AvgPoolΒΆ

Pools the input by taking the average over a window.

Abstract Signature:

AvgPool(inputs: Tensor, window_shape: Tuple[int, ...], strides: Tuple[int, ...], padding: str = VALID)

PyTorch

API: torch.nn.functional.avg_pool2d
Strategy: Direct Mapping

Keras

API: keras.ops.average_pool
Strategy: Direct Mapping

TensorFlow

API: tf.nn.avg_pool
Strategy: Direct Mapping

Flax NNX

API: flax.nnx.avg_pool
Strategy: Direct Mapping