Chen Wang

icon indicating a website link https://cwchenwang.github.io icon indicating an email [email protected]

icon company University of Pennsylvania icon location Philadelphia, PA icon location PhD Student at Penn

Results 20 comments of Chen Wang

Search for citation key

same problem here

Hard coded warning

We test one image for debugging and didn't clean up the code, see here at `nerf_downX_model.py`: We have uploaded a new commit to solve this issue, thanks for pointing out....

Unmatched performance with README

@ashawkey Thanks for your reply. It seems your results here also has some gaps with the videos in the readme. So why the results is such random?

Multiview models

Also interested

Access blocked

same problem here

Can the input four views of the gaussian unet be four random viewpoints and how to input the pose?

Thank you for your helpful answer! It would be even better if we can pass the pose as a parameter.

Can the input four views of the gaussian unet be four random viewpoints and how to input the pose?

Thank you! I found by converting elevation and azimuth to poses are much easier to do.

Weird results when testing on customized images

@ashawkey Thank you! It looks much better now, but seems not that good, is this expected? https://github.com/3DTopia/LGM/assets/23579918/29a8f98d-4981-4cf7-9f04-e2a69fc6b942

How do you render an image from flexicube geometry?

any update? @thuwzy

Looking forward to the custom demo

Also looking forward to the custom demo