UNPKG

greed.js

Version:

Run Python libraries in the browser with WebGPU acceleration - PyTorch, NumPy, and more. Modular architecture with full backward compatibility.

64 lines (50 loc) 2.18 kB
# Changelog All notable changes to this project will be documented in this file. The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/), and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html). ## [2.0.1] - 2024-07-25 ### Added - Comprehensive error handling for WebGPU operations with recovery mechanisms - Advanced memory management with automatic cleanup and pressure monitoring - Enhanced security validation with detailed threat detection - Performance monitoring and diagnostics capabilities - TypeScript declarations for better developer experience ### Changed - **BREAKING**: Extracted PyTorch polyfill from inline code to separate module for better performance - Replaced all console calls with proper structured logging system - Enhanced buffer management with emergency cleanup procedures - Improved modular architecture with better separation of concerns ### Fixed - Memory leaks in WebGPU buffer management - Missing error handling in GPU operations - Inconsistent logging throughout the codebase - Performance issues caused by massive inline PyTorch polyfill ### Performance - ~75% reduction in main thread PyTorch overhead through polyfill extraction - Intelligent memory pressure monitoring prevents out-of-memory errors - Optimized buffer pooling and cleanup strategies - Improved error recovery and graceful degradation ### Security - Enhanced input validation and threat detection - Safe polyfill installation with security validation - Comprehensive pattern matching for dangerous operations - Structured error reporting without sensitive data exposure ## [2.0.0] - 2024-07-23 ### Added - Complete rewrite with modular architecture - WebGPU acceleration support - Enhanced PyTorch compatibility - Comprehensive testing suite - Security validation system ### Changed - Migrated from monolithic to modular design - Improved performance and reliability - Better error handling and logging ### Removed - Legacy v1 compatibility (migration guide available) ## [1.0.0] - 2024-01-15 ### Added - Initial release - Basic Python execution in browser - PyTorch polyfill support - WebAssembly integration