Fix duplicate data loading for VOCASET

theEricMa / DiffSpeaker

This is the official repository for DiffSpeaker: Speech-Driven 3D Facial Animation with Diffusion Transformer

137 stars 18 forks source link

Closed leohku closed 7 months ago

leohku commented 8 months ago

data_splits is prepared twice in the original code for loading VOCASET.

leohku commented 8 months ago

Sorry, added additional files in this PR - please feel free to only refer to the changes in alm/data/vocaset.py.

leohku commented 8 months ago

Additional files are removed - sorry for that.