A simple python program to classify 102 flower data using Pytorch.
The program prompts user to choose a json file (which has a mapping to flower names), then a flower picture, then attempts to classify it. The top 5 classes are shown in a bar chart.
predict_out.py can be run directly using the attached checkpoint, or train.py can be used to generate a checkpoint first.
This program has been tested using 2 architectures: densenet, alexnet Although the checkpoint attached is for densenet only.
This program has been implemented as a part of Udacity's Pytorch Challenge.