This is how I do it …
“ Micro tile collage with cats and dogs in each tile wide image”
3 Likes
Ah the whole collage is generated. I thought it was made from single images.
1 Like
Idk how they do it, this is just me
20x20 Micro tile collage with cats and dogs in each tile wide image
I’m trying to figure out how to control grid size.
2 Likes
20x20 is something the image generator can’t do
1 Like
I know but I’m trying to get it to try, so it makes the squares smaller.
30x20 Micro collage with cats and dogs in each tile wide image
But I can’t figure out how to get it out of default
2 Likes
My collage was actually put together via good old Picasa from individual images.
2 Likes
My collage is single images and a c# script.
But you can easy create collections, it is useful to see for example a style. instead to make 10 pictures, i make 1 collage with many. don’t work perfect, but i do this mainly for testing.
1 Like
I think a website with a catalog of styles and example pictures would be a good thing to have to see what is possible and how to make it.
Anyone know a good one?
1 Like
I search for it but not really found any. the thing is too, that DallE not always follows all advice. it is very dependent what image you create and how strong a style is distinguishable, and how much training data is there for it, and if it is in the vocabulary DallE understand. Best you search a art website with styles and try them out. Water color for example works always, but other styles not. the styles must be in the vocabulary for the training data, and it is dependent if the system can detect and mark them (recap) or humans do the training-support.
You can create a post with the findings.
PS: My experiences are the prompts must be simple. otherwise the collage trick not work anymore (or it needs a other description then what i used).
1 Like
The most useful web resources I have found is the main Dalle pages on this forum. @PaulBellow and @Daller’s threads they are the closest I found to an image / prompt catalog.
3 Likes
I think DALL-E is one of the most profound models, if not the most profound, in terms of the artistic rendering of generated images. It has certainly been well-trained on images related to oil paintings, chalk on blackboard, watercolor, drawing, children’s illustrations, comics, and cartoons. The fantasy illustrations that can be achieved with DALL-E are extraordinary: you can create wonderfully expressive characters. DALL-E also excels at landscape representations that don’t rely on photo-realistic techniques. Highly detailed faces can be produced, as long as one isn’t seeking photo-realism in the rendering.
8 Likes
Prompt
A high-definition photograph of a man struggling to hold an enormous, young Golden Shepherd puppy (a mix of Golden Retriever and German Shepherd), which is so large it’s almost too heavy for him to lift. The puppy has a fluffy coat, friendly eyes, and giant playful paws, with an innocent, curious expression, and no extra paw. The man looks amused yet overwhelmed as he attempts to hold the massive puppy, standing outdoors with green natural scenery. The image is warm, humorous, and surreal, emphasizing the playful contrast between the giant puppy and the man’s effort.
or making things small out of things 
*ignore the man’s eyes. i had Dall-E do a couple of edits trying to fix the eyes but i think it’s like when you blur or smudge stuff in Photoshop that’s already been blurred or smudged 
5 Likes
Hah! I need to phrase it better maybe! 
Like things made out of match sticks or marshmallows or other “small objects”… Came up in the Halloween thread, I think?
9 Likes
Things made out of little things…
Image(Widescreen, RandomScene(), MadeOutOf(Matchsticks))
Image(Widescreen, Scene(a homage to little things), MadeOutOf(Matchsticks))
7 Likes
Phor(0-3, Image(Widescreen, RandomScene(), MadeOutOf(RandomThing())))
Phor(0-3, Image(Widescreen, RandomScene(), MadeOutOf(RandomThing([Item]))))
SUCCESS!!!
3 Likes
Do you ever feel, looking at these pictures… A little sad that these things don’t actually exist?
1 Like
Yes that’s why I want to build them. I love physical models. Game minis , cars, robots,
I build frame arm and other Japanese models too. I been building models since I was a kid. This is me in my “lad” reflected in the mirror behind my stack of ones I need to build. I build one a week lol.
4 Likes
OK guys you gotta think about this one before you check out the picture…
Think about what you are expecting to see based on the Phasm prompt before you look.
Did you see what you expected?
(Translated Prompt and Summary at end of images)
Phor(0-3, Image(Widescreen, Scene(Waves), MadeOutOf(Time([Item]))))
Age 0
A detailed widescreen scene featuring ocean waves, each wave creatively formed to represent the concept of time, specifically from a 0-year old perspective. The waves are gentle, calm, and translucent, capturing an early stage of time. In this image, the colors are soft and muted to evoke a sense of beginning.
to Age 3
A detailed widescreen scene featuring ocean waves, each wave creatively formed to represent the concept of time, specifically from a 3-year old perspective. The waves are energetic and full of motion, with a sense of adventure and joy, showing the vivid imagination of a young child. Colors are bright and lively, capturing the energy of youthful exploration and wonder.
All four images representing the concept of time as waves from perspectives of 0 to 3 years old are displayed. Each captures the evolving qualities of time and experience from gentle beginnings to dynamic exploration. Let me know if you’d like further adjustments.
1 Like