Image Classifier - TensorFlow Project

In this project, we will first develop code for an image classifier built with TensorFlow, then we will convert it into a command-line application.

We'll be using this dataset from Oxford of 102 flower categories. The data for this project is quite large - in fact, it is so large that you cannot upload it onto Github.

Requirements

The project requires Python 3.x and the following Python libraries installed :

• TensorFlow
• NumPy
• matplotlib

Before you run any commands in the terminal make sure to install TensorFlow 2.0 and TensorFlow Hub using pip as shown below:

$ pip install -q -U "tensorflow-gpu==2.0.0b1"

$ pip install -q -U tensorflow_hub

Note: In order to complete this project, you will need to use the GPU. since running on your local CPU will likely not work well. You should also only enable the GPU when you need it.

Project Structure

The project is constituted by two parts:

Part 1 - Developing an Image Classifier with Deep Learning

This part of the project is broken down into multiple steps:

1. Load the image dataset and create a pipeline.
2. Build and Train an image classifier on this dataset.
3. Use your trained model to perform inference on flower images.
4. Save the model, for later applications.

Part 2 - Building the Command Line Application

After we have built and trained a deep neural network on the flower data set, it's time to convert it into an application that others can use. The application is a Python script that runs from the command line. For testing, we would use the saved Keras model that has been saved in the first part.

Specifications

The predict.py file uses the saved trained network to predict the class for an input image. It should predict the top flower names from an image along with their corresponding probabilities also.

Basic usage:

$ python predict.py /path/to/image saved_model

Options:

• --top_k : Return the top K most likely classes:

$ python predict.py /path/to/image saved_model --top_k K

• --category_names : Path to a JSON file mapping labels to flower names:

$ python predict.py /path/to/image saved_model --category_names map.json

Example:

Assume that we have a file called orchid.jpg in a folder named /test_images/ that contains the image of a flower. We also assume that we have a Keras model saved in a file named my_model.h5.

Basic usage:

$ python predict.py ./test_images/orchid.jpg my_model.h5

Options:

• Return the top 3 most likely classes:

$ python predict.py ./test_images/orchid.jpg my_model.h5 --top_k 3

• Assume we use a label_map.json file to map labels to flower names:

$ python predict.py ./test_images/orchid.jpg my_model.h5 --category_names label_map.json

Images for Testing

In the Command Line Interface workspace, there are 4 images provided in the ./test_images/ folder for you to check your predict.py module. The 4 images are:

• cautleya_spicata.jpg
• hard-leaved_pocket_orchid.jpg
• orange_dahlia.jpg
• wild_pansy.jpg