Question on data NaN value

[MingCongSu]

When I run cal_mean_variance.ipynb, one of the data (007975.npy) and its mirror version (M007975.npy) contain much NaN value, which leads to a tiny double-check difference in double-check steps.

Execution result of cal_mean_variance.ipynb:

HumanML3D\new_joint_vecs\007975.npy is like:

Is this okay or did I do something wrong?

[EricGuo5513]

Hi, what is the size of this data matrix? Sometimes, it can be caused by different versions of numpy or scipy libraries.

…

On Tue, 16 Jan 2024 at 03:06, MingCongSu ***@***.***> wrote: When I run cal_mean_variance.ipynb, one of the data (007975.npy) and its mirror version (M007975.npy) contain much NaN value, which leads to a tiny double-check difference in double-check steps. Execution result of cal_mean_variance.ipynb: image.png (view on web) <https://github.com/EricGuo5513/HumanML3D/assets/113020932/f5c3ae03-8b92-4c10-b70a-9d4327f1e095> HumanML3D\new_joints\007975.npy is like: image.png (view on web) <https://github.com/EricGuo5513/HumanML3D/assets/113020932/c2719712-1a11-4980-97c3-683ad1b4b7c3> Is this okay or did I do something wrong? — Reply to this email directly, view it on GitHub <#118>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AKRYNB2AQSNEFFYHODPWE73YOZGLJAVCNFSM6AAAAABB4QZ7XKVHI2DSMVQWIX3LMV43ASLTON2WKOZSGA4DGNJSGQZTSMQ> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[MingCongSu]

Hi @EricGuo5513, thanks for replying to me.
The shape of this data matrix (HumanML3D\new_joint_vecs\007975.npy) is (84, 263).
In my environment, the versions of numpy==1.21.6 and scipy==1.7.3 .

[EricGuo5513]

I saw the numpy version used for creating this dataset is 1.18.5. Not sure if this makes big difference. Please email me at cguo2@ualberta.ca if this does not work.

[MingCongSu]

I reinstalled the numpy==1.18.5 but the result is the same.

I checked the result before running motion_representation.ipynb and producing new_joint_vecs and new_joints, which is the npy in /joints, but it seems normal?

007975.npy	M007975.npy