Releases: harvard-edge/cs249r_book
Releases Β· harvard-edge/cs249r_book
v0.4.2: Dark Mode & UX Enhancements
Actual Release Date: November 2, 2025
A feature-rich patch release introducing comprehensive dark mode support and extensive UX improvements.
π― Key Highlights
π Dark Mode Support (Major Feature)
- User-Controlled Toggle: Complete dark mode implementation with clear sun/moon icons
- Comprehensive Styling: Proper contrast for all UI elements including tables, code blocks, callouts, sidebar, and TOC
- System Preference Support: Respects OS theme preferences with manual override option
- Enhanced Accessibility: Improved legibility across all content types with careful color selection
- Persistent Settings: Theme choice remembered across sessions
π¨ UI/UX Enhancements
- Announcement Banner: Styled dismiss button and improved visibility in both light and dark modes
- Better Icons: Clear sun/moon icons for theme toggle replacing generic symbols
- TikZ Diagrams: White backgrounds in dark mode for optimal diagram visibility
- Footer Styling: Improved contrast and readability in dark mode
- Book Preview Card: Enhanced dark mode styling for better presentation
π§ Build System Improvements
- Windows Compatibility: Resolved PDF build errors affecting LaTeX/TikZ rendering
- Workflow Reliability: Enhanced Windows container and bare-metal build consistency
- Python 3.13 Support: Full compatibility with latest Python versions
- CI Enhancements: Better timeout handling and retry logic
π Content Refinements
- Documentation Updates: Improved README with special features and contribution information
- SocratiQ Clarity: Enhanced functionality descriptions
- Link Fixes: Corrected ML kits links and references
- Lab Improvements: Fixed IMU data descriptions and added required dependencies
π Developer Experience
- Intelligent Release Notes: AI-powered content analysis for automated changelog generation
- Version Automation: Streamlined publish workflow with automatic version updates
- Workflow Optimization: Removed concurrency restrictions for faster iterations
π Changes Summary
- 120+ commits focused on dark mode and UX
- 30+ styling fixes across all UI components
- 10+ workflow improvements for better developer experience
- Community contributions from multiple external reviewers
π Notable Bug Fixes
- Fixed critical dark mode readability issues in tables and code blocks
- Resolved callout header text visibility in dark mode
- Corrected footnote and margin text styling
- Fixed announcement banner transparency issues
- Improved sidebar text color specificity
- Resolved Windows PowerShell variable escaping issues
π Quick Links
- π Web Version
- π PDF Download
- π EPUB Download
- π Detailed Changelog
π Release Information
- Release Date: November 2, 2025
- Development Period: October 20 - November 2, 2025
- Previous Version: v0.4.1
- Type: Patch Release (UX Features & Polish)
This release dramatically improves the reading experience with comprehensive dark mode support while maintaining the high-quality content and build reliability standards. The dark mode implementation represents the most requested feature from the community and sets a new standard for technical textbook accessibility.
v0.4.1: Content Quality & Infrastructure Improvements
Actual Release Date: October 20, 2025
A focused patch release addressing content accuracy, infrastructure stability, and build system improvements.
π― Key Highlights
π Content Quality Improvements
- Enhanced Visualizations: New TikZ figures added to chapters 10, 15, 17, and 20 for improved clarity
- Quiz Enhancements: Added comprehensive quizzes to the Model Optimizations chapter
- Accuracy Fixes: Corrected equation formatting and CNN architecture question ordering
- Pedagogical Additions: Connected benchmarking concepts to Goodhart's Law with detailed footnote
π οΈ Infrastructure Enhancements
- Build System: Removed workflow concurrency bottlenecks for faster iteration
- Asset Management: Improved organization with dedicated
/assets/downloads/directory - Bibliography: Formatted all bibliography files with bibtex-tidy for consistency
- EPUB Generation: Resolved container compression failures
π Bug Fixes
- Fixed small typos in Chapter 10 (Model Optimizations) and chapter content
- Corrected transpose notation in equations (issue #974)
- Fixed LaTeX typo in Section 4.7.3
- Updated frameworks cross-reference data
- Resolved duplicate section issues with self-referential section checker
π Community Contributions
- Multiple typo fixes and formatting improvements from contributors
- Enhanced documentation based on user feedback
- Improved lab instructions clarity
π Quick Links
- π Web Version
- π PDF Download
- π EPUB Download
- οΏ½οΏ½ Detailed Changelog
π Release Information
- Release Date: October 20, 2025
- Previous Version: v0.4.0
- Type: Patch Release (Content Quality & Infrastructure)
This release represents focused improvements to content quality and build infrastructure, incorporating community feedback and enhancing the overall textbook experience.
v0.4.0: Enhanced Infrastructure and User Experience
Actual Release Date: October 9, 2025
This massive release represents the most comprehensive improvement to MLSysBook since its inception, featuring extensive content enhancements, infrastructure upgrades, and professional presentation refinements in preparation for MIT Press publication.
π― Key Highlights
π Comprehensive Content Improvements (62 files, +11 from v0.3.0)
- Educational Excellence: Extensive polish workflow applied across all 21 chapters
- Cross-Reference System: Comprehensive concept-driven links connecting related topics
- Glossary System: JSON-based glossary with interactive tooltips and cross-references
- Learning Objectives: Standardized format using Bloom's Taxonomy across chapters
- Fallacies & Pitfalls: Educational sections highlighting common misconceptions
π οΈ Infrastructure Excellence
- Smart Content Analysis: Enhanced changelog generation with intelligent filtering
- Citation Validation: Comprehensive validation hooks and bibliography improvements
- Table Formatting: Auto-alignment and professional styling automation
- Build System: Improved PDF configuration and cross-platform reliability
- Quality Assurance: Enhanced pre-commit hooks with content validation
π Professional Academic Presentation
- MIT Press Timeline: Updated announcement banner with 2026 publication schedule
- Mobile Navigation: Intelligent sidebar auto-collapse for better UX
- Cross-Platform: Enhanced compatibility across devices and formats
- Professional Polish: Publication-ready formatting and presentation
- Concept Maps: Detailed maps for enhanced understanding
π Educational Innovation
- Chapter Flow Optimization: Comprehensive narrative improvements
- Enhanced Quizzes: Context-aware generation with knowledge maps
- Interactive Elements: Improved engagement and learning tools
- Accessibility: Better mobile experience and content discovery
π Major Changes (4,842 commits)
- Content Files: 51 β 62 files (+22% growth)
- All 21 Chapters: Comprehensive polish and improvements
- Infrastructure: 100+ commits on build and deployment systems
- Quality Control: Automated formatting, citations, and validation
- Performance: Optimized build times and user experience
π Key Achievements
Content Excellence (200+ focused commits)
- Comprehensive Polish: Expert editorial workflows across all chapters
- Academic Rigor: Enhanced citations and cross-references
- Flow Optimization: Improved narrative progression
- Clarity: Removed redundancies, strengthened foundations
Technical Infrastructure (100+ commits)
- Release Automation: AI-powered content analysis systems
- Build Reliability: Enhanced Docker and cross-platform builds
- Quality Control: Automated table formatting and citation validation
- Performance: Optimized build times and user experience
Professional Standards (50+ commits)
- Publication Ready: Aligned with MIT Press requirements
- User Experience: Mobile-optimized navigation
- Community Tools: Enhanced contributor workflows
- Accessibility: Improved content discovery and navigation
π¬ Educational Impact
This release transforms MLSysBook into a professional academic resource:
- Students: Enhanced learning with improved flow and mobile optimization
- Educators: Professional-grade content with comprehensive cross-references
- Researchers: Robust citation system supporting advanced study
- Industry: Real-world perspectives with practical considerations
π Release Information
- Release Date: October 9, 2025
- Development Period: January - October 2025 (9 months)
- Previous Version: v0.3.0
- Commits: 4,842 improvements (largest release to date)
π Quick Links
- π Web Version
- π PDF Download
- π EPUB Download
- π Detailed Changelog
ποΈ Technical Details
- Build Platform: Linux with enhanced workflows
- Formats: HTML, PDF, EPUB
- Deployment: GitHub Pages
- PDF Engine: Quarto with LaTeX
This release represents over 600 commits of focused improvements in the final 2 months alone, incorporating feedback from educators, students, and academic reviewers. The enhanced infrastructure and content quality establish MLSysBook as the definitive resource for ML systems education, ready for professional academic publication by MIT Press in 2026.
v0.3.0: Enhanced Clarity and Accessibility
Actual Release Date: January 2, 2025
This release focuses on elevating content quality, improving accessibility, and strengthening the pedagogical foundation. Extensive refinements across all chapters enhance learning effectiveness for diverse audiences.
π― Key Highlights
π Content Refinements (51 files)
- Interactive Quizzes: New widget-based quiz system for active learning
- Enhanced Chapter Flow: Improved narrative progression across all chapters
- System Perspective: Strengthened systems thinking in Chapter 2 and throughout
- Clarified Explanations: Complex topics made more accessible without sacrificing depth
- Visual Improvements: Enhanced figures, diagrams, and visual aids
π οΈ Infrastructure Excellence
- Quiz System: Interactive widget-based assessment throughout chapters
- Build Reliability: Improved HTML, PDF, and EPUB generation stability
- Accessibility: Enhanced support for screen readers and alternative formats
- Performance: Faster page loads and optimized user experience
- Cross-platform: Better mobile and tablet experience
π Pedagogical Innovation
- Active Learning: Quiz widgets encourage engagement and self-assessment
- Better Readability: Improved typography and layout for comprehension
- Consistent Formatting: Unified style across all content
- Enhanced Navigation: Improved cross-references and chapter interconnections
- Multiple Formats: Optimized PDF and EPUB for offline study
π Major Changes (1,052 commits)
- Content Files: 54 β 51 files (consolidation for clarity)
- Quiz System: Interactive widget-based assessments added
- Chapter Improvements: Enhanced flow and systems perspective
- Infrastructure: Significant build and accessibility improvements
- Bug Fixes: Hundreds of refinements and corrections
π¬ Educational Impact
This release makes ML systems education more effective and accessible:
- Engaged Learning: Interactive quizzes promote active participation
- Broader Access: Accessibility features remove barriers for all learners
- Improved Comprehension: Clearer explanations support deeper understanding
- Flexible Study: Multiple high-quality formats for different preferences
π Release Information
- Release Date: January 2, 2025
- Development Period: June 2024 - January 2025 (7 months)
- Previous Version: v0.2.0
- Commits: 1,052 improvements focusing on quality and accessibility
π Quick Links
- π Web Version
- π PDF Download
- π EPUB Download
- π GitHub Repository
ποΈ Technical Details
- Build Platform: Quarto with enhanced workflows
- Formats: HTML, PDF, EPUB
- New Features: Interactive quiz widgets, improved accessibility
- Performance: Optimized for speed and reliability
This release represents a commitment to educational excellence, making ML systems education more engaging, accessible, and effective for students worldwide through interactive learning tools and enhanced content clarity.
v0.2.0: Scaling Beyond TinyML
Actual Release Date: June 20, 2024
This major release marks a pivotal transformation from TinyML-focused curriculum to comprehensive machine learning systems education. The scope expands far beyond embedded AI while preserving foundational principles, reflecting that ML systems principles are universal across deployment scales.
π― Key Highlights
π Content Expansion (54 files, +9 from v0.1.0)
- Broadened Scope: Evolution from TinyML-specific to full-spectrum ML systems
- Robust AI Chapter: New comprehensive chapter on fault-tolerant and resilient systems
- Enhanced Labs: Expanded hardware support with Seeed XIAO ESP32S3
- Shared Lab Content: Reusable modules across different hardware platforms
- Getting Started Guides: Improved onboarding for students and educators
π οΈ Infrastructure Improvements
- Build System: Enhanced reliability across HTML, PDF, and EPUB formats
- Navigation: Added GitHub stars counter and improved banner system
- Asset Management: Better organization of figures and media files
- Documentation: Expanded getting started and setup guides
π Expert Contributions
- Marcelo Rovai: Extensive lab updates and hardware platform support
- Community Contributors: Growing network of educators and practitioners
- Hardware Partners: Integration with Seeed Studio and other vendors
π Major Changes (932 commits)
- Content Files: 45 β 54 files (+20% growth)
- Hardware Platforms: Added Seeed XIAO ESP32S3 support
- Shared Labs: New reusable lab components
- Infrastructure: Numerous build and deployment improvements
- Bug Fixes: 100+ fixes for references, figures, and formatting
π¬ Educational Impact
This release positions the textbook as a comprehensive ML systems resource:
- Comprehensive Systems Thinking: Beyond individual algorithms to full systems
- Cross-Platform Expertise: From embedded devices to cloud-scale deployments
- Practical Skills: Real-world engineering and operations knowledge
- Industry Relevance: Preparation for production ML engineering roles
π Release Information
- Release Date: June 20, 2024
- Development Period: December 2023 - June 2024 (6 months)
- Previous Version: v0.1.0
- Commits: 932 improvements and additions
π Quick Links
- π Web Version
- π PDF Download
- π EPUB Download
- π GitHub Repository
ποΈ Technical Details
- Build Platform: Quarto with enhanced workflows
- Formats: HTML, PDF, EPUB
- New Features: Improved navigation, star counter, shared labs
- Performance: Faster builds and better reliability
This release represents a fundamental shift in scope, establishing MLSysBook as the definitive resource for ML systems education across computing disciplines, from embedded devices to cloud-scale deployments.
v0.1.0: Foundational TinyML Systems
Actual Release Date: December 12, 2023
This initial release establishes the foundations of Tiny Machine Learning (TinyML) systems education, presenting a comprehensive introduction to resource-constrained AI deployment and embedded machine learning.
π― Key Highlights
π Core Content (45 files)
- Complete TinyML Curriculum: 15+ foundational chapters covering embedded AI systems
- Deep Learning Primer: Neural network fundamentals optimized for embedded systems
- Hardware Acceleration: Coverage of ASICs, FPGAs, and microcontroller deployments
- Model Optimization: Quantization, pruning, and compression for resource constraints
- Benchmarking Framework: MLPerf and performance evaluation methodologies
π οΈ Technical Infrastructure
- Quarto-based Publishing: Modern, reproducible academic publishing framework
- Multi-format Output: Synchronized HTML, PDF, and EPUB generation
- Reference Management: Comprehensive bibliography system
- Interactive Labs: Practical examples with Arduino, ESP32, and Raspberry Pi
π Educational Foundation
- University-level Curriculum: Suitable for CS and engineering programs
- Hands-on Learning: Real hardware deployment on embedded platforms
- Open Source: Fully accessible content and collaborative development
- Academic Rigor: Peer-reviewed content with extensive citations
π Content Overview
- Total Files: 45 QMD source files
- Core Chapters: 15+ comprehensive chapters
- Hardware Platforms: Arduino Nicla Vision, ESP32, Raspberry Pi
- Lab Exercises: Multiple hands-on deployment tutorials
π Release Information
- Release Date: December 12, 2023
- Development Period: 2023
- Content Focus: TinyML fundamentals and embedded AI
- Target Audience: Students, educators, embedded systems engineers
π Quick Links
- π Web Version
- π PDF Download
- π EPUB Download
- π GitHub Repository
ποΈ Technical Details
- Build Platform: Quarto with R and Python
- Formats: HTML, PDF, EPUB
- License: Open source educational content
- PDF Engine: LaTeX with custom templates
This foundational release established the first comprehensive academic textbook dedicated to TinyML systems, empowering students and practitioners to master embedded ML from theory to deployment on resource-constrained devices.