Since the data link you provided does not contain image and mask, we processed the data according to InstantAvatar, but the mask result was in npy format, which was inconsistent with actual needs. We tried to use scripts/preprocess_PeopleSnapshot.py to process the data, but the camera.pkl parameter was not provided. can you update the parameter configuration file? There are still problems with the effect of the mask training we converted ourselves.
Since the data link you provided does not contain image and mask, we processed the data according to InstantAvatar, but the mask result was in npy format, which was inconsistent with actual needs. We tried to use scripts/preprocess_PeopleSnapshot.py to process the data, but the camera.pkl parameter was not provided. can you update the parameter configuration file? There are still problems with the effect of the mask training we converted ourselves.