This Web application implements the GUI for flower classification on the dataset from http://www.robots.ox.ac.uk/~vgg/data/flowers/102/ coded in Pytorch. It involves classification of flowers into 102 categories occuring mostly in United Kingdom. It is done as a part of Pytorch Deep Learning scholarship challenge lab project.
The step-by-step procedure of the Project:
- Collection of dataset from the link mentioned at the top;
- Data preprocessing: Augmentation being applied to train set;
- Training the classifier part of the model Densenet121 pretrained on Imagenet;
- The model built scores 98.3% on the validation set;
- Saving the checkpoint containing the models parameteres;
- Building a Flask Application using the inference from pretrained model;
NOTE : The whole Machine Learning pipeline is implemented in the jupyter notebook provided in the repository.
These instructions assume you have git installed for working with Github from command window.
- Clone the repository, and navigate to the downloaded folder. Follow below commands.
git clone https://github.com/pswaldia/Flower_identification
cd Flower_identification
- Creating python virtual environment using virtualenv package using following lines of code.
NOTE: For this step make sure you have virtualenv package installed.
virtualenv venv
source venv/bin/activate
- Install few required pip packages, which are specified in the requirements.txt file .
pip3 install -r requirements.txt
Run the following code:
flask runNow copy the URL of the local host that will appear on your terminal and run it in browser.
