search

Chuny1 / 3DGPT

765 stars 36 forks source link

Looking forward to this code #7

Open ShaneKirkbride opened 11 months ago

ShaneKirkbride commented 11 months ago

If there is a chance to beta test we’d be happy to do so!

matthewagi commented 11 months ago

same here

On Sun, 19 Nov 2023 at 21:56, ShaneKirkbride @.***> wrote:

If there is a chance to beta test we’d be happy to do so!

— Reply to this email directly, view it on GitHub https://github.com/Chuny1/3DGPT/issues/7, or unsubscribe https://github.com/notifications/unsubscribe-auth/AGM62KYT672233DCAMEJ7SDYFJW6PAVCNFSM6AAAAAA7R7U7P6VHI2DSMVQWIX3LMV43ASLTON2WKOZSGAYDCMBTGI3DCMA . You are receiving this because you are subscribed to this thread.Message ID: @.***>

TangSY commented 11 months ago

same here

rfan-debug commented 9 months ago

Same here. This is a great work.

Jiaqi-Chen-00 commented 9 months ago

same here

ForrestPi commented 7 months ago

same here

CAMOBAP commented 7 months ago

same here

darkknight0007 commented 6 months ago

same here

donghaiwang commented 6 months ago

same here

guohaoxiang commented 4 months ago

same here

scchess commented 2 weeks ago

There is no code - not sure this work was real or not. It could just be another academic misconduct. The paper looked too strong from the methods.

Chuny1 commented 2 weeks ago

There is no code - not sure this work was real or not. It could just be another academic misconduct. The paper looked too strong from the methods.

Thank you for your interest! I understand your concern regarding code availability.

The code release is tied to the acceptance of our paper, which has faced multiple rejections and is currently awaiting another decision. Once accepted, the code will be made available as planned.

Our method outperforms other 3D generation approaches by using procedural modeling, which builds real meshes, unlike others using neural networks that may produce lower-quality meshes. However, it does reduce diversity compared to prior work, as it’s limited to objects with procedural code.

I recommend trying the sky modeling case in the Infinigen library, using the prompt in our paper, it should give you a good sense of the results. You can implement this in a few hours or even use GPT then put parameters into Infinigen.

Thanks for your patience!