Model Serving Framework

Overview

This code is an attempt to explain in simple language how modern Model Serving frameworks works.

Features

• Customizable Model Serving: Create and deploy machine learning models easily.
• Integration with scikit-learn: Out-of-the-box support for scikit-learn models.
• Docker Integration: Build and package your models as Docker containers.
• CLI for Simplified Workflow: Command-line interface for building and deploying models.

Components

1. Framework ServingML:

   • Contains a base class for creating model instances.
   • Integrates with scikit-learn and other libraries.
   • Provides a Dockerfile.j2 template for creating Docker containers for specific models.

2. CLI (Command-Line Interface):

   • A tool to compile all necessary code into a single directory.
   • Includes a deploy command that passes the Dockerfile and files to the ServingML server.

3. ServingML Server:

   • Receives a directory with a Dockerfile, model, and other code.
   • Creates a Docker image and container for deploying the model.
   • Can run locally or as a remote service for REST API requests.

Prerequisites

• Python 3.9+
• Docker

Installation

1. Clone this repository:

   git clone https://github.com/your-username/model-serving-framework.git