Re:DALL-E 3: Improving image generation with better captions...
by visarga on Friday October 20, 2023 @12:18AM
(#37952345)
When I used GPT4-Image to get a full detailed description of an image, I was sure it was going to be used as a method for generating training data.
[ Parent
| on HN ]
Re:DALL-E 3: Improving image generation with better captions...
by cocoflunchy on Friday October 20, 2023 @04:54AM
(#37953768)
I just tried this and wow, I'm impressed. Input / output: https://imgur.com/a/nTSkJUK