TengdaHan / DPC

Video Representation Learning by Dense Predictive Coding. Tengda Han, Weidi Xie, Andrew Zisserman.

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Why are you shuffling the sequence of images?

sarimmehdi opened this issue · comments

I saw in your code that you shuffle your images around (main.py). Don't you want to preserve the sequence?

Hi, this RandomSampler is to randomly form batches from the entire video dataset (e.g. randomly form batch size 64 from 300,000 videos), not shuffle frames. Frames are definitely not shuffled.

Thank you very much, I forgot that you were using videos and not frames that were already provided before-hand. This clears up a lot, actually, my model seems to be performing much better now (before, the order of all the frames was random lol)