HumanML3D
HumanML3D copied to clipboard
HumanML3D: A large and diverse 3d human motion-language dataset.
Hello, I think `amass_to_pose` can be vectorized (not iterating a single pose at a time but all poses in a sequence (a npz file) at a single time.). I tried...
When I run this file, the fps value is 0 and pos_ Although a series of corresponding folders for datasets can be created under the data folder, no npy files...
Hi @EricGuo5513, I followed all the steps of raw_pose_processing.ipynb and I am facing this error which I cannot get rid of.  It is not clear from...
Hi, thank you very much for the great work. I notice that only a subset (18/24) of AMASS is used in HumanML3D. I am wondering why selecting these 18 subsets...
``` def get_rifke(positions): '''Local pose''' positions[..., 0] -= positions[:, 0:1, 0] positions[..., 2] -= positions[:, 0:1, 2] '''All pose face Z+''' positions = qrot_np(np.repeat(r_rot[:, None], positions.shape[1], axis=1), positions) return positions...
Thanks for the dataset. I want to generate a motion in Korean. so, I want to translate the text and make it Korean, is that okay?
At the end of `motion_representation.ipynb`, I get offsets of `0.0034465478` and `0.018278915`. Is that acceptable? Or should it be closer to `0.0`? The comment just says "If it's aligned with...
Thank you
Hi @EricGuo5513 ! Thanks for the great dataset. I am struggling to read how you convert SMPL to the 263 dimensional vector that is used with the AMASS dataset. I...
Hi, thanks for your great work. I am using this dataset and find that some data are in left-hand coordinate system, even though most data are in right hand coordinate...