Mastering Python for Finance - Second Edition

Phase 1 – assembling the graph

The following steps describe the process of setting up a TensorFlow graph:

1. Create placeholders for inputs and labels with the following codes:

In [ ]:    import tensorflow as tf     num_features = x_train.shape[1]     x = tf.placeholder(dtype=tf.float32, shape=[None, num_features])    y = tf.placeholder(dtype=tf.float32, shape=[None])

A TensorFlow operation starts with placeholders. Here, we defined two placeholders x and y for containing the network inputs and outputs, respectively. The shape parameter defines the shape of the tensor to be fed, with None meaning that the number of observations is unknown at this point. The second dimension of x is the number of features that we have, reflected in the num_features variable. ...