Saurav Gupta

Results 3 comments of Saurav Gupta

I am also facing the same issue. Did someone find the solution, please share.

Hi @Oussamab21, The first layer of the model is a convolution layer hence you don't need to change the model for processing different image dimensions. You need to create a...

> 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.