Rohit Girdhar

Hi @haofengac apologies for the delay in responding. So when running locally, no logs will be printed to the shell -- they will get stored in the output directory (which...

Hi @haofengac unfortunately I don't have a good sense of what might be going on either. Maybe try to submit a 1 GPU job and see if that works at...

Hi @ttran1904 Something seems wrong with the data setup -- the code doesn't find action labels for any of the samples. Was the data setup correctly as described in the...

Yes you shouldn't need videos for experiments on pre-extracted features. Can you share the full output log? It is strange that it is not able to read any labels --...

I don't see anything strange in the logs (it no longer complains about missing labels either, unlike the screenshot you shared earlier). I see you are running it with submitit,...

Hi @ttran1904 apologies for the delay in responding. Yes that should not happen -- the TSN model is fairly lightweight. Can you check what your GPU memory usage is before...

Hi, For EK100 I didn't report or optimize for top-1 accuracy since the official metric for that dataset is mean recall@5 (and I found that top-1 and recall@5 metrics behave...

Yes you can ignore that warning.

you are running it in `test_only` mode but not initializing with the trained checkpoint (it's initializing from the VIT checkpoint, not the AVT trained checkpoint). You can set the checkpoint...

Hi @CodyQ3 , ah yes you can ignore that part of the log; it should ideally print the running accuracy but I never got around to fixing it.