IDEA-Research / Motion-X

[NeurIPS 2023] Official implementation of the paper "Motion-X: A Large-scale 3D Expressive Whole-body Human Motion Dataset"
https://motion-x-dataset.github.io
Other
562 stars 15 forks source link

BMLhandball #8

Closed ChangeNext closed 1 year ago

ChangeNext commented 1 year ago

Thank for Huge Projects I have one question BMLhandball doesn't have SMPL-X G... only have SMPL+H G

shunlinlu commented 1 year ago

@ChangeNext, thanks for your interest.

Please refer to the following code: here. We actually ignore these datasets, as there is no SMPL-X G.

ChangeNext commented 1 year ago

Thank you for your reply. I have one more question. Can you provide the VQ-VAE code you ran with this dataset?

shunlinlu commented 1 year ago

@ChangeNext,

We have implemented VQ-VAE using the codebase from T2m-GPT. Please stay tuned for further updates, as we have integrated the implementations of MLD, T2M-GPT, and other related components into a unified codebase for our new motion generation framework called HumanTOMATO. We are in the process of preparing the release of the source code, pre-trained models, and pre-trained VQ tokens, which will be made available on the HumanTOMATO repository in the near future. Stay tuned for more updates and resources!

ChangeNext commented 1 year ago

Thank you for reply An error occurred while running the GRAP.py file.

File "grab.py", line 99, in data = {k: data[k].item() for k in data.files} File "grab.py", line 99, in data = {k: data[k].item() for k in data.files} AttributeError: 'bytes' object has no attribute 'item'

I download GRAB papameter s1~s10 but error occurred Why am I getting these errors?

shunlinlu commented 1 year ago

Hi, @ChangeNext

What's your format of the grab dataset? It should be npz files. You can use np.load to load the npz, and data.files to check the keys of the data.

ChangeNext commented 1 year ago

I'm sorry, that was my mistake, thank you so much for your kind words.