Flower Species Image Classifier

Project code for Udacity's AI Programming with Python Nanodegree program. In this project, students first develop code for an image classifier built with PyTorch, then convert it into a command line application.

Project Structure

The repository contains the following files:

• Image Classifier Project.ipynb: A Jupyter notebook where the image classifier is built using transfer learning with ResNet-18. This notebook includes visualizations of the training process, such as loss and accuracy metrics over time, to provide insights into the model's performance.

• train.py: A command-line script to train a new network on a dataset and save the model as a checkpoint.

• predict.py: A command-line script to use a trained network for predicting the class of an input image.

Usage

Train

To train a new network on a dataset, use the train.py command-line script. This script will print out the training loss, validation loss, and validation accuracy as the network trains.

Basic Usage:

python train.py data_dir --save_dir save_directory

Options:

• Set directory to save checkpoints:

python train.py data_dir --save_dir save_directory

• Choose architecture:

python train.py data_dir --arch "resnet18"

• Set hyperparameters:

python train.py data_dir --learning_rate 0.01 --hidden_units 250 --epochs 3

• Use GPU for training:

python train.py data_dir --gpu

Predict

Basic Usage:

python predict.py /path/to/image checkpoint

Options:

• Return top K most likely classes:

python predict.py input checkpoint --top_k 3

• Use a mapping of categories to real names:

python predict.py input checkpoint --category_names cat_to_name.json

• Use GPU for inference:

python predict.py input checkpoint --gpu

Instructions

1. Ensure that the necessary Python libraries are installed.

2. Prepare your dataset and place it in the specified directory.

3. Use train.py to train the model and save the checkpoint.

4. Use predict.py to make predictions on new images.

License

This project is licensed under the MIT License - see the LICENSE file for details.