nn.AdaptiveAvgPool2d¶

class lucid.nn.AdaptiveAvgPool2d(output_size: int | tuple[int, int])¶

The AdaptiveAvgPool2d performs adaptive average pooling on a 2D input tensor, dynamically determining kernel size, stride, and padding to produce a specified output size.

Class Signature¶

class AdaptiveAvgPool2d(nn.Module):
    def __init__(self, output_size: tuple[int, int] | int)

Parameters¶

output_size (tuple[int, int] | int): The desired output size \((H_{out}, W_{out})\). If an integer is provided, the same size is used for both dimensions.

Attributes¶

output_size (tuple[int, int] | int): Stores the target output size for the pooling operation.

Forward Calculation¶

The AdaptiveAvgPool2d module applies adaptive average pooling on a 2D input tensor of shape \((N, C, H, W)\), producing an output tensor of shape \((N, C, H_{out}, W_{out})\). The kernel size, stride, and padding are dynamically determined based on the input size \(H, W\) and the desired output size \(H_{out}, W_{out}\).

Behavior¶

The forward calculation for the module follows the same logic as the functional API (adaptive_avg_pool2d). It dynamically adjusts pooling parameters to ensure the target output size is achieved.

Examples¶

Basic Example

import lucid.nn as nn

# Input tensor with shape (1, 3, 32, 32)
input_ = Tensor([[[[1.0] * 32] * 32] * 3])

# AdaptiveAvgPool2d module with output size (8, 8)
pool = nn.AdaptiveAvgPool2d(output_size=(8, 8))

output = pool(input_)
print(output)  # Shape: (1, 3, 8, 8)

Output Explanation

The input tensor is adaptively pooled to produce a tensor with height and width of 8, averaging over evenly spaced regions.

Advanced Example with Variable Batch Size

# Input tensor with batch size 2
input_ = Tensor([
    [[[1.0] * 16] * 16],
    [[[2.0] * 24] * 24]
])

pool = nn.AdaptiveAvgPool2d(output_size=(4, 4))

output = pool(input_)
print(output)  # Shape: (2, 3, 4, 4)

Explanation

The pooling dynamically adjusts for each input’s height and width, producing a consistent output size of (4, 4) across all samples in the batch.