Implement Scaled Dot-Product Attention

Problem: Implement Scaled Dot-Product Attention Implement single-head, unmasked scaled dot-product attention: [ \operatorname{Attention}(Q,K,V)=\operatorname{softmax}\left(\frac{QK^T}{\sqrt{d_k}}\right)V ] Where: Q is a query matrix of shape q × d K is a key matrix of shape k × d V is a value matrix of shape k × dv Softmax is computed independently for each row of attention scores. Print every output value rounded to 6 decimal places. Input Format First line: four integers q k d dv Next q lines: matrix Q, each with d floating-point values Next k lines: matrix K, each with d floating-point values Next k lines: matrix V, each with dv floating-point values Output Format Print q lines. Each line contains dv floating-point values separated by spaces, representing the attention output matrix. Constraints 1 <= q, k <= 200 1 <= d, dv <= 128 Matrix values are in [-100, 100] Example Input 1 2 2 2 1 0 1 0 0 1 10 0 0 20 Output 6.697616 6.604767 Example Input 1 2 2 2 1 0 1 0 0 1 10 0 0 20 Output 6.697616 6.604767