C5_W4_A1_Transformer_Subclass_v1 Exercise 4

For Exercise 4, i am meant to pass Q,K,V matrices to multi attention layer, issue is i have no idea where these matrices are, they were not defined as instance or class variables and were not passed in the function call, the parameters in the function call are self, x, training, and mask, kindly help me understand how to go about this, thanks

Hi @faraday

Welcome to the community.

What course are you refereeing to?

It turns out you posted on the general category.

best regards

ohh thats true, i’ll change it i am referring to deep learning specialization sequence models

1 Like

For self-attention, the x matrix is used for each of K, Q, and V.