search

CVMI-Lab / Classifier-Score-Distillation

(ICLR2024) This is the official PyTorch implementation of ICLR2024 paper: Text-to-3D with Classifier Score Distillation
105 stars 4 forks source link

The weights of positive and negative terms of CSD used for text-to-3D generation #1

Closed lizhiqi49 closed 6 months ago

lizhiqi49 commented 7 months ago

I didn't see the description in paper, or I missed it. Can you tell me?

lingtengqiu commented 6 months ago

I didn't see the description in paper, or I missed it. Can you tell me?

maybe https://xinyu-andy.github.io/Classifier-Score-Distillation/ ?

lizhiqi49 commented 6 months ago

I mean the two hyper-parameters, $\omega_1$ and $\omega_2$ in equation (10) of the paper. I didn't see the corresponding description of them for text-to-3D generation either in paper or the project page.

lingtengqiu commented 6 months ago

I mean the two hyper-parameters, $\omega_1$ and $\omega_2$ in equation (10) of the paper. I didn't see the corresponding description of them for text-to-3D generation either in paper or the project page.

The details can be found in ICLR2024, openreviewer. I saw the author provide the link attached to partial CSD codes.

lizhiqi49 commented 6 months ago

The details can be found in ICLR2024, openreviewer. I saw the author provide the link attached to partial CSD codes.

I found it, thank you!