fendi
Results
2
issues of
fendi
@LTH14 I have not yet quite understood that why use `zero_module` in ResNet and SimpleMLP , it just zero out directly the tensors (input_rep_tensor + time_tensor+condition_tensor) ( torch.all(h==0)=> True), which...
How was the entry numbers configured in the codebook? For example, K = 8192 for 1M, and K = 16384 for 2M. Could you please give detailed illustration for that?