muelea / buddi Goto Github PK

Hi!

I'm trying to run the project, but I'm having problems setting up the environment.

Locally

First, I tried using it in my own computer. It's running Arch Linux, and I could follow the instructions from Install Conda Environment until I tried to install detectron2, which I couldn't manage to do.

I understand that this is an issue with detectron2 and not specifically with BUDDI, but maybe this wouldn't happen if I could match the rest of the software versions and the like. What OS do you use?

Colab

As I couldn't set up the environment, I tried using the linked Google Colab notebook. However, after running the cells, it failed because it couldn't find SMPLA_NEUTRAL.pth. As far as I can tell, this "smpla" model is a product of using ROMP, which is not running correctly.

Again, this may be ROMP's fault and not BUDDI's, but as I'm not experienced with this particular setup I think it's better to open an issue here and see if it can be easily fixed.

Thanks in advance,
Andrés

it can be real time if not vis it, how many fps ?

Can this work with just one person? If input image is for one person and want output for one person as well

Hi!

Great work! Is there a demo script for running the optimization on custom data? If not could you detail the steps needed?

Thanks

Hello, I can't download your pseudo-ground truth fits for dataset processing, could you check the pseudo-ground truth fits URL link in buddi/documentation
/DATA.md? Thanks!

Thanks for the wonderful work!

I saw that you reported the evaluation results on Hi4D. However, I could not find a way to reproduce that given your instructions in README? Could you help to indicate how to reproduce the reported Hi4D performance as in the paper? Thanks!

Hi,
Thanks for your great work! But it seems that there's no pseudo ground-truth fits for FlickrCI3D and auxiliary data in https://github.com/muelea/buddi/blob/e5825fe4d23707433ea83d0f95642e77757cf519/documentation/url.url! Could you please upload them if possible? Thanks in advance!!!

In llib/methods/hhc_diffusion/evaluation/sample.py, it seems that lines 17 and 20 import the functions build_renderer, render_360_views, and eval_diffusion. However, upon reviewing your released code, I couldn't locate the exact location of these functions. Could you please clarify where these functions are defined or provide additional information regarding their source?

Hi 👋 Thanks for the project ❤ I've created a colab notebook https://github.com/camenduru/buddi-colab

However, I'm unsure about how to generate vitpose keypoints. Could you please help me with this?

https://raw.githubusercontent.com/muelea/buddi/main/demo/data/FlickrCI3D_Signatures/demo/vitpose/college_232334_keypoints.json

Hi,

While you make the 'train_val_split.npz' files available, can you specify what the data splits are for all the datasets (Flickr, CHI3D and Hi4D)?

Thank you!