Bruno Korbar

Documenting current debugging: - [x] verify if the segfaults are caught in the test suite -- Feb 28th - [x] update test suite to check for the problematic video (#5543)...

Hi all, So I've dug quite a bit into this for the past two weeks (and am continuing to do so), and there are a few confusing factors. The error,...

Hi all, So I've dug quite a bit into this for the past two weeks (and am continuing to do so), and there are a few confusing factors. The error,...

^ This is likely due to the 3.9 python; for some reason that works.

https://github.com/bjuncek/VMZ hopefully will be merged soon

you can easily repurpose the reference script from torchvision for that :)

Also note that a LOT of the original kinetics-400 videos are actually not existing anymore. I can try to run them for you on their snapshot in a few days...

+ not sure if it's in the repo, but you'd have to implement the FCN testing scheme from the paper that I believe was used there.

> Regarding evaluation strategy. Reading > > https://research.fb.com/wp-content/uploads/2019/05/Large-scale-weakly-supervised-pre-training-for-video-action-recognition.pdf > > > 1. Experiments > > The fc-only experiments should be good enough for a first step here: extract features for...

Other than parameter fine-tuning (I'd say 15 epochs, LR:1e-4 for the convolutional blocks and 1e-3 for the FC layer), make sure the evaluation script is correct. If I recall correctly,...