DALL-E 2 – Day 27: Playing with Variations part 1

This video series shows how DALL-E 2 works. This program is in the research phase and comes to you from www.openai.com

Please comment on prompts you want to see and I will include them in future videos! Thanks for watching 🙂

Check out the best of the best created with DALL-E 2 on my instagram
www.instagram.com/jeremiahpetersen

#dalle #dalle2 #openai #artificialintelligence #ai

21 Comments

  1. I was looking at postage stamps made in Dall E the other day and they look great. Now imagine what paper money (bills/currency/notes/legal tender) from other planets or imaginary countries would look like! or even from fantasy universes like Studio ghibli or Adventure Time!

  2. you are almost 1k subs i wish u best of luck, thanks for showing us all these dall e images, but soon it will get bore. you can try doing some research onn art and try to explain that, so it will be more educational

  3. I would like to see variations of famous album covers and variations of like a random picture of someone’s kitchen or house. How does it make variations on real life, make it feel like an alternate universe.

  4. If only DALL-E could use the previous text that generated the image or a custom descriptor to help guide the variations. Since they get more abstract and further away from the original premise each time.

  5. Prompt Idea for the upcoming car video! A Mashup of a Glickenhaus SCG 003s and Brabham BT62. I have no idea how to word my prompt to get the 2 cars to mix correctly but love the idea of mixing car designs. I'm full with ideas for car mixing

  6. This prompt worked for me well in midjourney but sadly I don't have DALL-E 2. Could you please try "A small mountain river flowing rapidly through a late autumn forest. A proud stag, standing with its head high, is silhouetted by the sunbeams through the trees, cinematic, cgsociety, Brian froud, greg rutkowski , vray render, detailed, craig mullins style, early morning sun, realistic, atmospheric, lively"

  7. Have you tried recursively generating variations by using the first variation of the last output as the next input image? I think it would be wild to show the start and end images and then the sequence of 50 intermediates. I'm guessing they will be very different

Leave a Reply

Your email address will not be published. Required fields are marked *

© 2024 AI Art Video Tutorials