This was built with PyTorch/FastAI for the Machine Learning part and Flask as a Webserver. For containerization and easy deployment I used Docker. The dataset used for building the model is Food-101 Dataset which is provided by FastAI it contains many images of food and i choose the images of Samosa and Burger for building a CNN Classifier.
Food101 Dataset link : https://data.vision.ee.ethz.ch/cvl/datasets_extra/food-101/
Make sure you have installed latest libraries of Pytorch,fastai and your gpu is cuda enabled and its compute capability is above 3.0 check here : https://developer.nvidia.com/cuda-gpus
Clone the project
git clone https://github.com/ManthanShettigar/SamosaBurger-CNN-Classifier.git
Go to the project directory
cd my-project
Run model via jupyter-notebook
jupyer-notebook
Command to launch the container:
docker build -t cburger_samosa_classifier . && docker run --rm -it -p 5000:5000 burger_samosa_classifier