Hacker News Reader

Comment by meowtastic | Hacker News Reader

I haven't got access to DALL-E 2, but I did give Midjourney (https://www.midjourney.com/) a go. I found it really cool it created images that somewhat resembled my prompt, but I still felt it was way off what I really wanted. Maybe I didn't word the prompt correctly, maybe I didn't give it enough tries. Either way, I feel like we'll eventually move away from generic prompts to something that'll look a lot like...programming, funnily enough.

View on HN

If you're using Midjourney, you might like this prompt generator that I made after scraping prompts from their Discord server: https://huggingface.co/succinctly/text2image-prompt-generato...

Looks cool! One piece of feedback I have is that I personally wouldn't want to autocomplete prompts as shown in the image you have there. I'd want something that "fixes" my prompts so that they're understood by the model. Like say if I put "Afro-Asian samurai carrying two swords in a cyberpunk environment", for it to say how I could re-word it to generate something closer to what I actually want.

That's a really neat idea, since most of the time people already know what they want to generate. I'd have to think a little more about how I'd get the training data for that. LMK if you have any ideas.

Cool! Based on your description, I thought I can give the generator an image and it'll return the prompt. Would that be possible?

I really enjoy Midjourney for more abstract and vague vibe art. It's not as good at specific stuff (although Imagen/Parti are much better at it even than Dall-E too, which has its own shortcoming around text and numbers).

My favorite is coming up with two word prompts, like "endless beginnings" or "stressful shapes" or "happy anxiety".

Ah I see, thanks for the tips!

Reply on news.ycombinator.com