Hello, I'm hoping you can help me understand why is the dimension of A TDN, TDH, since the dimension of hidden_states is TD1, it becomes TD after dense.
ATA += W * (hidden_states @ hidden_states.t() ) #1956,1956 Why did ATA calculate it this way？
Can you recommend materials to solve this problem？ Looking forward to your reply.
liuxiaozhu01
commented
2 months ago
Hello, I'm hoping you can help me understand why is the dimension of A TDN, TDH, since the dimension of hidden_states is TD1, it becomes TD after dense.
ATA += W * (hidden_states @ hidden_states.t() ) #1956,1956 Why did ATA calculate it this way？ Can you recommend materials to solve this problem？ Looking forward to your reply.