TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

1 week of Stable Diffusion

457 点作者 victormustar超过 2 年前

18 条评论

fab1an超过 2 年前
I think most are vastly underestimating the impact of Synthetic AI media - this is at least as big as the invention of film&#x2F;photography (century-level shift) and maybe as big as the invention of writing (millenia-level shift). Once you really think through the consequences of the collapse of idea and execution, you likely to tend think the latter...<p>What we&#x27;re seeing now are toy-era seeds for what&#x27;s possible - e.g. I&#x27;ve been making a completely Midjourney-generated &quot;interactive&quot; film called SALT: <a href="https:&#x2F;&#x2F;twitter.com&#x2F;SALT_VERSE&#x2F;status&#x2F;1536799731774537733" rel="nofollow">https:&#x2F;&#x2F;twitter.com&#x2F;SALT_VERSE&#x2F;status&#x2F;1536799731774537733</a><p>That would have been completely impossible just a few months ago. Incredibly exciting to think what we&#x27;ll be able to do just one year from now..
评论 #32652080 未加载
评论 #32655230 未加载
评论 #32652072 未加载
评论 #32652234 未加载
评论 #32652157 未加载
评论 #32652194 未加载
评论 #32654055 未加载
评论 #32652549 未加载
评论 #32653693 未加载
syntaxing超过 2 年前
It’s really crazy how Stable Diffusion seems to be very on par with DALL-E and you can run it on “most” hardware. Is there an equivalent for GPT-3? I don’t even think I can run the 2M lite GPT-J on my computer…
评论 #32651799 未加载
评论 #32652209 未加载
评论 #32651721 未加载
评论 #32651909 未加载
评论 #32651682 未加载
评论 #32651841 未加载
评论 #32655378 未加载
评论 #32652650 未加载
metadat超过 2 年前
They forgot to mention the porn one..<p><a href="https:&#x2F;&#x2F;www.vice.com&#x2F;en&#x2F;article&#x2F;xgygy4&#x2F;stable-diffusion-stability-ai-nsfw-ai-generated-porn" rel="nofollow">https:&#x2F;&#x2F;www.vice.com&#x2F;en&#x2F;article&#x2F;xgygy4&#x2F;stable-diffusion-stab...</a><p>Why&#x27;d they &quot;overlook&quot; it? Probably more culturally significant and controversial than any of the others. It&#x27;s the natural elephant.
评论 #32651504 未加载
评论 #32650802 未加载
评论 #32651094 未加载
评论 #32651141 未加载
folli超过 2 年前
Funny of Reddit banning the mentioned Subs in a short amount of time.<p>Some years ago, the pendulum was very much on the other side.
评论 #32658209 未加载
评论 #32650898 未加载
评论 #32760804 未加载
评论 #32651325 未加载
评论 #32652100 未加载
gillesjacobs超过 2 年前
7 days and already that many UIs, plugins and integrations released. To be fair, developer&#x2F;researcher access was a bit earlier but that is impressive adoption speed.
评论 #32651658 未加载
danso超过 2 年前
Tangent discussion: What are people&#x27;s here experiences with running Stable Diffusion locally? I&#x27;ve installed it and haven&#x27;t had time to play around, but I also have a RTX 3060 8GB GPU -- IIRC, the official SD docs say that 10GB is the minimum, but I&#x27;ve seen posts&#x2F;articles saying it could be done with 8GB.<p>Mostly I&#x27;m interested in the processing time. Like, using a midrange desktop, what&#x27;s the average time to expect SD to produce an image from a prompt? Minutes&#x2F;Tens of minutes&#x2F;Hours?
评论 #32651716 未加载
评论 #32651674 未加载
评论 #32651919 未加载
评论 #32651942 未加载
评论 #32652905 未加载
评论 #32652467 未加载
评论 #32651652 未加载
评论 #32651916 未加载
评论 #32651750 未加载
motoboi超过 2 年前
Take a moment to appreciate the fact that in 4,2Gb (less than that actually) you have the English language somehow encoded.<p>This is mind blowing.
评论 #32653457 未加载
stephc_int13超过 2 年前
AI generated art is interesting and will probably be helpful.<p>I see it as a cheap and fast alternative to paying a concept artist.<p>But not a revolution. Creating precise and coherent assets is going to be a challenge, at least with the current architecture.<p>From a research perspective this is, I think, much more than a toy, those models can help us better understand the nature of our minds, especially related to their processing of text, images and abstraction.
评论 #32658953 未加载
评论 #32654593 未加载
评论 #32654877 未加载
m_ke超过 2 年前
After playing with it for a few hours I&#x27;m sold on it soon replacing all blog spam media and potentially flooding etsy with &quot;artists&quot; trying to pass the renders as their own art work.<p>Here&#x27;s some of the stuff I generated: <a href="https:&#x2F;&#x2F;imgur.com&#x2F;a&#x2F;mfjHNgO" rel="nofollow">https:&#x2F;&#x2F;imgur.com&#x2F;a&#x2F;mfjHNgO</a>
andruby超过 2 年前
It must be interesting being a graphic artist in 2020-2022. First NFT&#x27;s that enabled some to make millions of dollars. Less than 2 years later, Stable Diffusion, which will probably shrink the market significantly for human graphical artists.
ebabchick超过 2 年前
can someone recommend a good paper or blog post with an overview of the technical architecture of training and running stable diffusion?
评论 #32651823 未加载
ok_dad超过 2 年前
I guess we know where the new market for all those Ethereum miners&#x27; GPUs will come from. I have always been sort of bear-ish on the trend towards throwing GPU power at neural nets and their descendants, but clearly there are amazing applications for this tech. I still think it&#x27;s morally kinda wrong to copy an artist&#x27;s style using an automated tool like this, but I guess we&#x27;ll have to deal with that because there&#x27;s no putting this genie back in the bottle.
poisonborz超过 2 年前
Just imagine - you could write your own script of a series and have it realistically generated, especially cartoons, complete with voice acting. Popular generated Spongebob episodes could form canonical entries in the mind of the general public - after some information fallout, original episodes couldn&#x27;t be even told apart. Postmodern pastiche will accelerate and will become total.
revskill超过 2 年前
7 days trying to install python and their packages and failed. Have to remove those garbages , global dependencies from my machine. Such a waste of ecosystem.
评论 #32652769 未加载
评论 #32652882 未加载
评论 #32654542 未加载
评论 #32652718 未加载
评论 #32652077 未加载
评论 #32653225 未加载
评论 #32652991 未加载
lijogdfljk超过 2 年前
That Figma plugin is mind blowing to me. I&#x27;m also curious to see how the Blender integration pans out
coding123超过 2 年前
In 30 years everything AI generates will be a red circle, because at that point it will have just trained on itself repeatedly.<p>Instead of labeling data for what things are, we&#x27;ll have to label things as being generated or not.
throwaway888abc超过 2 年前
The collaboration,pace and progress is stunning. If this can applied to other fields such climate change etc.<p>Great write up
timost超过 2 年前
One use case I have in mind is manga drawing. I wonder if anybody has tested manga related generation.
评论 #32653920 未加载