Welcome to Neural-Scope’s documentation!

Neural-Scope is a comprehensive tool for analyzing and optimizing machine learning models. It provides a robust CI/CD integration that enables automated model optimization as part of ML workflows.

Contents:

• Installation
  • Requirements
  • Optional Dependencies
  • Development Installation
• Quickstart
  • Analyzing a Model
  • Optimizing a Model
  • Security Analysis
  • MLflow Integration
  • Command Line Interface
• User Guide
  • Model Analysis
• Neural-Scope MLflow Integration Guide
  • Table of Contents
  • Introduction
  • Prerequisites
  • Setting Up MLflow
  • Tracking Model Analysis
  • Tracking Model Optimization
  • Tracking Security Analysis
  • Tracking Robustness Testing
  • Using the Model Registry
  • Comparing Models
  • CI/CD Integration
  • Troubleshooting
  • Conclusion
• Neural-Scope CI/CD Integration Guide
  • Table of Contents
  • Introduction
  • Prerequisites
  • GitHub Actions Integration
  • GitLab CI Integration
  • Jenkins Integration
  • Azure DevOps Integration
  • CircleCI Integration
  • Travis CI Integration
  • Custom Workflows
  • Best Practices
  • Troubleshooting
  • Conclusion
• Neural-Scope Metrics Explained
  • Model Analysis Metrics
  • Optimization Metrics
  • Security Metrics
  • Adversarial Robustness Metrics
  • Attack-Specific Metrics
  • Versioning Metrics
  • How to Use These Metrics
• Neural-Scope Metric Analysis Document
  • Overview
  • 100% Accurate and Deterministic Metrics
  • Metrics with Potential Variability
  • Recommendations for CI/CD Integration
  • Conclusion
• Publishing Neural-Scope to PyPI
  • Prerequisites
  • Preparing for Publication
  • Publishing to Test PyPI
  • Publishing to PyPI
  • Automating Publication with GitHub Actions
  • Troubleshooting
  • Conclusion
• Contributing
  • Setting Up Development Environment
  • Code Style
  • Testing
  • Documentation
  • Pull Request Process
  • Reporting Issues
  • Code of Conduct
  • License
• Changelog
  • Version 0.3.0 (2023-07-15)
  • Version 0.1.0 (2023-01-10)

Features

• Automated Model Optimization: Analyze, optimize, and validate ML models automatically

• CI/CD Integration: Seamlessly integrate with GitHub Actions, GitLab CI, Jenkins, and Azure DevOps

• Pre-trained Model Support: Analyze models from PyTorch Hub, TensorFlow Hub, Hugging Face, SageMaker, and more

• Advanced Security Analysis: Detect vulnerabilities with sophisticated detection algorithms

• Adversarial Robustness: Test model resilience against FGSM, PGD, and other attacks

• Model Versioning: Track model versions and promote models through staging and production

• MLflow Integration: Track optimization results and model performance over time

Indices and tables

• Index

• Module Index

• Search Page