The project is live here check it out.
This project involves End-to-end ML-project. from developing a Convolutional Neural Network (CNN) model based on the ResNet50 architecture to detect COVID-19 from X-ray images to building the API with the help if FastAPI. with containerinzation using Docker
The model was trained on a dataset with 148 examples, with F1 score 0.8696 and achieved an accuracy of 97.5%.
-
code.ipynb
: Contains all the code for data preprocessing, model training, and evaluation. -
COVID DataSet.zip
: The dataset used for training and testing the model. -
Dockerfile
with the containerization of the project. -
The
/app
directory have all the file related to the Ops- With
/app/main.py
as the file containing FastAPI. - With
/app/model.h5
the trained model. - With
/app/requirement.txt
having the required libraries.
- With
- Accuracy: 97.5%
- F1 Score: 0.8696
- Training Data: 148 X-ray images
-
- Just download the
app/app.py
file - Run the following command in the terminal
by runing these commands a
pip install streamlit streamlit run app.py
this type of window will be shown and you can open the app by following the link provided.You can now view your Streamlit app in your browser. Local URL: http://localhost:8501 Network URL: http://172.25.97.164:8501
- Just download the
-
- Clone the Repository: Clone this repository
- Open the Repository: Go to the cloned repository
- Run Docker Image: use
$ docker run covid-19-detection
in the bash. - Check the FastAPI: Check the API at the
/predict/docs
-
- Clone the Repository: Clone this repository
- Open the Repository: Go to the cloned repository
- Go to the app:use
$ cd app
in the bash. - Install required libraries run
$ pip install -r requirements.txt
- Run the app: use
$ uvicorn main:app
in the bash. - Check the FastAPI: Check the API at the
/predict/docs
This project is licensed under the Apache License 2.0. See the LICENSE file for details.