the quality and coherence of the image:
plastic, deformed, blurry, bad anatomy, bad eyes, crossed eyes, disfigured, poorly drawn face, mutation, mutated, extra limb, ugly, poorly drawn hands, missing limb, floating limbs, disconnected limbs, malformed hands, blur, out of focus, long neck, long body, mutated hands and fingers, out of frame, blender, doll, cropped, low-res, close-up, poorly-drawn face, out of frame double, two heads, blurred, too many fingers, deformed, repetitive, black and white, grainy, extra limbs, bad anatomy.
Вот результат без негативной подсказки
Вот результат без с использованием негативной подсказки
↩️ Вернуться на Главную