HoLa B-Rep,也就是人工智能计算机辅助设计的恶作剧?
有人试过这个吗?使用整体潜在表示生成 HoLa-BRep:
https://huggingface.co/spaces/YuXingyao/HoLa-BRep
结果和 Spectral Labs 的 SGS-1 一样奇怪/不可用吗?
https://huggingface.co/spaces/spectral-labs/SGS-1
Transfigure 不使用奇怪的 DeepCAD 训练数据(幼儿园水平)。
我们使用的是来自硬件工程师的真实数据,他们像生命一样重视 CAD 的设计(因为这能支付我的房租)。
http://xfgr.ai
查看原文
Anyone try this? HoLa-BRep Generation using a Holistic Latent Representation:
https://huggingface.co/spaces/YuXingyao/HoLa-BRep<p>Same weird/un-usable results as Spectral Labs SGS-1? https://huggingface.co/spaces/spectral-labs/SGS-1<p>Transfigure does not use weird DeepCAD training data (kindergarten level)<p>We are using real data, born from hardware engineers who design CAD like their life depends on it (because it pays my rent)<p>http://xfgr.ai