1D Max Pooling

Perform 1D max pooling on an input tensor:

$$\text{output}[i] = \max_{m=0}^{k-1} \text{input}[S \cdot i + D \cdot m - P]$$

The max pooling operation slides a window of size $k \times k$ over the input tensor with stride $S$, dilation $D$, and padding $P$, computing the maximum value within each window position.

Input:

• Matrix input of size $\text{H}$ (input tensor)
• kernel_size ($k$): Size of the pooling window
• stride ($S$): Step size between window positions
• padding ($P$): Number of zero-padding elements added on all sides
• dilation ($D$): Spacing between kernel elements

Output:

• Matrix output of size $\text{H}_{\text{out}}$ where: $\text{H}_{\text{out}} = \left\lfloor\frac{\text{H} + 2P - D(k-1) - 1}{S}\right\rfloor + 1$

Notes:

• All tensors are stored in row-major order
• Zero padding is applied when specified by the padding parameter
• For values outside the input boundaries (after padding), use negative infinity
• Dilation controls the spacing between kernel elements, creating an effective kernel size of $D(k-1) + 1$
• This problem is adapted from KernelBench