Hey,
I have been using Dalle3 API to create images for my mobile application.
There is a limitation.
I can’t attach an image and tell Dalle3 to convert that image to a cartoon image for example. I can do this from front end by logging into my Open AI account.
I simply want to know, when will OpenAI provide the ability to attach images and play with the attached image via Dalle3 API?
Thanks.
DALL-E has no ability to modify images in a way you specify.
There is no vision.
The only way that ChatGPT would be able to complete such a task is by first using gpt-4-vision-preview’s machine vision to analyze the image to obtain a text description of the contents, and then send a prompt to dall-e, suitable for recreating it.
You can make the separate calls to the models yourself via API.
It only seems acceptable because of the significant transformation. Words are the middle step, not some kind of internal processing of imagery.