How to use windows created by the Dataset.window() method in TensorFlow 2.0?

Question

I'm trying to create a dataset that will return random windows from a time series, along with the next value as the target, using TensorFlow 2.0.

I'm using Dataset.window(), which looks promising:

import tensorflow as tf  dataset = tf.data.Dataset.from_tensor_slices(tf.range(10)) dataset = dataset.window(5, shift=1, drop_remainder=True) for window in dataset:     print([elem.numpy() for elem in window]) 

Outputs:

[0, 1, 2, 3, 4] [1, 2, 3, 4, 5] [2, 3, 4, 5, 6] [3, 4, 5, 6, 7] [4, 5, 6, 7, 8] [5, 6, 7, 8, 9] 

However, I would like to use the last value as the target. If each window was a tensor, I would use:

dataset = dataset.map(lambda window: (window[:-1], window[-1:])) 

However, if I try this, I get an exception:

TypeError: '_VariantDataset' object is not subscriptable 

Answer 1

The solution is to call flat_map() like this:

dataset = dataset.flat_map(lambda window: window.batch(5)) 

Now each item in the dataset is a window, so you can split it like this:

dataset = dataset.map(lambda window: (window[:-1], window[-1:])) 

So the full code is:

import tensorflow as tf  dataset = tf.data.Dataset.from_tensor_slices(tf.range(10)) dataset = dataset.window(5, shift=1, drop_remainder=True) dataset = dataset.flat_map(lambda window: window.batch(5)) dataset = dataset.map(lambda window: (window[:-1], window[-1:]))  for X, y in dataset:     print("Input:", X.numpy(), "Target:", y.numpy()) 

Which outputs:

Input: [0 1 2 3] Target: [4] Input: [1 2 3 4] Target: [5] Input: [2 3 4 5] Target: [6] Input: [3 4 5 6] Target: [7] Input: [4 5 6 7] Target: [8] Input: [5 6 7 8] Target: [9]