DALL·E 2, by OpenAI, is a new AI system that can create realistic images and art from a description in natural language. With this system, you can create images by just describing how you want them to be. Without any further ado, let us jump in to see how this is going to revolutionise the digital art world.
How to generate images in DallE?
There are five ways to generate images in DallE:
1.Entering text to generate images
One way to generate images text in DallE is by simply entering descriptive text. For example:
A bowl of soup that is a portal to another dimension
2. Generating variations
DALL·E 2 can take an image and create different variations of it inspired by the original. For example:
3. Directly uploading an image and editing it
Below the text box on the DALL·E 2 website, there is an upload image button which allows you to upload an image and edit it. After uploading, with the eraser tool at the bottom, you can erase specific areas of your image and add some textual description to fill in that empty spot. With this you have your image intact + some AI suggestions for the empty spots.
4. Using the previously generated image and editing it
You can also do the erasing and editing with previously generated images by clicking the edit button in the top right corner of the image.
5. Adding frames on or around the image
During editing, another tool you might come across is Frames. Frames are squares that allow you to generate images with more information. You can place a frame on top of an image to add more detail on them or around them to create a bigger image with more background descriptions. Note that placing a frame on top of an image will overlap that specific area and generate an image.
If you want to make the overlapped part look a bit more better, you can use keywords like “continuous” and “smooth” to make them have a continuous transition.
Helpful Resources
These two resources were the most helpful to me when deciding what text to enter. DALL·E 2 allows you to create a maximum of two accounts under one phone number. Each account gives you only 50 free credits, where 1 credit is equal to clicking on the generate button once. That is why wisely typing in the right text to generate the right images is important.
- Dallery Gallery
Dallery Gallery provides helpful guides and articles as to how to use descriptive words to create prompts that will give you the exact images you want to generate. In fact, they have depicted various styles, attributes and properties with the help of necessary descriptive words and the corresponding images they generate. Another cool thing I came across is their prompt book (download the PDF), which contains guides to aesthetics, vibes, emotional prompt language, photography, film styles, illustration styles, art history, 3D art, prompt engineering, out-painting, in-painting, editing, variations, creating landscape and portrait images in DALL·E, merging images, and more!
There is no better place to learn about the specifications of prompts than the DALL·E 2 website itself. There are plenty of resources available on their website as well so feel free to explore!
My NFT collection — MetaSkulk
Lastly, I would really appreciate if y’all can view my first ever NFT collection, the MetaSkulk. The MetaSkulk is a digital art collection of 50 unique MetaFox NFTs -unique digital collectibles living on the Ethereum blockchain. Feel free to onboard them on your Metaverse journey! These took quite a lot of credits to make but I am content with the progress I have made. I will be releasing more NFT collections using generative art and Solidity soon, so keep a tab on my collections in my linktree.
MetaFox #42
Conclusion
DALL·E 2 has created a multitude of opportunities for creators and artists to use AI and explore realms that we have not visited yet. With AI tools like these, we can automate a lot of digital content being made nowadays and capitalize the chances it creates for us.
I hope you enjoyed reading this article and found it to be helpful. Please follow, share and subscribe to get updates on more interesting and insightful articles in the future :)