how.wtf

How to prompt DALL-E effectively

· Thomas Taylor

Prompting DALL-E for the best results

DALL-E is an OpenAI model that generates images from textual descriptions. It combines elements of creativity and technology to produce unique imagery based on human prompts.

Tip 1 - DALL-E Prompt Book

To kick things off, I recommend reviewing the DALL-E prompt book released by the dallery gallery. It contains many useful techniques for describing tone, emotion, angles, lightning, illustration style, etc. It has been a wonderful resource for myself and will likely give you enough information to get started.

The mascot for https://how.wtf is a bald character, depicted frustratingly slamming his hands on a desk while shouting:

angry how.wtf mascot

Tip 2 - Upload an image and prompt DALL-E

Seeking an image of the same character riding a llama, I:

  1. Uploaded the image
  2. Prompted DALL-E with a brief request: “Given the uploaded image, render an animated version riding a llama”

prompt mascot to ride llama joyfully

Tip 3 - Ask DALL-E to describe an image

If you wish to reproduce or modify an image, DALL-E can help describe it. For instance, I asked DALL-E to describe my mascot riding the llama:

Me: How should I prompt you to get the image attached?

DALL-E: To create an image similar to the one attached, you could use a prompt like this:

“Create an image of a joyful, cartoonish, bald character with large expressive eyes, riding on the back of a fluffy, stylized alpaca. The character has a wide, excited smile and is holding onto reins as if they are having the time of their life. The alpaca appears to be in mid-gallop, creating small clouds of dust behind it to show its speed. The environment is a soft, minimalistic background that doesn’t draw attention away from the action of the ride. The lighting should be bright and cheerful, capturing the sense of fun and adventure in the scene.”

With DALL-E’s guidance, I then requested a Simpsons-style version in a 16:9 aspect ratio:

Me: Create a 16:9 aspect ratio image of a joyful, cartoonish, bald character with large expressive eyes, riding on the back of a fluffy, stylized alpaca in the style of “The Simpsons”. The character has a wide, excited smile and is holding onto reins as if they are having the time of their life. The alpaca appears to be in mid-gallop, creating small clouds of dust behind it to show its speed. The environment is a soft, minimalistic background that doesn’t draw attention away from the action of the ride. The lighting should be bright and cheerful, capturing the sense of fun and adventure in the scene.

This resulted in the featured image for this blog post!

Conclusion

DALL-E’s remarkable ability to generate customized images from textual prompts is evident. By using specific descriptors and understanding the nuances of effective prompting, as detailed in the DALL-E prompt book, users can create vibrant, expressive images that vividly capture their ideas with striking accuracy and creativity.

#Generative-Ai  

Reply to this post by email ↪