Serverless Deep Learning Deploy examples

Training the Malaria detection model

1. Install the dependencies of this project with:

pip install -r requirements.txt

WARNING! Make sure to match your Python version according to this:

• Azure Functions uses Python 3.6

• AWS Lambda uses Python 3.6 or Python 3.7

• GCP Functions uses Python 3.7

2. Download the dataset from Kaggle, using the Kaggle API:

kaggle datasets download -d iarunava/cell-images-for-detecting-malaria

3. Unzip the downloaded .zip file to the root of this project with:

unzip cell-images-for-detecting-malaria.zip

4.Rename the directories and copy random files to validation split:

sh prepare_dataset.sh

4. Train the model:

python train.py

Running the demo server

Enter the /example_web_page directory and install the libs:

npm i

To start the server:

node server.js

The demo page will be served in localhost:4000