You don't have access to this class

Keep learning! Join and start boosting your career

Aprovecha el precio especial y haz tu profesión a prueba de IA

Antes: $249

Currency
$209
Suscríbete

Termina en:

0 Días
16 Hrs
54 Min
35 Seg

Parametrización con Midjourney práctica

14/20
Resources

How does chaos influence image generation?

In artificial intelligence imaging, understanding the chaos parameter is crucial. This parameter controls the level of randomness in the generated results. A low chaos, such as 20, offers images that show small variations, keeping closer to the original concept. However, a high chaos of 95 boosts the creativity of the artificial intelligence, leading the images to unexpected and different paths from the initial request. This parameter, which ranges from 0 to 100, allows creators to experiment intensively with variations and explore how far their imagination can go.

Examples of chaos in action

  • Chaos 20: generates variant images while retaining the main theme and message.
  • Chaos 95: images lead to unpredictable and creative results, such as people in unexpected scenarios.

What is stylized and how to optimize its use?

Stylized is a parameter that controls how much styling is applied to the generated image and is set in a range from 0 to 1000. Experimenting with stylized levels gives visibly different results, transforming the aesthetics of the image. A high value such as 700 gives a more visually appealing finish, while lower values offer greater visual simplicity. Opting for a mid-range allows you to balance the styling without losing sight of the original intent.

Why not always use the maximum stylized value?

By applying the maximum stylized value, images can lose clarity and draw attention to the wrong elements. So while the parameter offers flexibility, finding the perfect balance is key.

Tips for managing MidJourney effectively

Working in MidJourney offers multiple tools to optimize creative processes, from channel management to advanced settings, and here I'll share strategies for getting the most out of it.

Why create custom channels?

Having dedicated channels for different projects allows you to organize your creations without mixing them with other images and losing history. So, keeping your work by categories helps you easily go back to what you really liked or needs revisions.

How to use /info and /settings in MidJourney?

  • Command /info: provides an overview of your account, such as the number of images created and the remaining duration of your subscription. This way, you can plan and adjust your creation rate.
  • Settings /settings: Adjust parameters such as styling and chaos by default. Activating modifications like remix and using fast or relax according to your needs will give you more creative control.

How to choose the right aspect ratio?

Changing the aspect ratio of images generates optimal results depending on their purpose. For example, for a video, an aspect ratio of 16:9 is ideal and modern. Experiment with these dimensions for various projects and analyze the visual impact.

What are other important parameters in image creation?

In addition to chaos and stylized, several parameters enrich image creation with artificial intelligence. Researching, testing and adjusting these parameters can make a big difference in the final quality of your productions.

What is the function of weirdness?

The weirdness parameter, which ranges from 0 to 3000, introduces unusual elements into images. Although it serves to explore unconventional results, it is important to use it sparingly to maintain harmony and coherence in the final image. For example, an image generated with weirdness may acquire strange or dislocated characteristics.

When to use high stylized or weirdness values?

Raising these values can offer a unique and bold touch. However, it is recommended to reserve its use for situations where you want to break new ground or explore new stylistic directions. Using it constantly could result in outdated or excessively altered images.

Test and experiment: the key to learning

Trial and error with different parameters is essential. Although the first result may not be ideal, iterating with different settings will allow you to discover the richness that different combinations provide. When iterating, focus on:

  1. Small alterations: changing details in the prompt or subtly adjusting parameters can improve the results.
  2. Analyze and adjust: evaluate the results obtained and experiment with different combinations.
  3. Feedback and practice: sharing your creations and getting feedback broadens your perspective and provides new ideas for improvement.

At the end of the day, your creativity is your best resource. By following these tips, you will be able to create impressive and exciting images when working with artificial intelligence, so be encouraged to experiment and share your findings with the community!

Contributions 5

Questions 1

Sort by:

Want to see more contributions, questions and answers from the community?

Se entiende perfecto; las clases son espectaculares, pero si no tenemos suficiente presupuesto para los créditos no podemos hacer la practica con esa profundidad
### **Parámetros explicados en la página  “Midjourney Cheat Sheet - Parameter List and Examples”:** * **--aspect / --ar**: Cambia la relación de aspecto de la imagen. El valor predeterminado es 1:1. * **--stylize / --s**: Influye en la fuerza del estilo estético predeterminado de Midjourney en el trabajo. El valor predeterminado es 100, con un rango de 0 a 1000; valores más altos significan más estilización. * **--chaos**: Ajusta el nivel de variación en los resultados. El valor predeterminado es 0, con un rango de 0 a 100; valores más altos crean resultados más inusuales e inesperados. * **--quality / --q**: Determina la calidad de renderizado y el tiempo. El valor predeterminado es 1, con opciones como 0.25, 0.5, 1 y 2. La versión 6.1 introduce un nuevo modo --q 2 que toma un 25% más de tiempo y puede añadir más textura a costa de la coherencia de la imagen. * **--weird / --w**: Utiliza el parámetro experimental --weird para explorar estéticas inusuales. El valor predeterminado es 0, con un rango de 0 a 3000. * **--iw**: Peso de imagen; ajusta el balance entre los elementos de imagen y texto del mensaje. El valor predeterminado es 1, con un rango de 0 a 2.0; valores más altos enfatizan más la imagen. * **--sw**: Peso de estilo; ajusta la fuerza de la estilización al usar referencias de estilo. El valor predeterminado es 100, con un rango de 0 a 1000. * **--cw**: Peso de personaje; modifica la fuerza de la referencia de personajes, afectando cuánto impacta en la imagen final. El valor predeterminado es 100, con un rango de 0 a 100. * **--repeat / --r**: Crea múltiples trabajos a partir de un solo mensaje. El valor predeterminado es 1, con un rango de 1 a 40. * **--version / --v**: Usa diferentes versiones del algoritmo de Midjourney. La versión más reciente es la predeterminada, con opciones como 1, 2, 3, 4, 5.0, 5.1, 5.2 o 6. * **--niji**: Cambia a un modelo alternativo enfocado en imágenes de estilo anime. * **--no**: Mensaje negativo; por ejemplo, --no plants intenta eliminar elementos específicos de la imagen. * **--sref**: Usa una URL de imagen para hacer referencia a un estilo, aplicando sus elementos estéticos a la imagen generada. Se pueden usar varias URL y asignar pesos a cada referencia. * **--sref random**: Agrega un código de ajuste de estilo de más de mil millones de estilos base de forma aleatoria al mensaje. * **--fast**: Anula los ajustes actuales para ejecutar un solo trabajo en modo rápido. * **--relax**: Anula los ajustes actuales para ejecutar un solo trabajo en modo relajado. * **--seed**: Número de semilla para generar el ruido visual de la cuadrícula inicial de la imagen, con un rango de 0 a 4294967295. * **--stop**: Detiene el trabajo en un cierto porcentaje de finalización, con un rango de 10 a 100. * **--tile**: Genera imágenes adecuadas para mosaicos continuos. * **--turbo**: Anula los ajustes actuales para ejecutar un solo trabajo en modo turbo.
He sido un poco friki y he realizado captura del prompt utilizado por el profesor, pasando por chatgpt que es: Young Latina woman sitting gracefully in a breathtaking Andes mountain landscape, Colombian highlands in the background with lush green valleys and rugged peaks, soft golden sunlight illuminating her natural features, wearing casual and culturally inspired clothing with earthy tones, serene and contemplative expression, gentle breeze moving her hair, vibrant details of native flora and rustic textures around her, cinematic composition, high-definition photography, atmospheric and warm tones capturing the essence of the Colombian Andes Y en este caso a mi me gusta utilizar REPLICATE AI y ahí utilizar black-forest-labs/flux-1.1-pro-ultra ![](https://i.postimg.cc/wBRMGS0N/latina-1.png) ![](https://i.postimg.cc/Wp8NXNTK/latina-2.png)
![](https://static.platzi.com/media/user_upload/78d076a7-1c81-4321-884f-2123a3eb973a-e2d882b5-ec58-4dfa-838e-ab028a889e0e.jpg) yo segui iterando y creando versiones con la pagina de prompt hero se puede entender mucho los prompts y la generacion y la estructura del prompt y como debe ser muy especifico para poder expresar lo que quieres
Un tal random ![](https://static.platzi.com/media/user_upload/image-e17b8d15-30ab-4690-8619-952dd6c057a0.jpg)