Function Differences with tf.image.random_flip_left_right

tf.image.random_flip_left_right

tf.image.random_flip_left_right(
    image,
    seed=None
)

For more information, see tf.image.random_flip_left_right.

mindspore.dataset.vision.RandomHorizontalFlip

class mindspore.dataset.vision.RandomHorizontalFlip(
    prob=0.5
)

For more information, see mindspore.dataset.vision.RandomHorizontalFlip.

Differences

TensorFlow: Randomly flip the image horizontally with a probability of 0.5. The random seed can be specified by the input parameter.

MindSpore: Randomly flip the image horizontally with the specified probability. The global random seed can to be set through mindspore.dataset.config.set_seed.

Code Example

# The following implements RandomHorizontalFlip with MindSpore.
import numpy as np
import mindspore.dataset as ds

ds.config.set_seed(57)
image = np.random.random((28, 28, 3))
result = ds.vision.RandomHorizontalFlip(prob=0.5)(image)
print(result.shape)
# (28, 28, 3)

# The following implements random_flip_left_right with TensorFlow.
import tensorflow as tf

image = tf.random.normal((28, 28, 3))
result = tf.image.random_flip_left_right(image, seed=57)
print(result.shape)
# (28, 28, 3)