Closed hasibzunair closed 3 years ago
Hi,
Thanks for the suggestion. I think the paper is more of a supervised pretraining approch using a SSL method, a bit different than the ones listes.
Alright. Closing this then!
Hi @hasibzunair
You were right, I've just seen that the method is effective in SSL real-world settings, so the paper is definitely relevant to SSL, just added it, sorry for my earlier response,
Thanks.
No problem!
Hi,
Maybe this can be added.
Self-training with Noisy Student improves ImageNet classification Published in CVPR 2020