Open diaodeyi opened 2 years ago
By the way the self.proj = nn.Linear(dim_out, dim_out)
Means FFN only projection with same dimension?
By the way the
self.proj = nn.Linear(dim_out, dim_out)
Means FFN only projection with same dimension?
Yes, it is a part of the attention mechanism.
I want to konw why there isn't any to the function get_cls_model and compute_macs