This project demonstrates the usage of ZenML for MLOps (Machine Learning Operations). It includes a ZenML pipeline for training a model and a Streamlit web application for making predictions using the trained model.
Docker Installation 🔗
Use the this link to install docker desktop.
To get started, clone this Git repository to your local machine:
git clone https://github.com/yo-harsh/MLOps.git
cd MLOps
Build and run the Docker containers using Docker Compose:
docker compose up --build
This command will build the necessary Docker images and start the services.
As we know making the ml model involve many steps and in big projects new data might come with new trends and relations in some case the old model predictions accuracy might drop and the old model might not perform well; this is where MLOps come in play.
- We can make process form cleaning data to predicting the data fully auto and with little efforts we can even devlop it.
- ZenML helps with caching a pipeline so if the error happens we might end up saving the time.
- We can test the different model accuracy and performance and select the best one.
- As times passes new data will keep bundling up therefor using MLOps is best way to be sure about your prediction.
- MLOps not only helps with new data but also help keep your development cycle structured giving you CI/CD.
Visit localhost:8080 to access the ZenML Dashboard. Here, you can view and manage your machine learning pipelines.
The Streamlit web application is available at localhost:8501. Use this app to interact with the trained model and make predictions.
don't run docker compose up --build again just first time after that run "docker compose up"
Please note that Docker is required to run this project. Make sure to install Docker before attempting to build and run the project.
Feel free to provide feedback and suggestions.