AIPND Final Project

PyTorch transfer learning example developed as part of Udacity's AI Programming with Python Nanodegree program.

Getting Started

Environment Python 3.6.5:

Numpy PyTorch TorchVision

Usage

python train.py [-h] [--save_dir SAVE_DIR] [--arch ARCH]
                [--learning_rate LEARNING_RATE]
                [--hidden_features HIDDEN_FEATURES] [--gpu]
                data_dir

positional arguments:

data_dir Directory used to locate source images

optional arguments:

-h, --help show this help message and exit --save_dir SAVE_DIR Directory used to save checkpoints --arch {vgg19, alexnet} Model architecture to use for training --learning_rate LEARNING_RATE Learning rate hyperparameter --hidden_features HIDDEN_Features Number of hidden units hyperparameter

--gpu Use GPU for training

Examples

The following will train a vgg model on the GPU:

python train.py flowers --arch VGG --gpu 

Usage

python predict.py  [-h] [--top_k TOP_K] [--gpu] image_path filepath

positional arguments:

image_path Input image filepath Model checkpoint file to use for prediction

optional arguments:

-h, --help show this help message and exit --top_k TOP_K Return top k most likely classes --gpu Use GPU for prediction

Examples

The following will return the most likely class using a VGG checkpoint executing on the GPU:

python predict.py flowers/test/14/image_0814.jpg VGG_checkpoint.pth --gpu