Haodong Duan

Hi, @Kewenjing1020 , recently I'm working on my research project. Perhaps I will add X3D several weeks later. You are welcome to raise a PR if you decide to implement...

Thanks for pointing out the bug, I will provide a fix PR.

Seems the error log is not full. Please paste the complete error log and provide the command you run which lead to the error log.

Hi, what do you mean by 'the name of the data read by each GPU', do you mean 'img_group_0'?

@gengenkai , you can check if there are some features you are interested in.

Hi, @hukkai , if you are eager to implement this feature in MMAction2, you can provide me a list of missing videos to be. I will see if these videos...

> Hi @kennymckormick I can implement the AVA-Kinetics dataset and training. Training on AVA-Kinetics indeed helps the performance on AVA a lot. > > For the AVA-Kinetics, I plan to...

> Hi @kennymckormick I can implement the AVA-Kinetics dataset and training. Training on AVA-Kinetics indeed helps the performance on AVA a lot. > > For the AVA-Kinetics, I plan to...

Hi, cheesezoella, I recommend you benchmark the inference speed by running testing scripts in PYSKL (by default it's 10-clip testing).

Unfortunately, you cannot use GCN models in demo_video_structualize, since currently we use k400 action classes for action recognition demo and AVA action classes for skeleton action recognition demo. We only...