Problems

1 / 1

Max Pooling Forward

Neural Networks

Medium

Max pooling is a downsampling operation commonly used in convolutional neural networks to reduce spatial dimensions while retaining the most prominent features. It slides a window over the input and takes the maximum value in each window position.

Given a 2D input matrix X (height H, width W), a pool size p, and a stride s, compute the max pooling output.

Algorithm

Compute the output dimensions:

H_{out} = \left\lfloor \frac{H-p}{s} \right\rfloor + 1

W_{out} = \left\lfloor \frac{W-p}{s} \right\rfloor + 1

For each output position (i, j), extract the p x p window and take the maximum:

\text{out}[i][j] = \max_{0 \le a < p,\; 0 \le b < p} X[i \cdot s + a][j \cdot s + b]

Return the pooled two-dimensional list.

Loading visualization...

Examples

Input: X = [[1, 2, 3, 4], [5, 6, 7, 8], [9, 10, 11, 12], [13, 14, 15, 16]], pool_size = 2, stride = 2

Output: [[6, 8], [14, 16]]

Explanation: Each non-overlapping 2 by 2 window contributes its maximum.

Input: X = [[1, 5, 3], [4, 2, 6], [7, 8, 0]], pool_size = 2, stride = 1

Output: [[5, 6], [8, 8]]

Hint 1

Compute output height and width with integer division before scanning windows.

Hint 2

Initialize a window maximum from its top-left value, then compare the remaining entries.

Requirements

Apply 2D max pooling with the given pool size and stride
Each output element is the maximum value in its pooling window
Handle both overlapping (stride < pool_size) and non-overlapping (stride = pool_size) cases
Return the pooled output as a list of lists of floats or ints

Constraints

X has at least pool_size rows and pool_size columns
pool_size >= 1, stride >= 1
Return a list of lists
Time limit: 300 ms

Try Similar Problems

Max Pooling 2d Average Pooling 2d Global Avg Pooling Simple Cnn Layer Conv2d Image Filtering

Prev Next

Max Pooling Forward

Neural Networks

Medium

Given a 2D input matrix X (height H, width W), a pool size p, and a stride s, compute the max pooling output.

Algorithm

Compute the output dimensions:

H_{out} = \left\lfloor \frac{H-p}{s} \right\rfloor + 1

W_{out} = \left\lfloor \frac{W-p}{s} \right\rfloor + 1

For each output position (i, j), extract the p x p window and take the maximum:

\text{out}[i][j] = \max_{0 \le a < p,\; 0 \le b < p} X[i \cdot s + a][j \cdot s + b]

Return the pooled two-dimensional list.

Loading visualization...

Examples

Input: X = [[1, 2, 3, 4], [5, 6, 7, 8], [9, 10, 11, 12], [13, 14, 15, 16]], pool_size = 2, stride = 2

Output: [[6, 8], [14, 16]]

Explanation: Each non-overlapping 2 by 2 window contributes its maximum.

Input: X = [[1, 5, 3], [4, 2, 6], [7, 8, 0]], pool_size = 2, stride = 1

Output: [[5, 6], [8, 8]]

Hint 1

Compute output height and width with integer division before scanning windows.

Hint 2

Initialize a window maximum from its top-left value, then compare the remaining entries.

Requirements

Apply 2D max pooling with the given pool size and stride
Each output element is the maximum value in its pooling window
Handle both overlapping (stride < pool_size) and non-overlapping (stride = pool_size) cases
Return the pooled output as a list of lists of floats or ints

Constraints

X has at least pool_size rows and pool_size columns
pool_size >= 1, stride >= 1
Return a list of lists
Time limit: 300 ms

Try Similar Problems

Max Pooling 2d Average Pooling 2d Global Avg Pooling Simple Cnn Layer Conv2d Image Filtering

Case 1

Case 2

X =

Accepts: array

pool_size =

Accepts: number

stride =

Accepts: number

Given a 2D input matrix X (height H, width W), a pool size p, and a stride s, compute the max pooling output.

Algorithm

Compute the output dimensions:

H_{out} = \left\lfloor \frac{H-p}{s} \right\rfloor + 1

W_{out} = \left\lfloor \frac{W-p}{s} \right\rfloor + 1

For each output position (i, j), extract the p x p window and take the maximum:

\text{out}[i][j] = \max_{0 \le a < p,\; 0 \le b < p} X[i \cdot s + a][j \cdot s + b]

Return the pooled two-dimensional list.