Details about encoder pretrain

I am very interested in your work. I was wondering if you could kindly share how "encoder_canopus.pt" is pretrained. Any details regarding the pretraining parameters would be greatly appreciated.