Padding and stride are two simple knobs in convolutional layers that control output size and how much neighboring pixels influence each other. In this task, you’ll compute the output height/width of a 2D convolution given an input image size and conv settings.

## Requirements

Implement the function
```python
import numpy as np

def conv2d_output_shape(h, w, k_h, k_w, stride, padding):
    \"""
    Computes the output (height, width) of a 2D convolution.

    Uses the standard formula:
        out_h = floor((h + 2*padding - k_h) / stride) + 1
        out_w = floor((w + 2*padding - k_w) / stride) + 1

    Args:
        h (int): Input height.
        w (int): Input width.
        k_h (int): Kernel height.
        k_w (int): Kernel width.
        stride (int): Stride (same for height and width).
        padding (int): Zero-padding on all sides (same for height and width).

    Returns:
        np.ndarray: Array of shape (2,) containing [out_h, out_w].
    \"""
    # Your solution below
```

**Rules:**
- Use $$\text{out} = \left\lfloor \frac{\text{in} + 2p - k}{s} \right\rfloor + 1$$ for both height and width.
- Return the output shape as a NumPy array: `np.array([out_h, out_w])`.
- Do not simulate convolution over pixels; just compute the shape.
- Use only NumPy and Python built-in libraries.
- Keep it in a single Python function.

## Example

```python
h, w = 5, 5
k_h, k_w = 3, 3
stride = 2
padding = 1

conv2d_output_shape(h, w, k_h, k_w, stride, padding)
```

Output:
```python
np.array([3, 3])
```

Batch Retry Queue

Partition Labels

Padding and stride effects

Solve this easy level ML coding question from General. The premier LeetCode for ML platform to master Python, SQL, and Machine Learning algorithms.

658. Padding and stride effects

Examples

Constraints

Similar Questions

Argument	Type
h	int
w	int
k_h	int
k_w	int
stride	int
padding	int