icon Υπότιτλοι
Φόρτωση υπότιτλων...

HiDream is here. NEW and UNCENSORED Image Model!

youtube μετάφραση youtube μεταφραστής youtube μεταγραφή youtube υπότιτλοι μετάφραση youtube στα ελληνικά youtube μετάφραση βίντεο youtube translate to greek translate youtube to greek youtube transcript to greek translate youtube video to greek

YouTube transcript, YouTube translate

32/32

A quick preview of the first subtitles so you know what the video covers.

This amazing local image model supports over 3800 different art styles. It's uncensored and currently it outranks flux and mid journey in version 6. I'm talking about high dream. In this video I'm going to show you how you can generate amazing images like this on your local PC and also how you can upscale your images to increase the details even further. Try chicken nugget superhero use one of the art styles mentioned earlier. B in style style of B one by one image and generate our image. I find it really funny chicken McNugget as superhero. You can see it took art style of B and yeah it actually applied it. Make sure to download your image otherwise it will be gone if you refresh your browser. Yeah, we are going to use Comfy UI. So, but the good news is there is a Windows installer. Now you can download it from this URL. Click on download. Select the version you need. And yeah, unless you have Nvidia 50 series card then you have to download the GitHub version. So in that case click install from GitHub and down here pixel Windows portable package. If you need help installing the portable version then watch this video here where I explain the basics and the installation in the end of the video. Make sure to run the latest version of comfy UI. So in order to do that, click on manager click update all. So after the update restart if travel updating confi UI then go to the update folder in your [Musik] confi UI update stable batch file. To replicate what I'm doing. Head over to Patreon download my workflows are for free. And then just drag them into comfy UI. High dream comes in three flavors full death and fast. Each needs less steps the last but basically trade between quality and speed. First grab the text encoders. Transform your prompts into signals for the image model. You may have already the T5 XXL from stable diffusion 3.5 or flux. Same goes for the vae that transforms images from the latent space to the pixel space.

Ρυθμίσεις

100%

Γλώσσα προορισμού μετάφρασης

🔊 Αναπαραγωγή ήχου
Αναπαραγωγή μεταφρασμένου ήχου