Preparing IEMOCAP dataset

[joeoct93]

I am trying to try out your USEV code, and I have a few questions about preparing the dataset.
I read that for the iemocap, "you need to adjust the --num_failed_det; the face detector S3FD/MTCNN; the face detector confidence threshold for different videos". Could you tell me how I need to adjust these?
And what should I put in for the video reference? Also, do we need to do this (0_crop_face,py) for just Ses02F_impro01.avi, since that's the default, or is it for other avi files as well?
I would greatly appreciate it if you answered these questions for me. Thank you.