search

shashankvkt / DoRA_ICLR24

This repo contains the official implementation of ICLR 2024 paper "Is ImageNet worth 1 video? Learning strong image encoders from 1 long unlabelled video""
58 stars 8 forks source link

Running on a multi-video dataset #3

Open kunalmessi10 opened 3 months ago

kunalmessi10 commented 3 months ago

Hi,

Thanks for sharing the code for your work. I was able to get the training running with the Venice video. Could you suggest how can I it run on a dataset with multiple videos, similar to the WT-All videos case in your paper. Thanks in advance!

kristinat8 commented 2 months ago

Hi, Would you mind sharing the configuration you used for training on Venice? I’m really curious about the time it takes you to train one epoch, as my training duration is quite long

kunalmessi10 commented 2 months ago

Hi, for me it was taking 2-3 days per epoch. So yes, quite long.