MLAD

Implementation of paper "Modeling Multi-Label Action Dependencies for Temporal Action Localization"

Here are the trained models

Charades : https://drive.google.com/file/d/1tna5PLkFm2A9RA45sOtCnG6yx6mGHw4j/view?usp=sharing MultiTHUMOS : https://drive.google.com/file/d/1vXq-y68hC4Qe6N1PBk3DlqGjOWhP9Vsc/view?usp=sharing

These are the models with 5 MLAD layers and give the best results on the corresponding datasets.