fine-tune my own video dataset

[ustc-wlw]

hello yoosan:
I plan to classify the 'fight behavior' with fine-tuning the pre-trained I3D model on my own dataset. Can you release the training and test code? thanks a lot!