Below is a flowchart demonstrating the process of creating generative prompts and images dreived from brief descriptions, and edits applied to those descriptions.
#1 Initial Input
The user types a brief idea of what prompt they want chatGPT to generate
#2 ChatGPT Response
ChatGPT generates a verbose description, which will be the input for the Dall-E image generation
#3 Image Generation
An image is generated, and an edit box appears for editing the created prompt
#4 Image Editing
User adds an edit they want to have applied to the prompt ChatGPT made originally
#5 Prompt Augmentation
Another api call is made with the original prompt and the edit, which ChatGPT combines. The new prompt including the edit is generated and then sent to Dall-E again