Classification of images using TensorFlow. Pre-trained Inception v3 model is used to implement Transfer learning that retrains from the existing weights of fully-trained model for a set of categories like ImageNet for new classes of images.
The file retrain.py is to be run to train the model and generate graph and label files.
The command is:
python retrain.py --output_graph=retrained_graph.pb --output_labels=retrained_labels.txt --image_dir=path_to_dataset
The file label_image.py is to predict the result for test image
The command is:
python label_image.py path_to_image_image.jpg