Closed mactavish91 closed 6 months ago
Hello, may I ask if the learning rates of different layers of vit and text encoders are the same during pre training?
@mactavish91 in default training, they are the same, EVA CLIP for example was a fork of this that changed LR across the layers in phases of their training
Hello, may I ask if the learning rates of different layers of vit and text encoders are the same during pre training?