gd3kr

Understandable. For context, the GPT in the name comes from an earlier version of this project (https://github.com/gd3kr/blendergpt) which actually used GPT-4 to write python scripts that Blender would then execute. This would allow GPT-4 to program operations like instantiating primitives with the Blender Python API given only a text prompt (ex. "create 50 cubes")

The new version of BlenderGPT (lets call this v2) doesn't use an any autoregressive token prediction for the actual mesh generation part, so I understand why it sounds dishonest. I really just chose to stick with the name because artists really didn't seem to care about how the meshes are generated, and the term GPT became closely associated with AI.

As for the technical stuff, I've been working on BlenderGPT v2 for the past several months, and until a week ago, i had been using a custom pipeline I built borrowing and re-implementing bits of Unique3D (https://wukailu.github.io/Unique3D/) and combining it with optimized models (flow matching diffusion models etc) for intermediate steps (text to image generation). My optimizations reduced inference time from >2 minutes to only about 20 seconds. This is the model used in this demo i shared: https://x.com/gd3kr/status/1853645054721606100

And then Microsoft released Trellis (https://github.com/microsoft/TRELLIS), and it seemed to leapfrog my model's capabilities on most things. Integrating it into the pipeline wasn't too hard and so I went forward with it.

All of this is just to say that there really was a lot of effort put into the core pipeline, and the landing page was mostly an afterthought. Actively working on a more comprehensive one that covers all the points I talked about.

Hey HN! I'm aarya (twitter.com/gd3kr) and I built BlenderGPT

Really overwhelmed with the traffic, will be defaulting to a less compute intensive model for a bit. Will try to keep up with this thread and respond to questions/comments but mostly focused on not letting BlenderGPT crash.

BlenderGPT is entirely bootstrapped and i'm really the only one on the team. Also, The required google sign in is only put in place to prevent botting/account creation abuse which is really not ideal when dealing with expensive GPU associated compute costs while generating every model.

I hope you have fun with it! DM me on twitter if you enjoy it and would like more credits to try it out.

Hey, no it’s not a parody lmao. I shipped it because realised it could have at least some utility while building it. To be fair, the wrong examples would be an oversight on my part, but I thought of them being more of genuine benchmark of what GPT3 was capable of; sort of like an experimental feature. I’m working hard on refining the results by fine tuning DaVinci and getting the output as close to ideal as possible — I think there’s a lot of potential there. In the end, even boosting a user’s productivity marginally is a win in my book, and as a lot of people have pointed out, it’s already doing that to some extent.

Hey! I’m the creator of www.autoregex.xyz (@gd3kr on twitter) I originally built it as a small side project in a couple days, I was absolutely not expecting a response this massive. I realise concerns about email and password sign up, and I built it in an hour-ish with firebase auth as a temporary solution to capping the sudden surge in GPT3 requests to the server after the twitter post gained traction. Im working on a better approach to that involving not having to create an account. Otherwise, all suggestions are absolutely welcome. Please tell me how I can make this a better experience for everyone.