Deep Learning with TensorFlow

Using the retrained model

The retraining script that we just used will write out a custom version of Inception V3 network with the final layer retrained to our flower dataset to tensorflow_files/output_graph.pb and you will find a text file that contains the labels in tensorflow_files/output_labels.txt.

The following are the steps to classifying an image:

1. We are going to write a Python script that can use the previous network and use it to classify an image.

        import tensorflow as tf, sys

2. You will be sending the image to be classified as a parameter:

        provided_image_path = sys.argv[1]

3. Then we will read the image data: 

        provided_image_data = tf.gfile.FastGFile(provided_image_path,         'rb').read()

4. Load labels file:

 label_lines = [line.rstrip() ...