写了个支持连续对话的ChatGPT Bot,跑在Cloudflare Workers上。代码简陋但能用。https://github.com/Envl/workers-chatgpt-tgbot
👍4❤1
enjoy https://twitter.com/eranhill/status/1601988444498034688
介绍: https://cdm.link/2022/08/chantlings-an-app-where-animated-forest-creatures-sing-along-with-you-is-a-marvel-of-design/
App: https://www.iorama.studio/chantlings
介绍: https://cdm.link/2022/08/chantlings-an-app-where-animated-forest-creatures-sing-along-with-you-is-a-marvel-of-design/
App: https://www.iorama.studio/chantlings
X (formerly Twitter)
eranhill on X
a portal to dreamland on my desk #chantlings @ioramastudio
除了标题比较糟糕,其他都很好。Xmind创始人的一次分享。 https://mp.weixin.qq.com/s/MLM9x8skNKG-9sEruVhW-Q
Weixin Official Accounts Platform
神仙公司是怎样炼成的?Part 1
【以下文字由园长2022年11月12日在上海圆桌活动录音整理而成】
👍1
工作中有点想搞个code figma联动的东西。市面上类似概念的产品看了有很多,但这个公司做的东西是目前看到最酷的。https://divriots.com/
Divriots
‹div›RIOTS
We craft revolutionary Figma plugins!
❤1
终于1.0。实际上已经默认用它做项目一年多了,期间经历了很多bug和breaking change,但收益依然大过成本。它能提供等于原生HTML+JS+CSS的符合直觉的开发体验&性能。我认为SvelteKit值得成为绝大多数情况下的首选,无论全栈or纯前端项目。 https://svelte.dev/blog/announcing-sveltekit-1.0
svelte.dev
Announcing SvelteKit 1.0
Web development, streamlined
👏2❤1
Oculus之前的CTO,从Meta离职了。他说自己无法忍受Meta内部运作的低效且对此感到无力。 https://m.facebook.com/story.php/?story_fbid=pfbid0iPixEvPJQGzNa6t2x6HUL5TYqfmKGqSgfkBg6QaTyHF5frXQi7eLGxC7uPQv5U5jl&id=100006735798590&s=09
Facebook
Log in to Facebook
Log in to Facebook to start sharing and connecting with your friends, family and people you know.
阿根廷🇦🇷人在布宜诺斯艾利斯庆祝🎉夺冠🏆。今年最爱的视频,这两天看了好多次,头皮发麻。 https://petapixel.com/2022/12/19/amazing-drone-footage-of-200k-argentinian-fans-celebrating-world-cup/
(Instagram原版视频不知为何没声音了。用电脑打开链接能直接从这个网页播放,有声音。手机打开只能跳转ig。)
(Instagram原版视频不知为何没声音了。用电脑打开链接能直接从这个网页播放,有声音。手机打开只能跳转ig。)
Peta Pixel
Amazing Drone Footage of 200K Argentinian Fans Celebrating World Cup
A special moment for the country.
❤🔥1
世界杯直播中出现了一些关键场景(比如越位)的3D重建画面,nyt在他们的网站中也做了类似事情,这篇文章介绍了实现流程。
关键大概四步:
1. 从现场照片中足球大小&禁区形状计算出摄影师位置,然后求出投影矩阵,把照片映射到真实足球场模型去,于是得到各球员位置
2. 用机器学习模型单独计算照片中各球员的3D模型,结合同时刻其他角度的照片,手工校正姿势&其他细节
3. 参照原始照片,把球员模型摆到球场模型中对应位置,调整角度、大小
4. 添加shader、运镜、渲染
https://rd.nytimes.com/projects/modeling-key-world-cup-moments-with-machine-learning
关键大概四步:
1. 从现场照片中足球大小&禁区形状计算出摄影师位置,然后求出投影矩阵,把照片映射到真实足球场模型去,于是得到各球员位置
2. 用机器学习模型单独计算照片中各球员的3D模型,结合同时刻其他角度的照片,手工校正姿势&其他细节
3. 参照原始照片,把球员模型摆到球场模型中对应位置,调整角度、大小
4. 添加shader、运镜、渲染
https://rd.nytimes.com/projects/modeling-key-world-cup-moments-with-machine-learning
Nytimes
Modeling Key World Cup Moments with Machine Learning
R&D used reference photos and videos of the venue, photos of the play, and player models created using a machine learning model called ICON to recreate Pulisic’s goal in 3D.
Forwarded from Hacker News
Omg.lol – A lovable web page and email address, just for you (🔥 Score: 156+ in 3 hours)
Link: https://readhacker.news/s/5v9Se
Comments: https://readhacker.news/c/5v9Se
Link: https://readhacker.news/s/5v9Se
Comments: https://readhacker.news/c/5v9Se
home.omg.lol
omg.lol - A lovable web page and email address, just for you
Treat yourself to an awesome web address, a devastatingly gorgeous profile page, a stellar email address, and tons more