3 Issues closed by 1 person

• Unable to reproduce the effect when generating dataset/horsejump/masks in preprocess with vitL

  #13 closed Jul 28, 2024

• Benchmark results different from the original papers

  #15 closed Jul 28, 2024

• The process of generating the trajectory is much slower than cotracker. How to accelerate the process?

  #18 closed Jul 28, 2024

3 Issues opened by 3 people

• Cycle-Consistency Loss

  #26 opened Jul 28, 2024

• Insufficient memory during data preprocessing stage

  #25 opened Jul 26, 2024

• ModuleNotFoundError: No module named 'yaml' and 'einops'

  #24 opened Jul 25, 2024