This repo comprises of training DenseNet from scratch on cifar10 and then using Flask for Model serving.
The directory is defined as
To setup the environment
python3 -m pip install -U virtualenv # install the virtualenv
virtualenv -p python3 dev_env # creating an environment
source dev_env/bin/activate # activate environment
git clone https://github.com/entiretydotai/Image-Classification-and-Deployment #clone the repo
cd Image-Classification-and-Deployment
pip install -r requirements.txt # install the requirements
we need Cifar-10 data for inference. Execute the below to generate 10 random images from cifar-10 test data.
cd data
python generate_valid_data.py
For Serving this model as a website. First generate test images by executing above commands and then.
python flask_api/app.py