search

Pilhyeon / Learning-Action-Completeness-from-Points

Official Pytorch Implementation of 'Learning Action Completeness from Points for Weakly-supervised Temporal Action Localization' (ICCV-21 Oral)
MIT License
84 stars 15 forks source link

For GTEA and BEOID #4

Open PHDJieFu opened 3 years ago

PHDJieFu commented 3 years ago

Hello, thanks for your excellent job. I am interested in your work so much! May I ask if the extracted features on ActivityNet, GTEA and BEOID will be released？

Pilhyeon commented 3 years ago

Hello, thanks for your interest! For the experiments on GTEA and BEOID, we utilized the features provided by SF-Net and you can find them here. On the other hand, we extracted the features of ActivityNet videos by ourselves, but they are not supposed to be publicly available due to some internal issues. As far as I know, there are several public repositories that provide the features for ActivityNet, and you may want to use them.

PHDJieFu commented 3 years ago

Hello, thanks for your interest! For the experiments on GTEA and BEOID, we utilized the features provided by SF-Net and you can find them here. On the other hand, we extracted the features of ActivityNet videos by ourselves, but they are not supposed to be publicly available due to some internal issues. As far as I know, there are several public repositories that provide the features for ActivityNet, and you may want to use them.

Thanks a lot! Can you realease the code of dataset and data_loader class (like the thumos_features.py) for GTEA and BEOID?

chenrxi commented 2 years ago

Hello, thanks for your interest! For the experiments on GTEA and BEOID, we utilized the features provided by SF-Net and you can find them here. On the other hand, we extracted the features of ActivityNet videos by ourselves, but they are not supposed to be publicly available due to some internal issues. As far as I know, there are several public repositories that provide the features for ActivityNet, and you may want to use them.

Thanks a lot! Can you realease the code of dataset and data_loader class (like the thumos_features.py) for GTEA and BEOID?

Do you reproduce the result of GTEA and BEOID? Thank you.

mn-liu commented 10 months ago

@PHDJieFu Can you reproduce the result of GTEA and BEOID? I try but there is a big gap between the results and the paper