Skip to content

Latest commit

 

History

History
96 lines (77 loc) · 2.92 KB

File metadata and controls

96 lines (77 loc) · 2.92 KB

Changelog

All notable changes to this project will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

[0.1.10] - 2025-04-13

Added

  • Enhanced benchmarking capabilities with CPU performance analysis
  • New optimization improvements for data loading

Changed

  • Updated benchmarking infrastructure for better performance analysis
  • Improved README documentation

[0.1.2] - 2025-03-30

Changed

  • Updated package name to match PyPI repository
  • Fixed GitHub Actions workflow for automated releases

[0.1.1] - 2025-03-30

Added

  • Comprehensive examples demonstrating key features
  • Memory management with allocation stack
  • Enhanced batch size calculation for multi-GPU scenarios
  • Improved error handling and recovery
  • Progress tracking with detailed statistics
  • Data caching with automatic cleanup
  • Support for various data formats (CSV, JSON, Images)

Changed

  • Improved memory management with better allocation tracking
  • Enhanced batch size calculation for multi-GPU scenarios
  • Better error handling and recovery mechanisms
  • Updated documentation with examples and tutorials
  • Optimized data loading performance

Fixed

  • Memory leaks in cleanup operations
  • Batch size calculation for device distribution
  • Progress tracking accuracy
  • Test reliability and coverage
  • Multi-GPU batch distribution issues

Documentation

  • Added comprehensive examples
  • Updated README with installation and usage instructions
  • Added API documentation
  • Included example requirements and setup guide
  • Added detailed feature documentation

Examples

  • Added data loading demo with multiple formats
  • Included sample data generation scripts
  • Added configuration examples
  • Demonstrated key features with real-world scenarios
  • Added multi-GPU usage examples

[0.1.0] - 2025-03-19

Added

  • Initial release of JAX DataLoader
  • Support for multiple data formats (CSV, JSON, Images)
  • Multi-GPU support with automatic batch distribution
  • Memory management with automatic batch size tuning
  • Progress tracking and statistics
  • Data caching and prefetching
  • Type hints and documentation

Changed

  • Improved memory management with allocation stack
  • Enhanced batch size calculation for multi-GPU scenarios
  • Better error handling and recovery
  • Updated documentation with examples

Fixed

  • Memory leaks in cleanup operations
  • Batch size calculation for device distribution
  • Progress tracking accuracy
  • Test reliability and coverage

Documentation

  • Added comprehensive examples
  • Updated README with installation and usage instructions
  • Added API documentation
  • Included example requirements and setup guide

Examples

  • Added data loading demo with multiple formats
  • Included sample data generation scripts
  • Added configuration examples
  • Demonstrated key features with real-world scenarios