search

szq0214 / Un-Mix

Un-Mix: Rethinking Image Mixtures for Unsupervised Visual Representation Learning.
149 stars 14 forks source link

L_m Loss #4

Closed juingzhou closed 2 years ago

juingzhou commented 3 years ago

Hello, I'm very interested in your work. I want to know L_M loss， can you tell me the calculation method ?

szq0214 commented 3 years ago

Hi @juingzhou, the calculation method can be l_m = lam criterion(ori_images, mixed_images) + (1 - lam) criterion(ori_images, mixed_images_flip).