Hello Dear Shells, Knowledge Lovers!
We would like to show you how easy it is to create amazing images using our diverse set of tools, using the example of the **Hiroshinagai Style bot! **
And so, having come up with the scene, style, keywords and essence of the image we can ask GPT to help us compose a Prompt for our image!
My message to GPT:
Hi GPT! I'm using a Hiroshi Nagai-trained bot to generate images.
Can you help me compose a Prompt to generate? I want to see the following: "Cinematic image, beautiful girl, standing on the slope of a mountain. The girl is standing half-sideways. She is tall and slender, her face illuminated by the moonlight".
Also use these keywords "Movie shot style, ultra detailed, masterpiece, best quality, girl in black elegant dress, moon, stars, mountains".
Prompt:
Generate a cinematic masterpiece in the style of a movie shot, ultra-detailed, of the highest quality. The scene is set on the slope of a mountain under a night sky. A beautiful, tall, and slender girl stands in a poised half-sideways position, her elegant black dress flowing with the contours of the landscape. Her face is softly illuminated by the gentle glow of the moonlight, enhancing her serene expression. The moon hangs prominently in the sky, surrounded by a tapestry of stars, casting a silver light over the mountains and creating a harmonious contrast between the girl.
Negative Prompt:
Smudged or illegible face, twisted fingers, mutations
First, let's understand what a Prompt is:
Simply put, a prompt can be considered a guiding question, suggestion, or instructions aiming to stimulate the bot to generate a corresponding response or creation.
So how do you create a prompt that will yield the best generation result?
- Be specific: Describe in detail the type of image you want to create. Specify objects, the scene, mood, and other important details.
- Use clear instructions: Specify any specific features that should be included or excluded from the image. For example, colors, composition, style, etc.
P.S. Don't forget, you can always use the help of another tool-bot to achieve your goals
First and second link if you want to know more.
Now that we have somewhat covered Prompts, what is a Negative Prompt?
A Negative Prompt is an addition to a Prompt that specifies what should NOT be included in the image. It indicates elements or aspects that you would not like to see in the generated image. For example, "Do not include people in the image" or "Exclude insects from the image." This helps to avoid unwanted details and clarify your expectations.
IMG2IMG
Shall we spice up the monotony with some image?
I think this example clearly illustrates what is IMG2IMG
Prompt strength when using img2img
Determines the degree of change from the initial image, so the lower the value, the weaker the change to the original image and vice versa.
Num Inference Steps/Number of denoising steps
Responsible how many times (iterations) the bot should improve the image before completing the generation. The higher the value, the higher the quality of the final image.
**Guidance scale/Scale for classifier-free guidance **
Determines how closely the bot will follow your prompt/instructions. The lower the value of scale, the more creative bot will be, and vice versa.
Lora Scale
Parameter that determines how much the generation will be altered according to the bot used, in our case, how similar the image will be to the style of the artist Hiroshi Nagai.
Such titles as "scheduler," "refine," and their corresponding parameters deserve a separate article. Try experimenting with them yourself and see how the results change!
Thx for sharing! just tried this widget after reading it. Looking forward to the next guide on detailed parameters



