languid flower
If you have a specific illustration or a concept in mind, you can insert it to Vision, and the use Dall-E 3 to generate illustrations based on that description. If you're trying to replicate or get inspired by a particular style, providing a detailed description of the visual elements, colors, and composition can help achieve that, too. You need only 4 simple steps.
tall cradle
Vision?
languid flower
Yes now available directly from gpt-4, image upload option
tall cradle
So I just upload a image?
dense mesa
Unfortunately this doesn’t work properly especially with text
languid flower
Unfortunately this doesn’t work properly especially with text
So how did you try to use? Prompts are never 100% correct since it always depends the request and what you're trying achieve. So you may have to adjust accordingly
halcyon canyon
It's hard because dalle has biases due to its training data. The prompt in your first step isn't detailed enough for some use cases, like people or faces.