Tensorflow.js is an open-source library developed by Google for running machine learning models and deep learning neural networks in the browser or node environment.
The tf.data.Dataset.shuffle() method randomly shuffles a tensor along its first dimension.
Syntax:
tf.data.Dataset.shuffle(
buffer_size, seed=None,
reshuffle_each_iteration=None
)
Parameters:
- buffer_size: This is the number of elements from which the new dataset will be sampled.
- seed[optional]: It is an optional parameter used to create a random seed for the distribution, to see the same results use same seed.
- reshuffle_each_iteration: A Boolean, which is true indicates that the dataset should be pseudo randomly reshuffled each time it is iterated over. Default value is true. It is Optional parameter.
Return Value: A tensor with same shape and data type as value, but shuffled along its first dimension.
Example 1: In this example first we will create a tensor and then shuffle it, In this example reshuffle_each_iteration is True
Javascript
async function shuffle() { // Creating a Tensor const a = tf.data.array([1, 2, 3, 4, 5, 6]).shuffle(3); await a.forEachAsync(e => console.log(e)); //print 1 await a.forEachAsync(e => console.log(e)); //print 2 } shuffle(); |
Output:
3 4 1 2 5 6 3 4 2 5 6 1
Example 2: In this example, seed is set to an Integer, whenever a specific integer is used it will generate that specific output
Javascript
async function shuffleseed() { const a = tf.data.array([1, 2, 3]).shuffle(3, seed = 42); await a.forEachAsync(e => console.log(e)); const b = tf.data.array([1, 2, 3]).shuffle(3, seed = 42); await b.forEachAsync(e => console.log(e)); } shuffleseed(); |
Output:
2 1 3 2 1 3
Reference: https://js.tensorflow.org/api/3.6.0/#tf.data.Dataset.shuffle
