{"id":25463,"date":"2022-12-22T22:23:22","date_gmt":"2022-12-22T22:23:22","guid":{"rendered":"https:\/\/mailinvest.blog\/index.php\/2022\/12\/22\/new-openai-tool-turns-text-into-3d-models\/"},"modified":"2022-12-22T22:23:22","modified_gmt":"2022-12-22T22:23:22","slug":"new-openai-tool-turns-text-into-3d-models","status":"publish","type":"post","link":"https:\/\/mailinvest.blog\/index.php\/2022\/12\/22\/new-openai-tool-turns-text-into-3d-models\/","title":{"rendered":"New OpenAI Tool Turns Text Into 3D Models"},"content":{"rendered":"

<\/a>\n<\/p>\n

This site may earn affiliate commissions from the links on this page. Terms of use<\/a>.<\/div>\n
\n
\n
\n\"\" <\/div>\n<\/div>\n

It\u2019s unusual for an innovation to make equally large waves in the art and artificial intelligence communities, but that\u2019s what OpenAI accomplished with the release of the DALL-E image generator<\/a>. Simply type in a description, and DALL-E can make it real. The company\u2019s new algorithm has a similar pitch, but instead of making a 2D image, Point-E creates a 3D model of your description.<\/p>\n

Point-E doesn\u2019t go right from text to a 3D mesh \u2014 it\u2019s actually composed of two different AI models. First, a text-to-image AI processes the prompt to make a standard 2D image. Then, an image-to-3D AI swings into action to turn the flat rendering into a 3D model. So, if you were to ask Point-E to make a traffic cone, it would start with a stripey triangle. It\u2019s up to the second AI to understand that traffic cones are cones to produce the correct 3D shape.<\/p>\n

This isn\u2019t a completely novel idea \u2014 Google has a tool called DreamFusion that can do something similar. However, DreamFusion was designed to run on a machine with four of Google\u2019s custom TPU v4 AI processors, and it takes that hardware 90 minutes to generate an image. You\u2019re looking at multiple hours of GPU time per image with DreamFusion.\u00a0Point-E is much faster<\/a>, and it can run on a computer with a single GPU.<\/p>\n

Many of us have had a blast telling DALL-E to make off-the-wall renderings, but Point-E isn\u2019t yet ready for that instant AI gratification. The study<\/a> explains this is the first step in foundational technology that could eventually become as quick and easy as DALL-E. OpenAI says that Point-E still falls far short of commercial 3D modeling \u2014 the results are more akin to a cloud of points.<\/p>\n

\"\"<\/p>\n

Point-E models converted to a mesh.<\/p>\n<\/div>\n

When smoothed and processed, the Point-E models can produce a passable representation of a real object, as seen above. The real leap forward here is the efficiency, which OpenAI claims is \u201cone to two orders of magnitude faster\u201d than existing systems. Perhaps in the future, Point-E will intrude upon the domain of 3D modeling the way DALL-E has in art. Adobe recently decided it would allow AI-generated art<\/a> in its stock image library, which not all artists like.<\/p>\n

If you want to tinker with Point-E, all the code is available on GitHub. However, you\u2019ll need to have Python, as well as some experience with programming and command line tools to make it work. However, the relatively modest hardware requirements mean it\u2019s more accessible than DreamFusion.<\/p>\n

Now read: <\/strong><\/p>\n<\/p><\/div>\n