SBYB - Step-By-Your-Byte

A comprehensive machine learning library that unifies the entire ML pipeline.

Overview

SBYB (Step-By-Your-Byte) is a Python library designed to provide a unified, offline-capable machine learning toolkit that outperforms existing solutions like TensorFlow and Keras. It integrates the entire ML pipeline from data preprocessing to model deployment in a single, cohesive package.

Key Features

• Unified Data Preprocessing: Automatic handling of missing values, outliers, encoding, and scaling
• Task Type & Data Type Auto-detection: Intelligent identification of ML tasks and data characteristics
• AutoML Engine: Automated model selection, hyperparameter tuning, and ensemble creation
• Evaluation & Explainability: Comprehensive metrics and model interpretation tools
• Deployment & Serving: Easy model export and deployment options
• Zero-code UI Generation: Automatic creation of user interfaces for models
• Project Scaffolding: Quick setup of new ML projects with best practices
• EDA Tools: Powerful data profiling and visualization capabilities
• Plugin System: Extensible architecture for custom components
• Local Experiment Tracking: Track, compare, and visualize ML experiments
• CLI & Programmatic API: Multiple interfaces for different workflows

Installation

pip install sbyb

Quick Start

Using the CLI

# Create a new project
sbyb project create --name my_project --template classification

# Run AutoML on a dataset
sbyb automl run --data data.csv --target target_column

# Generate a UI for a model
sbyb ui generate --model model.pkl --output ui_app

Using the API

from sbyb.api import SBYB

# Initialize SBYB
sbyb = SBYB()

# Preprocess data
import pandas as pd
data = pd.read_csv("data.csv")
preprocessed_data = sbyb.preprocess_data(data)

# Run AutoML
result = sbyb.run_automl(
    data=preprocessed_data,
    target="target_column",
    output_dir="output"
)

# Generate UI
sbyb.generate_ui(
    model=result.model,
    output_dir="ui_app",
    ui_type="dashboard",
    framework="streamlit"
)

Documentation

For detailed documentation, visit https://sbyb.readthedocs.io/

License

MIT License