Function Differences with tf.nn.avg_pool2d

tf.nn.avg_pool2d

tf.nn.avg_pool2d(
    input,
    ksize,
    strides,
    padding,
    data_format='NHWC',
    name=None
) -> Tensor

For more information, see tf.nn.avg_pool2d.

mindspore.nn.AvgPool2d

mindspore.nn.AvgPool2d(
    kernel_size=1,
    stride=1,
    pad_mode='valid',
    data_format='NCHW'
)(x) -> Tensor

For more information, see mindspore.nn.AvgPool2d.

Differences

TensorFlow: Performs average pooling on the input Tensor.

MindSpore: MindSpore API implements the same function as TensorFlow, and only the parameter names and the way of using input Tensor are different.

Categories	Subcategories	TensorFlow	MindSpore	Differences
Parameters	Parameter 1	input	x	Same function, used to input a 4-dimensional Tensor. The data input format is different
	Parameter 2	ksize	kernel_size	Same function, different parameter names, no default values for TensorFlow
	Parameter 3	strides	stride	Same function, different parameter names, no default values for TensorFlow
	Parameter 4	padding	pad_mode	Same function, different parameter names, no default values for TensorFlow
	Parameter 5	data_format	data_format	Same function, different default values of parameters
	Parameter 6	name	-	Not involved

Code Example

The two APIs achieve the same function and have the same usage.

# TensorFlow
import tensorflow as tf
import numpy as np

y = tf.constant([[[[1, 0, 1], [0, 1, 1]]]], dtype=tf.float32)
out = tf.nn.avg_pool2d(input=y, ksize=1, strides=1, padding='SAME')
print(out.numpy())
# [[[[1. 0. 1.]
#    [0. 1. 1.]]]]

# MindSpore
import mindspore
import mindspore.nn as nn
from mindspore import Tensor

pool = nn.AvgPool2d(kernel_size=1, stride=1, pad_mode='SAME')
x = Tensor([[[[1, 0, 1], [0, 1, 1]]]], dtype=mindspore.float32)
output = pool(x)
print(output)
# [[[[1. 0. 1.]
#    [0. 1. 1.]]]]