Convolution Output Shape

Compute the output size of a convolution along one dimension from the input size W, kernel size K, padding P (added on both sides) and stride S: $\lfloor (W+2P-K)\mathrm{/}S\rfloor +1\backslash lfloor\; (W+2P-K)/S\backslash rfloor+1$.

Implement solve(W,K,P,S).

Returns. Return an integer. A kernel larger than the padded input gives a value of 0 or less, meaning no valid position.

Examples

Example 1

Input:

solve(5, 3, 0, 1)

Output:

3

Example 2

Input:

solve(7, 3, 1, 2)

Output:

4