Hi,

Thanks for your organized repository and for sharing your code. Your paper about Self-Supervised Meta-Learning is interesting. I have a few questions and I appreciate it if you could help me understand more about the details of your paper.

In your paper, you set "Support samples per task = 80", "Query samples per task = 10", "Adaptation Steps (G) = 7" , "Meta-training Epochs = 1". Meanwhile, in your codes, you fix eval epochs to 5 by setting "inner_epochs = 5". I wonder how you choose these hyperparameters.

Besides, is it means that every SMLMT task used for meta training contains exactly 80 + 10 examples ?

Thanks.