imprinted-weights

imprinted-weights copied to clipboard

Hi, Is the magic number[100] in new_weight a class number? new_weight = torch.zeros(100, d_emb) for i in range(100): tmp = output_stack[target_stack == (i + 100)].mean(0) if not random else torch.randn(d_emb,...

ghost

About the pre-training

3

It seems that you pre-train the model with all the training images, including the base and novel classes. I argue it is unfair to use the novel classes to train...

kikyou123

Hi, I just wanted to ask about num_classes adjustment. If I want to use 120 classes as base and 80 as novel, what should the num_classes value be in models.py

MohammadBelalIrshaid