Media is too big
VIEW IN TELEGRAM
Google announced VideoPoet, a language model for video generation.
The AI is able to convert text into video, as well as edit, voice and stylize.
It looks pretty good. Waiting for the release.
The AI is able to convert text into video, as well as edit, voice and stylize.
It looks pretty good. Waiting for the release.
👍5❤3
This media is not supported in your browser
VIEW IN TELEGRAM
PixelLab - pixel art from sketches
A neural network running on the discord platform that upgrades a rudimentary sketch to a high-quality pixel-art image. There's a handy panel for editing the results.
Here you can create landscapes, characters, stylized maps and individual objects in motion. Up to 20 generations are free. More - by subscription from $9 per month.
#service
A neural network running on the discord platform that upgrades a rudimentary sketch to a high-quality pixel-art image. There's a handy panel for editing the results.
Here you can create landscapes, characters, stylized maps and individual objects in motion. Up to 20 generations are free. More - by subscription from $9 per month.
#service
👍1
👍5
The alpha version of the Midjourney V6 is out
What we noticed:
MJ now generates text.
The neural network has "forgotten" some celebrities. For example, Florence Pugh, Jenna Ortega or Henry Cavill are unrecognizable.
Wes Anderson's style works even better.
In some moments, the model is very pumped (private examples: Shrek looks like Shrek, and Rick Sanchez has a monobrow). The Pixar style has become less Pixar.
MJ's strong suit is photorealistic images with a bias towards faces. Neurophotos are sometimes impossible to distinguish from real photos! Every eyelash and every vasculature in the eye is drawn (but sometimes pupil shoals come out). MJ has moved away from neural net beauties and now more often generates unconventionally attractive people and adds freckles, skin imperfections, etc.
#news #midjourney
What we noticed:
MJ now generates text.
The neural network has "forgotten" some celebrities. For example, Florence Pugh, Jenna Ortega or Henry Cavill are unrecognizable.
Wes Anderson's style works even better.
In some moments, the model is very pumped (private examples: Shrek looks like Shrek, and Rick Sanchez has a monobrow). The Pixar style has become less Pixar.
MJ's strong suit is photorealistic images with a bias towards faces. Neurophotos are sometimes impossible to distinguish from real photos! Every eyelash and every vasculature in the eye is drawn (but sometimes pupil shoals come out). MJ has moved away from neural net beauties and now more often generates unconventionally attractive people and adds freckles, skin imperfections, etc.
#news #midjourney
👍6