/TensorFlow 2.4


Randomly crops a tensor to a given size in a deterministic manner.

Slices a shape size portion out of value at a uniformly chosen offset. Requires value.shape >= size.

If a dimension should not be cropped, pass the full size of that dimension. For example, RGB images can be cropped with size = [crop_height, crop_width, 3].

Guarantees the same results given the same seed independent of how many times the function is called, and independent of global seed settings (e.g. tf.random.set_seed).

Usage Example:

image = [[[1, 2, 3], [4, 5, 6]], [[7, 8, 9], [10, 11, 12]]]
seed = (1, 2)
tf.image.stateless_random_crop(value=image, size=(1, 2, 3), seed=seed)
<tf.Tensor: shape=(1, 2, 3), dtype=int32, numpy=
array([[[1, 2, 3],
        [4, 5, 6]]], dtype=int32)>
value Input tensor to crop.
size 1-D tensor with size the rank of value.
seed A shape [2] Tensor, the seed to the random number generator. Must have dtype int32 or int64. (When using XLA, only int32 is allowed.)
name A name for this operation (optional).
A cropped tensor of the same rank as value and shape size.

© 2020 The TensorFlow Authors. All rights reserved.
Licensed under the Creative Commons Attribution License 3.0.
Code samples licensed under the Apache 2.0 License.