Trying to generate image via DALL E for educating kids about parts of plants. The image is coming with labels in some weird language. Is there any way to generate images with English labels.
DALL-E cannot reliably produce language in images. It is trained on visual labeled imagery, and unless you pass in words yourself or have the AI write them, there is no hope of having anything meaningful that is not scrambled at the least. Especially it should not be used for education.
You won’t find “perfectianiom” in the text of a DALL-E image. (unless you look closely)
I think using the exact text we want in the image in the prompt also
doesn’t work. I tried adding those words in prompt but its coming out in a different manner.
Was anyone able to resolve this issue?
Hi, welcome to the community!
With DALL-E, it is very hard.
But with new 4o Image Generation in ChatGPT, it is better.
Sorry but I wanted to know, if its possible to do it using the OpenAI API not ChatGPT
You can use API, but it is same.
This is the topic from yesterday. The user uses API.
If you use gpt-image-1, its result is better about text. But not OK, everytime.