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。

Meta Open-Sources Computer Vision Foundation Model DINOv2

224 点作者 thread_id大约 2 年前

12 条评论

behnamoh大约 2 年前
What's the benefit for Meta? They are now the true open source AI providers (after OpenAI got closedAI), but I wonder why they keep releasing such models for free and kinda open source?
评论 #36054323 未加载
评论 #36054256 未加载
评论 #36054304 未加载
评论 #36054125 未加载
评论 #36054131 未加载
评论 #36054094 未加载
评论 #36056841 未加载
评论 #36054177 未加载
评论 #36054427 未加载
评论 #36054482 未加载
评论 #36054134 未加载
评论 #36054537 未加载
评论 #36054577 未加载
评论 #36054756 未加载
评论 #36054196 未加载
TekMol大约 2 年前
The license is &quot;Attribution-NonCommercial 4.0 International&quot;:<p><a href="https:&#x2F;&#x2F;github.com&#x2F;facebookresearch&#x2F;dinov2&#x2F;blob&#x2F;main&#x2F;LICENSE">https:&#x2F;&#x2F;github.com&#x2F;facebookresearch&#x2F;dinov2&#x2F;blob&#x2F;main&#x2F;LICENSE</a><p>The license seems to only apply to &quot;sharing&quot; the code. Probably because it addresses copyright issues.<p>But what about commercial <i>use</i> of the code and weights?<p>Can one incorporate those into a server side system which lets users pay for using it?
评论 #36054908 未加载
评论 #36054895 未加载
Pandabob大约 2 年前
Alas the license is CC-BY-NC 4.0.
评论 #36053759 未加载
评论 #36054755 未加载
评论 #36055043 未加载
评论 #36053819 未加载
rocauc大约 2 年前
Original release from Meta (April 17): <a href="https:&#x2F;&#x2F;ai.facebook.com&#x2F;blog&#x2F;dino-v2-computer-vision-self-supervised-learning&#x2F;" rel="nofollow">https:&#x2F;&#x2F;ai.facebook.com&#x2F;blog&#x2F;dino-v2-computer-vision-self-su...</a><p>Demos: <a href="https:&#x2F;&#x2F;dinov2.metademolab.com&#x2F;" rel="nofollow">https:&#x2F;&#x2F;dinov2.metademolab.com&#x2F;</a>
orangepurple大约 2 年前
Meta Releases Restricted Source Code for Computer Vision Foundation Model DINOv2
forgingahead大约 2 年前
Interesting to see if&#x2F;how this is different (or maybe built upon) Segment Anything[0].<p>For folks in the know: I often see segmentation models on video frames producing patchy results (see the DinoV2 video of the running dog, the body gets black patches randomly, so the segmentation fails for certain frames). What methods are folks using to deal with this - standard fine-tuning, or is there a way to &quot;force&quot; the area to be cleanly segmented (ie, add a bounding box around the class to supplement the data)?<p>And is it something that can be implemented in foundation models, or are we always going to have patchy zero-shot results like this on video files?<p>[0]<a href="https:&#x2F;&#x2F;github.com&#x2F;facebookresearch&#x2F;segment-anything">https:&#x2F;&#x2F;github.com&#x2F;facebookresearch&#x2F;segment-anything</a>
gnabgib大约 2 年前
Meta-meta-meta, this low content post references one of the (8 and counting) hn posts on meta&#x27;s post about DINO (36 days ago, 176 points)[0]<p>[0]: <a href="https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=35605039" rel="nofollow">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;item?id=35605039</a>
评论 #36053954 未加载
RobotToaster大约 2 年前
It&#x27;s not open source, the noncommercial licence violates several points of the open source definition.
评论 #36057212 未加载
_gtly大约 2 年前
see: <a href="https:&#x2F;&#x2F;dinov2.metademolab.com&#x2F;" rel="nofollow">https:&#x2F;&#x2F;dinov2.metademolab.com&#x2F;</a>
sacnoradhq大约 2 年前
And there&#x27;s another round of layoffs tomorrow. Security alert for everyone to WfH.
bg24大约 2 年前
I am not an expert, but Meta might be missing out on a business opportunity by being blindsided on everything social media. Over time, every company (Google, Apple, Amazon) diversified themselves. With their AI firepower, they could have explored a number of new avenues. Every business need not boot with billion users.
评论 #36054254 未加载
eclectic29大约 2 年前
I feel Meta, Google and OpenAI are the real AI leaders of this industry. Definitely not Microsoft who is riding on integrations with OpenAI APIs. Just saying.
评论 #36054337 未加载