Trained an image classifier to classify different species of flowers. It is a command line application that takes in input the location of dataset, the classifier architecture (VGG, Densenet, AlexNet), and other parameters to train a classifier on flower dataset.
Basic usage using default settings
python train.py ./flowers
To change the architecture
python train.py ./flowers --arch "densenet"
To change other parameters
python train.py ./flowers --learning_rate 0.01 --hidden_units 512 --epochs 20 --dropout 0.5 --gpu --save_dir checkpint.pth
Basic usage using default settings using a test image sample
python predict.py ./flowers/test/20/image_04910
To change other parameters using a test image sample
python predict.py ./flowers/test/20/image_04910 --category_names cat_to_name.json --top_k 10 --gpu
Didn't include the flowers dataset here as it has large size