Requirements:

1.Mxnet installed

2.Download UCF-101

3.Run avi2jpg.py to split the video of UCF-101

4.Run train_ucf101.py to train the model

Reference:

1.Thanks the author Du tran's code: C3D-caffe

2.C3D: Generic Features for Video Analysis