Rendered results not meeting expectation on training set

[475651582]

It is really a fantastic work. Congrats! I managed to deal with all the data preprocessing and testing. However I found the rendered results on avatar_lbn1 is not quite ideal, even on the training smpl dataset. The rendered image rendered below ith command [python main_avatar.py -c configs/avatarrex_lbn1/avatar.yaml --mode=test]: Is that sounds reasonable to you? Thanks very much!

[lizhe00]

Could you attach your yaml file?

[lizhe00]

And which frame?

[475651582]

frame id is 00001080 avatar.txt

[lizhe00]

My result seems to be correct. Do you download our preprocessed files (https://github.com/lizhe00/AnimatableGaussians/blob/master/PREPROCESSED_DATASET.md)?

[475651582]

My result seems to be correct. Do you download our preprocessed files (https://github.com/lizhe00/AnimatableGaussians/blob/master/PREPROCESSED_DATASET.md)?

yes, here is the dataset structure preview:

[lizhe00]

What about other frames?

[475651582]

What about other frames?

mostly faild. most of the faces looks strange like they are stitched to the head.

I only tried your released pretrained model. Do I have to train from scratch?

[475651582]

Note that I mosaic the face for the sake of privacy :)

[lizhe00]

Thanks for protecting the privacy. Sorry, I don't know the reason. The results I just ran are correct. Maybe you can check if some codes are changed or train from scratch.

[475651582]

Thanks for protecting the privacy. Sorry, I don't know the reason. The results I just ran are correct. Maybe you can check if some codes are changed or train from scratch.

Thanks, I will review that. BTW, there is an extra information I need to confirm. For SMPL-X model, Did you use 10-shape version or 300-shape version?

[lizhe00]

10-shape version.

[trThanhnguyen]

Hi, @475651582. Thanks for raising the issue. Could you please describe how you managed to get the 'smpl_params.npz' for the subject? Many thanks.