¡Midjourney v4 está aquí y es lo mejor! / Midjourney v4 is here and it's the best!

Hola amigos, desde que salió midjourney (una IA competencia de dall-e 2) siempre hubo mucha expectativa al respecto ya que fue una de las primeras en hacerse pública y de libre acceso al público, sin embargo sus resultados no eran los mejores. Aun así con todo y eso fue aprovechada para un concurso de arte, y aunque ganó debido a que los jueces desconocían su origen demostró ser una de las IAs más creativas del mercado.

Pues hace poco midjourney ha dado un gran salto liberando la versión 4 de su algoritmo que incluye varías mejores visuales. Para esto les mostrare un prompt que ya también había usado con dall-e hace un tiempo y lo probé con las 4 versiones del algoritmo para mostrarles las mejoras. El prompt ha sido "a frog wearing a witch hat doing magic in front of a cauldron --v 1" el comando "--v 1" sirve para que genere imágenes con una versión especifica en este caso la versión 1 y como podemos ver los resultados son terribles.

Ahora con la versión 2 vemos un pequeño esbozo de lo que queremos pero aun no.

Con la versión 3 ya se ven un poco más de mejoras y podemos ver la figura de la rana frente al caldero.

Y finalmente la versión 4, donde los resultados son muy coherentes y visualmente atractivos.

Como verán los resultados con esta nueva versión son muy buenos. A continuación algunas imágenes que he hecho con este nuevo algoritmo.

"a cat driving a spaceship --v 4"

"a girl wearing earrings in the shape of a full moon, realistic --v 4"

"a boy holding a glowing pill surprised, photorealistic style --v 4"

"a man sitting in front of a campfire watching an alien spaceship in the night sky --v 4"

"an astronaut floating in a spaceship --v 4"

"a giraffe having tea with Queen Elizabeth --v 4"

"a giant turtle in space carrying the earth in its shell --v 4"

"a giant cat leaning on the eiffel tower --v 4"

"a man looking at the moon who has an evil face --v 4"

"a ballet dancer dancing in the hell --v 4 --chaos 10"

y bueno hasta aquí el post de hoy, espero que les haya gustado y si les gustó apóyenme con un voto arriba, hasta un próximo blog.


English

Hi friends, since midjourney (an AI competition of dall-e 2) was released there was always a lot of expectation about it since it was one of the first to be made public and freely available to the public, however its results were not the best. Even so, it was used for an art contest, and although it won because the judges did not know its origin, it proved to be one of the most creative AIs in the market.

Recently midjourney has taken a big leap by releasing version 4 of its algorithm that includes several better visuals. For this I will show you a prompt that I had already used with dall-e some time ago and I tested it with the 4 versions of the algorithm to show you the improvements. The prompt has been "a frog wearing a witch hat doing magic in front of a cauldron --v 1" the command "--v 1" is used to generate images with a specific version in this case version 1 and as we can see the results are terrible.

Now with version 2 we see a little sketch of what we want but not yet.

With version 3 we can see a little more improvements and we can see the figure of the frog in front of the cauldron.

And finally version 4, where the results are very coherent and visually attractive.

As you will see the results with this new version are very good. Here are some images I have made with this new algorithm.

"a cat driving a spaceship --v 4"

"a girl wearing earrings in the shape of a full moon, realistic --v 4"

"a boy holding a glowing pill surprised, photorealistic style --v 4"

"a man sitting in front of a campfire watching an alien spaceship in the night sky --v 4"

"an astronaut floating in a spaceship --v 4"

"a giraffe having tea with Queen Elizabeth --v 4"

"a giant turtle in space carrying the earth in its shell --v 4"

"a giant cat leaning on the eiffel tower --v 4"

"a man looking at the moon who has an evil face --v 4"

"a ballet dancer dancing in the hell --v 4 --chaos 10"

And well so far today's post, I hope you liked it and if you liked it please support me with a vote above, see you next blog.

Translated with www.DeepL.com/Translator (free version)

image.png
Imagen hecha por @fclore22