Questions tagged [tensorflow-datasets]
The functional API in TensorFlow's `tf.data` module enables the creation of input pipelines by utilizing the `tf.data.Dataset` and `tf.data.Iterator` classes.
Consider the dataframe df: revenue 2016-11-05 -0.352631 2016-11-06 -0.438142 2016-11-07 -0.470228 2016-11-08 -0.487672 2016-11-09 -0.491773 ... ... 2020-11-16 -0.624413 2020-11-17 -0.640770 2020-11-18 -0.606660 2020-11-19 -0.596660 20 ...
After following a tutorial on creating custom datasets for TensorFlow at this link, I attempted to create my own dataset using the `tfds` command. Upon entering the command `tfds new my_dataset`, I encountered the following error: tfds new my_dataset Trac ...
I am attempting to import the MNIST dataset and separate it into three distinct datasets: train, test, and validation. The correct syntax for this operation is as follows: (train, test, valid) = tfds.load('mnist', split=['train[:90%]','train[90%:]', 'test' ...
Even after reading the documentation on tf.keras.Model.fit and tf.data.Dataset, I am still unsure about whether I should call repeat and batch on the dataset object when passing it to the fit function. Should I provide the batch_size and epochs arguments t ...