C4W2 - Encoder Layer: self.mha call arguments

Hello, sirs!
I wonder if I correctly follow the logic behind the argument in self.mha call.

self_mha_output = self.mha(x, x, x, mask)

Do we have to use the same x parameter for all query, keys and values tensors because in step 6.1 Encoder Layer we build a self-attention mechanism?

Yes.

1 Like

@TMosh, thank you!