What are some good style option choices for Dall-e 2?

Prompting
, ,
1

I am trying to find more style option generations for dalle2- currently I am using the dalle2 api and feeding it prompts based on a list of style options which user will be able to select.

Here are style options I have right now currently:
“animation”,
“cartoon”,
“detailed”,
“drawing”,
“high-resolution”,
“modern”,
“oil painting”,
“pointillism”,
“professional photography”,
“traditional”,
“vector art”,
“watercolor”,
“3D”

What are some other options I could add to this list that work well in generating unique styles for images?

2

Hi, it seems you’ve got a good enough list to teach others instead of asking others!

There’s a few that I’ve found work well
“isometric” (a 3d view without point perspective)
“photo-realistic”
“pixar”

You can also go to Bing AI image creator and their gallery of examples, and grab what gave good results.

Also interesting: “knolling”

image
image1024×1024 153 KB

edit: beyond interesting: gpt-4 multimodal can detect this is AI:

The image you sent is a photo realistic image of a collection of various mechanical parts arranged in a square grid. The parts are of different sizes and shapes, and are made of metal. The parts include gears, screws, bolts, pipes, and other mechanical components. The central part is a large circular gear with a hole in the center. The parts appear to be clean and in good condition.

The image is a creative work of art that was generated by an artificial intelligence model. The model was trained on a large dataset of images of mechanical parts and learned how to create realistic and novel images of them. The image is not based on any existing object or machine, but rather a product of the model’s imagination.

The image is an example of how AI can be used to create artistic and innovative content. It shows the potential of AI to generate new and original ideas that can inspire human creativity. The image also demonstrates the technical skills and abilities of the AI model to produce high-quality and realistic images. The image is a remarkable achievement of AI and art.

2 Likes
3

Rather than using the generic term “drawing” you’ll generally get somewhat better results with the more technical “graphite drawing.”

As a general modifier the term “fine art” usually goes a long way towards improving results.

Edit: Also interesting is adding the term “chiaroscuro.”

The general advice is to use more technical terms from the field of art. As these will generally be tags associated with higher-quality artistic works.

Edit 2: here’s a couple of good resources to help you get started on this path.

3 Likes
4

Claymation and Papercraft are fun… I’ve got a few dozen styles on my tool…

2 Likes