All the codes were made during my course with udacity #AIPND .
Using transfere learning in pytorch to create a classifier to classify flowers and dog breeds.
the main project program is in the /FinalProject file.
Use this program via command line to train a neural network using transfere learning to classify what you want by using train.py use it to train the models you need,
I've put densenet and vgg as options for the archetictures available for you to chose from.
Put the arch. and know the last layer so that you can create the classifying layer as the first layer and the number of classes for the last layer .
Choose the dataset location and deduce the cpu or gpu extention if you want and start training .
Then you can use the checkpoint created by the train.py to predict any kind of images you want, just insert the checkpoint address and the image address and wait
Sorry that the code is not perfectly clean i made it in a hurry because i was having my exams and 2 other courses at the same time.
Thanks :).