Hi, thanks for your great work and releasing such a valuable dataset for the sign language community!
I noticed that the current phonex_pkls_cropFalse_shapeFalse folder only contains around 596 .pkl files (seems to be part of the test set).
According to the paper, the phonex dataset should include about 8.25K samples.
Could you please provide the full train/dev/test .pkl annotation files of the avatar dataset?
Hi, thanks for your great work and releasing such a valuable dataset for the sign language community!
I noticed that the current phonex_pkls_cropFalse_shapeFalse folder only contains around 596 .pkl files (seems to be part of the test set).
According to the paper, the phonex dataset should include about 8.25K samples.
Could you please provide the full train/dev/test .pkl annotation files of the avatar dataset?