sthalles / PyTorch-BYOL

PyTorch implementation of Bootstrap Your Own Latent: A New Approach to Self-Supervised Learning
480 stars 72 forks source link

loss is a neg value? #5

Closed BrettLL closed 4 years ago

BrettLL commented 4 years ago

when I print loss value in training phase, Found that loss is a neg value after epoch 40(about -3.45), is that right? confused~

sthalles commented 4 years ago

It has been fixed. Thanks for the message.