TE
TechEcho
Home24h TopNewestBestAskShowJobs
GitHubTwitter
Home

TechEcho

A tech news platform built with Next.js, providing global tech news and discussions.

GitHubTwitter

Home

HomeNewestBestAskShowJobs

Resources

HackerNews APIOriginal HackerNewsNext.js

© 2025 TechEcho. All rights reserved.

A prompt engineering guide for DALLE-2

242 pointsby kevemanalmost 3 years ago

13 comments

indiv0almost 3 years ago
I was expecting some clickbait&#x2F;spam (the layout of the website has that feel) but this was surprisingly super in-depth and 100% matches up with my experience doing prompt engineering.<p>There&#x27;s a fine line between so descriptive that the AI hits an edge case and can&#x27;t get out of it (so every attempt looks the same) and not being descriptive enough (so you can&#x27;t capture the output you&#x27;re looking for). DALL-E is already incredibly fast compared to public models and I can&#x27;t wait for the next order-of-magnitude improvement in generation speed.<p>Real-time traversal of the generation space is absolutely key for getting the output you want. The feedback loop needs to be as quick as possible, just like with programming.
评论 #32091626 未加载
评论 #32096722 未加载
评论 #32091363 未加载
o_____________oalmost 3 years ago
Great document.<p>Damn I am salivating to get access to Dall-E for some projects. Been on the waiting list for quite a while.<p>I&#x27;ve been experimenting with Midjourney, which is amazing for spooky&#x2F;ethereal artwork, but it struggles with complex prompts and realism.
评论 #32090088 未加载
评论 #32089751 未加载
评论 #32090244 未加载
评论 #32091766 未加载
评论 #32094357 未加载
评论 #32096755 未加载
评论 #32093010 未加载
alexjrayalmost 3 years ago
The Open AI clear content policy is quite interesting to me. It&#x27;s reasonable but clearly controlling.
评论 #32089997 未加载
skybrianalmost 3 years ago
Nice! I was wondering why there are example images of real-looking people, but it seems this is allowed now:<p><a href="https:&#x2F;&#x2F;www.vice.com&#x2F;en&#x2F;article&#x2F;g5vbx9&#x2F;dall-e-is-now-generating-realistic-faces-of-fake-people" rel="nofollow">https:&#x2F;&#x2F;www.vice.com&#x2F;en&#x2F;article&#x2F;g5vbx9&#x2F;dall-e-is-now-generat...</a>
评论 #32093037 未加载
tracyhenryalmost 3 years ago
Based on this, an interesting project would be paraphrasing any regular prompt into a prompt that works for DALLE-2.
yoyopaalmost 3 years ago
i really don&#x27;t understand how people can appreciate something like this. to me it just filling the world with literally mindless garbage.
评论 #32092789 未加载
评论 #32093278 未加载
评论 #32092421 未加载
评论 #32093032 未加载
aantixalmost 3 years ago
Dall-e still has a lot of work to be done with face construction.<p>Maybe that’s a feature not a bug.
评论 #32091606 未加载
评论 #32091599 未加载
godmode2019almost 3 years ago
Can anybody recommend a prompt engineering resource for language models?<p>Interesting topic
评论 #32091901 未加载
alana314almost 3 years ago
This is great, lots of good ideas in the deck.
totetsualmost 3 years ago
There is some shared Google docs in the dalle2 discord community about this too.
seydoralmost 3 years ago
What&#x27;s the copyright situation for images from dalle&#x2F;imagen?
评论 #32094610 未加载
codeshauntedalmost 3 years ago
This would be super useful if I actually had access :P
评论 #32093137 未加载
trentionalmost 3 years ago
Calling this &quot;engineering&quot; is just beyond parody.
评论 #32094209 未加载