Going forward, AI algorithms will be incorporated into more and more everyday applications. For example, you might want to include an image classifier in a smart phone app. To do this, you'd use a deep learning model trained on hundreds of thousands of images as part of the overall application architecture. A large part of software development in the future will be using these types of models as common parts of applications. This classifier has been tested on a dataset containing 102 different species of flowers. The dataset can be found here. This is the final project for Udacity's AI Programming with Python Nanodegree.
This project has been written in Python 3.6.5. It should work with Python versions 3.* , Pytorch, PIL. To work on or view the project, it is recommended to use Jupyter Notebooks. The classifier.pth file contains the trained classifier, therefore it can be directly used to classify a flower on the terminal also.
-
Image Classifier Project.ipynb: Contains the Python code for the classifier developed on Jupyter Notebook.
-
classifier.pth: It is the finished classfier trained on a dataset of 102 flower species using GPU. It can be directly used for predicition purposes or can be retrained. Since it is a huge file I have compressed them into multiple smaller files which can be extracted from classifier.rar.
-
cat_to_name.json: This JSON file is used to map the flower numbers and flower names.
-
train.py: Application to provide new hyperparameters, training dataset etc., to retrain the network and save it as classifier.pth
-
predict.py: It is used to classify an input image on the basis of the trained network.
- Utkarsh Lath
- Udacity
An interesting and insightful project which would not have been possible without Udacity and its mentors.
The project is licensed under MIT License.