Dream4.0 has recently been on the line, and many calls have been stronger than banana
In addition to capacity enhancement, a number of functions have been added to 4.0
Select Edit Area
Smart Dimensions
Multigraph Upload
Let's see what dream 4.0 is capable of
Contents
01. Listed editing additions, deletions
Selected regional editors
03. Smart size to generate public cover
04. Make a little red book and cover it
The creation of characters
06. Cameras point to generation
Image generation manual
08. Generation of Four-Master Cartoons
Generating emoticons
10. Comic-turning real people
11. person p-chart
12. Product migration
14. Style migration
15. Forecast future appearance
16. Knowledge labelling
images of this article are produced by ai and written manually
Edit & Delete
Upload a picture, add or delete it by simple command
Common routines, such as adding, deleting, changing the background, going to text and changing the text, are also performing well, and simple instructions can do a good job
The new smart size and "specify the editor" features of this update are also very useful
Select Edit Area
Specifies where to edit, and when uploading a picture, you can paste and frame the picture
It's a very useful function, which can be better achieved if you can't handle simple commands
For example, it's been a few unsuccessful attempts to remove the flowers from the model's clothes, but then the box can remove them
But there'll still be some flaws here, removing the clothes from the arm after removing the flowers
Smart Size Generates Public Cover
This size scale allows you to select intelligence and automatically generate the right sizes according to the context and use of the command
Prompt word: remove the word “vogue” and add to the left side of the image the art word “i.e. dream 4.0” “super-strength assessment”, in two rows, for the public-name cover
It worked very well, and the instructions were completed with precision
Little Red's on the cover
try to make a little red book cover, take clothes out of a model and make a ootd cover
psych: make a 3:4 cover, remove the text of the picture, add "otd" uppercase english above, do not change models, remove clothes, bags, shoes, and lay them on the right
It is important to note here that the little red book should be a forbidden word, directly to create a small red book cover, showing that "the words entered do not conform to the Platform's rules" and are deleted before they are properly generated
The character is created
The ability to be consistent on the subject of dream 4.0 is greatly enhanced
Indictment: Put the girl in a hood, change the background to a red wall brick, and keep her face the same
People's faces are very consistent. They can be used for writing
It's hard to create a hair when you're doing a real picture, so you can upload a reference to make up
Multigraph Upload
Dream 4.0 also updates a new capability, which is to upload multiple maps so that it can upload master maps and multiple reference maps
Psych: Let the girls in figure I take a look at the makeup in figure 2, put on a Tibetan costume, make up a Tibetan make-up, with a background in the grassland, with a swab, a real picture
The lens points to generation
Try to turn people around, face-to-face, or you can do that
Picture Generation Manual
And before it was hot, I could do it
PLEASE MAKE THIS CARTOON IMAGE A PLASTIC MODEL. BEHIND THE MODEL, A COMPUTER SHOWS ITS MODEL, NEXT TO A MODEL BOX WITH THIS PICTURE ON IT. IN FRONT OF THE COMPUTER, ON A ROUND PLASTIC BASE, PUT ON THE MODEL I GAVE YOU. THE MODEL POSITION REMAINS UNCHANGED. IT IS HOPED THAT THE PVC MATERIAL WILL BE CLEARLY MARKED. BACKGROUND IN THE ROOM。
By comparison, it can be seen that the scale is not quite right. The model is somewhat too large, but the details are clear
Generates four-megagram comics
It's good to create a four-gauge comic, but there's still a few flaws, and most of them are not good, and they need a draw card
Generate emoticon bag
Psych: Generates four different facebags for girls, asking for fun, humor, impeccable, cute
The face bag produces multiple multiuterine expressions
Now you can generate four different images in one-time, and the images don't repeat
I'm a human
personal p-chart
Product migration
Replace the pillows in figure 2 with the pillows in figure 1 and keep the scene consistent
style migration
TRY USING A REFERENCE MAP TO MAKE A PPT COVER OF THE SAME STYLE
MAKE A PPT COVER BASED ON THE STYLE OF THE PICTURE, AND ASK FOR A BIG TITLE ENTITLED "THE POWER OF DREAM 4.0" AND A SUBTITLE ENTITLED "HI-HAI-I-I-I-I-I-I-I."
Turn Figure 1 into Figure 2 style, keep the characters the same
Image reasoning
You can identify images and understand their content in depth before they can be generated
For example, it can be extrapolated in time. It can be extrapolated from photos
Forecast future appearance
30 years later
It also includes a knowledge-based ability to label pictures in terms of words
Knowledge label
Mark the name of the flower in the picture and give a simple introduction to the bottom left corner
Fill out the words in the picture
That's the test of dream 4.0, and it's really fun to use it as a whole. It's a lot of fun to use
With the recent increase in the number of users, the rate of generation has slowed。
Interested can open a dream。