L-Sun/LGTM

The difference of performance

Closed this issue · 3 comments

Hi! I'm confused about why is the performance of MDM, MLD, and MotionDiffuse in Table 2 different from that in the original paper. Could you tell me which paper the performance in Table 2 comes from?

The Table.2 use evaluation/run.py to calculate metrics. The difference of metrics may be from these reasons:

  • since the implement of dataloader has some differences from those paper, the different order of loading makes the performance not exactly same
  • This result was obtained through distributed computing, so there may be some differences.
  • random seed influence

For more details, you can review the implement of evaluation/run.py and dataloader

Thank you for your feedback.