hello, how can i realize that the feature_encoder in tinyimagenet_training_contrastive.py has been train enough

[yigguo]

this is a wonderful job! but when i look through your code, i have a problem:

only can i judge from the loss in tinyimagenet_training_contrastive.py�, if i want to get the classify result of this encoder with the close-set data, what should i do? looking for your reply.

thanks a lot.

[Raymondkeepsgoing]

请问你在训练的时候loss一般是多大呀，我这边loss出现了很大的情况。