How DALL-E 2 Works

Re:How DALL-E 2 Works by SleekEagle on Tuesday April 19, 2022 @12:15PM (#31085176)

GPT-3 was released 2 years ago, and in that time CLIP, GLIDE, and DALL-Es 1 and 2 have been released. All of this is just from OpenAI too! DL research is cranking along as quickly as ever imo!

[ Parent | on HN ]

Re:How DALL-E 2 Works by lurker619 on Tuesday April 19, 2022 @02:27PM (#31086922)

Just need a music one please.

[ Parent | on HN ]

Re:How DALL-E 2 Works by donkarma on Tuesday April 19, 2022 @07:04PM (#31090531)

You're going to be waiting a while, even now these models can't get the details right. Look at all the example pictures, everything is just wrong. You can obviously see what it is trying to get at, but it can't get there. Another example of the last 10% taking 90% of the time.

[ Parent | on HN ]

Re:How DALL-E 2 Works by gwern on Tuesday April 19, 2022 @03:11PM (#31087511)

Jukebox. If you listen to Jukebox samples, recall that that was quite a while ago in dog/DL years, and imagine what the DALL-E 2 equivalent would be for a Jukebox 2...

[ Parent | on HN ]

Re:How DALL-E 2 Works by p1esk on Tuesday April 19, 2022 @03:15PM (#31087551)

I'm surprised no one has tried to launch a music generation startup based on Jukebox. I'd be interested in collaboration if anyone wants to work on it (and has compute resources).

[ Parent | on HN ]