Inference on caption activitynet test dataset

[Leo-Yuyang]

Dear author, thank you for your great work! I want to inference on the test set of Dense-Captioning Events in Videos(https://cs.stanford.edu/people/ranjaykrishna/densevid/).
What should I do?

Following the above instruction and just modifying the video folder path to the test set video folder, is this the right way?
Or should I download some C3D features from some where and then do something else?
Thank you for your help and look forward to your reply!