Deep Learning for Computer Vision

Training the model with data

Now you have defined the model and training operation. The next step is to start training the model with the data. During training, the gradients are calculated and the weights are updated. The variables have not yet been initialized. Next, start the session and initialize the variables using a global variable initializer:

session = tf.Session()session.run(tf.global_variables_initializer())

The preceding two lines are required for most of the examples in this book. It is assumed that the reader will use these two lines wherever required. Now the graph is ready to be fed with data and start training. Over a loop, read the data in batches and train the model. Training the model is carried out by running the session ...