HumanML3D

HumanML3D copied to clipboard

This can take a (very) few minutes. This way, you know what is taking so long.

felixdivo

Skip LICENSE file that exists in the downloaded datasets form the AMASS website. This previously stopped the processing until it was removed manually. Now it just skips it. Closes #54.

felixdivo

Fix an error where a directory was missing Allow to seamlessly take up interrupted work

felixdivo

question about data

4

In raw_pose_processing.ipynb, all data is saved in joints folder. What is the meaning of the data in the files under the joints folder? Data's shape is like (36, 24, 3)...

leisurehippo

humanml3d convert

1

I want to convert my human motion dataset into a format like humanml3d, what information should my dataset contain ？I have position information.

Deante-dx

Forward Direction in inverse kinematic

1

hello! thanks for your great job! I have some question about the forward direction in inverse kinematics。 The forward direction is calculated by four joints (r_hip, l_hip, sdr_r, sdr_l) and...

huangsheng95

Missing subdataset "humanact12" when run script raw_pose_processing.ipynb

5

During run raw_pose_processing.ipynb, It report mistake. ```python # codes in raw_pose_processing.ipynb for i in tqdm(range(total_amount)): source_path = index_file.loc[i]['source_path'] new_name = index_file.loc[i]['new_name'] data = np.load(source_path) start_frame = index_file.loc[i]['start_frame'] end_frame = index_file.loc[i]['end_frame']...

spzhuang

Joint rotation output

2

Hi, Eric! I'm trying to make a real time AI controlled motion agent. Because the skeleton of the agent is different, I have to convert the XYZ coordinate to the...

XinandYu

Running the repository in Torch CPU. In mdm.py for the class MDM in the method "load_and_freeze_clip". def load_and_freeze_clip(self, clip_version): clip_model, clip_preprocess = clip.load(clip_version, device='cpu', jit=False) # Must set jit=False for...

JohanValero

how can we perform more data augmentation?

2

Hi, Sorry, I am a little bit new to this area. I am wondering how to augment the current HumanML3D data by simply concatenating the same or different motion sequences...

sukun1045