I have a question of the shape of weights of lstm.
If the input to the lstm layer is (batch_size, seq_length, input_size), then x_t would be (batch_size, input_size).
And as the documentation say so, the shape of W_ii is (hidden_size*4, input_size).
But here how possibly matrix multiplication of W_ii and x_t can be happen?
What did I miss here? Can anyone help me please